Handbook of Research on Web 2.0, 3.0, and X.0: Technologies, Business, and Social Applications (Advances in E-Business Research Series (Aebr) Book Series) [1 ed.] 1605663840, 9781605663845, 9781605663852

As the Web continues to evolve, advances in Web technology forge many new applications that were not previously feasible

232 28 18MB

English Pages 1116 [1038] Year 2009

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

Handbook of Research on Web 2.0, 3.0, and X.0: Technologies, Business, and Social Applications (Advances in E-Business Research Series (Aebr) Book Series) [1 ed.]
 1605663840, 9781605663845, 9781605663852

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Handbook of Research on Web 2.0, 3.0, and X.0: Technologies, Business, and Social Applications San Murugesan Multimedia University, Malaysia & University of Western Sydney, Australia

Volume I

InformatIon scIence reference Hershey • New York

Director of Editorial Content: Senior Managing Editor: Assistant Managing Editor: Publishing Assistant: Typesetter: Cover Design: Printed at:

Kristin Klinger Jamie Snavely Michael Brehm Sean Woznicki Carole Coulson, Ricardo Mendoza, Kurt Smith Lisa Tosheff Yurchak Printing Inc.

Published in the United States of America by Information Science Reference (an imprint of IGI Global) 701 E. Chocolate Avenue Hershey PA 17033 Tel: 717-533-8845 Fax: 717-533-8661 E-mail: [email protected] Web site: http://www.igi-global.com/reference

Copyright © 2010 by IGI Global. All rights reserved. No part of this publication may be reproduced, stored or distributed in any form or by any means, electronic or mechanical, including photocopying, without written permission from the publisher. Product or company names used in this set are for identiication purposes only. Inclusion of the names of the products or companies does not indicate a claim of ownership by IGI Global of the trademark or registered trademark. Library of Congress Cataloging-in-Publication Data Handbook of research on Web 2.0, 3.0, and X.0 : technologies, business, and social applications / San Murugesan, editor. p. cm. Includes bibliographical references and index. Summary: "This book provides a comprehensive reference source on next generation Web technologies and their applications"--Provided by publisher. ISBN 978-1-60566-384-5 (hardcover) -- ISBN 978-1-60566-385-2 (ebook) 1. Web 2.0. 2. Social media. I. Murugesan, San. TK5105.88817.H363 2010 025.042'7--dc22 2009020544 British Cataloguing in Publication Data A Cataloguing in Publication record for this book is available from the British Library. All work contributed to this book is new, previously-unpublished material. The views expressed in this book are those of the authors, but not necessarily of the publisher.

Oh that my words were now written! Oh that were printed in a book! —JOB: 19:23 This Handbook of Research on Web 2.0, 3.0 and X.0 is our contribution to commemorate the: • • •

40th Anniversary of the Internet 20th Anniversary of the Web 5th Anniversary of Web 2.0

* The mystical symbol that appear on page is “Om”, or “ Aum”.

Editorial Advisory Board Steve Andriole, Villanova University, USA Athula Ginige, University of Western Sydney, Australia Dion Hinchcliffe, Hinchcliffe & Company, USA Narayanan Kulathuramaiyer, Universiti Malaysia Sarawak, Malaysia In Lee, Western Illinois University, USA Graeme Philipson, Connection Research Services, Australia Prasad Ram, Google India, Bangalore, India Bhuvan Unhelkar, Method Science, Australia Ed Yourdon, Software Consultant, USA Martin Gaedke, Chemnitz University of Technology, Germany

List of Reviewers Aaron Bond Alan Rea Andrea Pesoli Andreas Schmit Annet Mills Antonina Dattolo Bertalan Masko Bhuvan Unhelkar Carlo Tasso Chak Chaka Christian Futches Christoph Lattemann Christoph Schroth Dan J. Kim Daniel Memmi Darren Nicholson David Grifin David Kerr David Wyld Donald Kerr

Epaminondas Kapetanios Fabio Vitali Frédéric Adam Gustavo Rossi Hansel Burley Hesoke Lee Indrit Troshani James Richard Jan vom Brocke John M Artz Jongho Kim Jorg Rech Jorge Marx Katsumi Tanaka Kristina Lerman Lance Fung Luis Olsina Marco Brambilla Mariano Corso Oscar Pastor Pankaj Kamthan Phillip O Reiley Phillip Olla Piero Fraternali Rafael A. Calvo Richard Hartshorne Robert Sassano Sara Comi Sebestian Wheber Shakib Manouchehri Sotiris Christodoulou Stefano Picascia Steve Weller Steven A. Demurjian Steven Burgess Susanne Draheim T. Andrew Yang Terry Daugherty Udo Winand Valentin Zacharias Wail M. Omar Werner Beuschel Woojong Suh Yasmin Ibrahim Young Yu

List of Contributors

Agresta, T. / University of Connecticut Health Center, USA ............................................................. 682 Ajjan, Haya / University of North Carolina at Charlotte, USA ........................................................ 593 Balkesen, Çağrı / ETH Zurich, Switzerland ...................................................................................... 720 Bao, Shenghua / Shanghai Jiao Tong University, China ................................................................... 260 Becerra-Ortiz, I. / Fair Haven Community Health Center, USA ...................................................... 682 Berhe, S. / University of Connecticut, USA........................................................................................ 430 Boder, Gautier / ETH Zurich, Switzerland ........................................................................................ 720 Bozzon, Alessandro / Politecnico di Milano, Italy .............................................................................. 75 Brambilla, Marco / Politecnico di Milano, Italy................................................................................. 96 Braun, Simone / FZI Research Center for Information Technology, Germany................................. 225 Burley, Hansel / Texas Tech University, USA .................................................................................... 613 Calvo, Rafael A. / The University of Sydney, Australia ..................................................................... 817 Carter, S. / Community Health Centers, Inc., USA ............................................................................ 682 Carughi, Giovanni Toffetti / Università della Svizzera Italiana, Switzerland ................................... 75 Casoto, Paolo / University of Udine, Italy ......................................................................................... 312 Chaka, Chaka / Walter Sisulu University, South Africa .................................................................... 630 Chang, Maiga / Athabasca University, Canada ................................................................................ 613 Christodoulou, Sotiris P. / University of Patras, Greece .................................................................. 192 Comai, Sara / Politecnico di Milano, Italy .......................................................................................... 75 Cook, M. J. / University of Connecticut Health Center, USA ............................................................ 682 Cress, Ulrike / Knowledge Media Research Center, Germany .......................................................... 573 Crowell, R. / University of Connecticut Health Center, USA ............................................................ 682 Dattolo, Antonina / University of Udine, Italy .......................................................................... 312, 349 Demurjian, S. / University of Connecticut, USA ....................................................................... 430, 682 Derham, Richard / University of Canterbury, New Zealand ............................................................ 206 Devineni, M. / Serebrum Cooperation, USA...................................................................................... 430 Dhalwani, Vishal / University of Houston-Clear Lake, USA ............................................................ 647 Di Iorio, Angelo / University of Bologna, Italy.................................................................................. 329 Dindar, Nihal / ETH Zurich, Switzerland .......................................................................................... 720 Duca, Silvia / University of Bologna, Italy ........................................................................................ 349 Fei, Ben / IBM China Research Lab, China ....................................................................................... 260 Ferdig, Richard E. / Kent State University, USA .............................................................................. 593 Fiield, J. / University of Connecticut Health Center, USA ................................................................ 682

Fraternali, Piero / Politecnico di Milano, Italy............................................................................. 75, 96 Fuchs, Christian / University of Salzburg, Austria ........................................................................... 764 Ginzburg, Jeronimo / FCEyN, UBA, Argentina ................................................................................. 59 Grifin, David / Leeds Metropolitan University, UK ......................................................................... 496 Hartshorne, Richard / University of North Carolina at Charlotte, USA.......................................... 593 Hong, Jinwon / Inha University, South Korea ................................................................................... 387 Ibrahim, Yasmin / University of Brighton, UK ................................................................................. 828 Jatowt, Adam / Kyoto University, Japan ........................................................................................... 242 Kamthan, Pankaj / Concordia University, Canada .................................................................. 472, 733 Kapetanios, Epaminondas / University of Westminster, UK ............................................................ 277 Keusch, Florian / ETH Zurich, Switzerland ...................................................................................... 720 Kim, Dan J. / University of Houston-Clear Lake, USA ..................................................... 647, 662, 804 Kim, Jongho / Hyundai Research Institute, South Korea .................................................................. 387 Kimmerle, Joachim / University of Tuebingen, Germany ................................................................ 573 Koutsomitropoulos, Dimitrios A. / University of Patras, Greece .................................................... 192 Kromwijk, Katinka / ETH Zurich, Switzerland................................................................................ 720 Lattemann, Christoph / University of Potsdam, Germany............................................................... 699 Lee, Heeseok / Korea Advanced Institute of Science and Technology, South Korea ......................... 387 Lerman, Kristina / USC Information Sciences Institute, USA.......................................................... 296 Li, Rui / Shanghai Jiao Tong University, China ................................................................................ 260 Linnenfelser, Marcel / Synlag Web Engineering, Germany ............................................................. 135 Mahaley, Steve / Duke Corporate Education, USA ........................................................................... 556 Manouchehri, Shakib / University of Kassel, Germany ................................................................... 673 Marmo, Samuele / LABSS-ISTC-CNR, Italy ..................................................................................... 411 Memmi, Daniel / University of Quebec in Montreal, Canada .......................................................... 790 Mendes, Emilia / The University of Auckland, New Zealand ............................................................ 449 Mich, Luisa / University of Trento, Italy............................................................................................ 371 Mills, Annette / University of Canterbury, New Zealand .................................................................. 206 Molteni, Emanuele / Web Models S.r.l., Italy ...................................................................................... 96 Moskaliuk, Johannes / University of Tuebingen, Germany .............................................................. 573 Murugesan, San / Multimedia University, Malaysia & University of Western Sydney, Australia ........ 1 Naik, Ninad / University of Houston-Clear Lake, USA ..................................................................... 804 Nakamura, Satoshi / Kyoto University, Japan .................................................................................. 242 Nikolakopoulos, Ioannis G. / National Technical University of Athens, Greece.............................. 863 O’Rourke, Stephen T. / The University of Sydney, Australia ........................................................... 817 Olaniran, Bolanle A. / Texas Tech University, USA .......................................................................... 613 Olla, Phillip / Madonna University, USA........................................................................................... 522 Olsina, Luis / National University of La Pampa, Argentina.............................................................. 371 Omar, Wail M. / Sohar University, Sultanate of Oman ..................................................................... 119 Omero, Paolo / University of Udine, Italy ......................................................................................... 312 Pang, Minseok / University of Michigan at Ann Arbor, USA ............................................................ 387 Paolucci, Mario / LABSS-ISTC-CNR, Italy ....................................................................................... 411 Papatheodorou, Theodore S. / University of Patras, Greece ........................................................... 192 Pastor, Oscar / Universidad Politécnica de Valencia, Spain ............................................................... 40

Patrikakis, Charalampos Z. / National Technical University of Athens, Greece ............................ 863 Pelechano, Vicente / Universidad Politécnica de Valencia, Spain ...................................................... 40 Picascia, Stefano / LABSS-ISTC-CNR, Italy...................................................................................... 411 Plangprasopchok, Anon / USC Information Sciences Institute, USA............................................... 296 Polineni, K. / Serebrum Cooperation, USA ............................................................................... 430, 682 Pomonis, Tzanetos / University of Patras, Greece ............................................................................ 192 Pudota, Nirmala / University of Udine, Italy .................................................................................... 312 Qureshi, Elena / Madonna University, USA ...................................................................................... 522 Rea, Alan / Western Michigan University, USA ................................................................................. 159 Rech, Jörg / Fraunhofer Institute for Experimental Software Engineering (IESE), Germany.... 12, 135 Ren, H. / University of Connecticut, USA .......................................................................................... 430 Richards, James / Heriot-Watt University, UK ................................................................................. 846 Rossi, Gustavo / UNLP and Conicet, Argentina .................................................................................. 59 Sassano, Roberto / University of Trento, Italy .................................................................................. 371 Schmidt, Andreas / FZI Research Center for Information Technology, Germany ............................ 225 Şengül, Ali / ETH Zurich, Switzerland ............................................................................................... 720 Sonnenberg, Christian / University of Liechtenstein, Principality of Liechtenstein ........................ 699 Stieglitz, Stefan / University of Potsdam, Germany .......................................................................... 699 Su, Zhong / IBM China Research Lab, China ................................................................................... 260 Suh, Woojong / Inha University, South Korea ................................................................................... 387 Tanaka, Katsumi / Kyoto University, Japan ..................................................................................... 242 Tanase, Diana Irina / University of Westminster, UK ....................................................................... 277 Tasso, Carlo / University of Udine, Italy ........................................................................................... 312 Tatbul, Nesime / ETH Zurich, Switzerland ........................................................................................ 720 Teigland, Robin / Stockholm School of Economics, Sweden ............................................................. 556 Tomasi, Francesca / University of Bologna, Italy ............................................................................. 349 Tracey, L. / StayWell Health Care, Inc., USA .................................................................................... 682 Trivedi, Bharti / DD University, India .............................................................................................. 748 Unhelkar, Bhuvan / University of Western Sydney & MethodScience.com, Australia ............. 178, 748 Urbieta, Matias / UNLP and Conicet, Argentina ................................................................................ 59 Valderas, Pedro / Universidad Politécnica de Valencia, Spain ........................................................... 40 Valverde, Francisco / Universidad Politécnica de Valencia, Spain .................................................... 40 Vegad, S. / Serebrum Corporation, USA ............................................................................................ 682 Vegad, Sushil / Serebrum Cooperation, USA..................................................................................... 430 Vitali, Fabio / University of Bologna, Italy................................................................................ 329, 349 vom Brocke, Jan / University of Liechtenstein, Principality of Liechtenstein .................................. 699 Voulodimos, Athanasios S. / National Technical University of Athens, Greece ............................... 863 Vu, Tri / University of Houston-Clear Lake, USA.............................................................................. 647 Weber, Sebastian / Fraunhofer Institute for Experimental Software Engineering (IESE), Germany ................................................................................................. 12, 135 Weinberger, Hadas / HIT – Holon Institute of Technology, Israel .................................................... 539 Wheeler, Steve / University of Plymouth, UK.................................................................................... 511 Winand, Udo / University of Kassel, Germany ................................................................................. 673 Wu, Ming-Chien (Mindy) / University of Western Sydney, Australia .............................................. 178

Yanbe, Yusuke / Kyoto University, Japan.......................................................................................... 242 Yang, T. Andrew / University of Houston-Clear Lake, USA ............................................. 647, 662, 804 Yu, Yong / Shanghai Jiao Tong University, China ............................................................................. 260 Zacchiroli, Stefano / Universitè Paris Diderot, France .................................................................... 329 Zacharias, Valentin / FZI Research Center for Information Technology, Germany ......................... 225

Table of Contents

Preface .................................................................................................................................................. xl Acknowledgment ............................................................................................................................... xliv

Volume I Section 1 Overview Chapter 1 Web X.0: A Road Map ............................................................................................................................ 1 San Murugesan, Multimedia University, Malaysia & University of Western Sydney, Australia Chapter 2 An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement: The Web Before and Beyond 2.0 .......................................................................................................... 12 Sebastian Weber, Fraunhofer Institute for Experimental Software Engineering (IESE), Germany Jörg Rech, Fraunhofer Institute for Experimental Software Engineering (IESE), Germany

Section 2 Web Modeling and Design Chapter 3 A Model-Driven Engineering Approach for Deining Rich Internet Applications: A Web 2.0 Case Study........................................................................................................................... 40 Francisco Valverde, Universidad Politécnica de Valencia, Spain Oscar Pastor, Universidad Politécnica de Valencia, Spain Pedro Valderas, Universidad Politécnica de Valencia, Spain Vicente Pelechano, Universidad Politécnica de Valencia, Spain

Chapter 4 Modular and Systematic Interface Design for Rich Internet Applications ........................................... 59 Gustavo Rossi, UNLP and Conicet, Argentina Matias Urbieta, UNLP and Conicet, Argentina Jeronimo Ginzburg, FCEyN, UBA, Argentina Chapter 5 Towards Web 2.0 Applications: A Conceptual Model for Rich Internet Applications ......................... 75 Alessandro Bozzon, Politecnico di Milano, Italy Sara Comai, Politecnico di Milano, Italy Piero Fraternali, Politecnico di Milano, Italy Giovanni Toffetti Carughi, Università della Svizzera Italiana, Switzerland Chapter 6 A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX ............................... 96 Marco Brambilla, Politecnico di Milano, Italy Piero Fraternali, Politecnico di Milano, Italy Emanuele Molteni, Web Models S.r.l., Italy Chapter 7 Web 2.0: Self-Managing System Based on SOA Model and Grid Computing Overlay .................... 119 Wail M. Omar, Sohar University, Sultanate of Oman

Section 3 Web Architecture Chapter 8 An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures ............... 135 Marcel Linnenfelser, Synlag Web Engineering, Germany Sebastian Weber, Fraunhofer Institute for Experimental Software Engineering (IESE), Germany Jörg Rech, Fraunhofer Institute for Experimental Software Engineering (IESE), Germany Chapter 9 The Layered Virtual Reality Commerce System (LaVRCS): Proposing an Immersive Web X.0 Framework for E-Commerce ............................................................................................... 159 Alan Rea, Western Michigan University, USA Chapter 10 Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era .......................... 178 Ming-Chien (Mindy) Wu, University of Western Sydney, Australia Bhuvan Unhelkar, University of Western Sydney & MethodScience.com, Australia

Chapter 11 Towards Web 3.0: A Unifying Architecture for Next Generation Web Applications ......................... 192 Tzanetos Pomonis, University of Patras, Greece Dimitrios A. Koutsomitropoulos, University of Patras, Greece Sotiris P. Christodoulou, University of Patras, Greece Theodore S. Papatheodorou, University of Patras, Greece

Section 4 Information Search, Bookmarking, and Tagging Chapter 12 Web 2.0—Social Bookmarking: An Overview of Folksonomies ....................................................... 206 Richard Derham, University of Canterbury, New Zealand Annette Mills, University of Canterbury, New Zealand Chapter 13 Social Semantic Bookmarking with SOBOLEO ................................................................................ 225 Valentin Zacharias, FZI Research Center for Information Technology, Germany Simone Braun, FZI Research Center for Information Technology, Germany Andreas Schmidt, FZI Research Center for Information Technology, Germany Chapter 14 Social Bookmarking and Web Search ................................................................................................. 242 Yusuke Yanbe, Kyoto University, Japan Adam Jatowt, Kyoto University, Japan Satoshi Nakamura, Kyoto University, Japan Katsumi Tanaka, Kyoto University, Japan Chapter 15 Social Tagging: Properties and Applications ...................................................................................... 260 Yong Yu, Shanghai Jiao Tong University, China Rui Li, Shanghai Jiao Tong University, China Shenghua Bao, Shanghai Jiao Tong University, China Ben Fei, IBM China Research Lab, China Zhong Su, IBM China Research Lab, China Chapter 16 Improving Cross-Language Information Retrieval by Harnessing the Social Web............................ 277 Diana Irina Tanase, University of Westminster, UK Epaminondas Kapetanios, University of Westminster, UK

Chapter 17 Leveraging User-Speciied Metadata to Personalize Image Search ................................................... 296 Kristina Lerman, USC Information Sciences Institute, USA Anon Plangprasopchok, USC Information Sciences Institute, USA

Section 5 Semantic Analysis and Semantic Web Chapter 18 Accessing, Analyzing, and Extracting Information from User Generated Contents .......................... 312 Paolo Casoto, University of Udine, Italy Antonina Dattolo, University of Udine, Italy Paolo Omero, University of Udine, Italy Nirmala Pudota, University of Udine, Italy Carlo Tasso, University of Udine, Italy Chapter 19 Wiki Semantics via Wiki Templating.................................................................................................. 329 Angelo Di Iorio, University of Bologna, Italy Fabio Vitali, University of Bologna, Italy Stefano Zacchiroli, Universitè Paris Diderot, France Chapter 20 Towards Disambiguating Social Tagging Systems ............................................................................. 349 Antonina Dattolo, University of Udine, Italy Silvia Duca, University of Bologna, Italy Francesca Tomasi, University of Bologna, Italy Fabio Vitali, University of Bologna, Italy

Section 6 Web Quality, Trust, Security, and Effort Estimation Chapter 21 Modeling Content Quality for the Web 2.0 and Follow-on Applications ........................................... 371 Roberto Sassano, University of Trento, Italy Luis Olsina, National University of La Pampa, Argentina Luisa Mich, University of Trento, Italy

Chapter 22 A New Web Site Quality Assessment Model for the Web 2.0 Era...................................................... 387 Minseok Pang, University of Michigan at Ann Arbor, USA Woojong Suh, Inha University, South Korea Jinwon Hong, Inha University, South Korea Jongho Kim, Hyundai Research Institute, South Korea Heeseok Lee, Korea Advanced Institute of Science and Technology, South Korea Chapter 23 Electronic Reputation Systems ........................................................................................................... 411 Mario Paolucci, LABSS-ISTC-CNR, Italy Stefano Picascia, LABSS-ISTC-CNR, Italy Samuele Marmo, LABSS-ISTC-CNR, Italy Chapter 24 Improving the Information Security of Collaborative Web Portals via Fine-Grained Role-Based Access Control................................................................................................................. 430 S. Demurjian, University of Connecticut, USA H. Ren, University of Connecticut, USA S. Berhe, University of Connecticut, USA M. Devineni, Serebrum Cooperation, USA Sushil Vegad, Serebrum Cooperation, USA K. Polineni, Serebrum Cooperation, USA Chapter 25 Web 2.0 Effort Estimation .................................................................................................................. 449 Emilia Mendes, The University of Auckland, New Zealand

Volume II Section 7 Educational Applications Chapter 26 A Social Web Perspective of Software Engineering Education .......................................................... 472 Pankaj Kamthan, Concordia University, Canada Chapter 27 University 2.0: Embracing Social Networking to Better Engage the Facebook-Generation in University Life ................................................................................................................................ 496 David Grifin, Leeds Metropolitan University, UK

Chapter 28 On Using Wiki as a Tool for Collaborative Online Blended Learning ............................................... 511 Steve Wheeler, University of Plymouth, UK Chapter 29 Integration of Web 2.0 Collaboration Tools into Education: Lessons Learned .................................. 522 Phillip Olla, Madonna University, USA Elena Qureshi, Madonna University, USA Chapter 30 ECHO: A Layered Model for the Design of a Context-Aware Learning Experience ......................... 539 Hadas Weinberger, HIT – Holon Institute of Technology, Israel Chapter 31 Advancing Learning Through Virtual Worlds .................................................................................... 556 Steve Mahaley, Duke Corporate Education, USA Robin Teigland, Stockholm School of Economics, Sweden Chapter 32 Virtual Reality 2.0 and Its Application in Knowledge Building ......................................................... 573 Johannes Moskaliuk, University of Tuebingen, Germany Joachim Kimmerle, University of Tuebingen, Germany Ulrike Cress, Knowledge Media Research Center, Germany Chapter 33 Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education ................. 593 Haya Ajjan, University of North Carolina at Charlotte, USA Richard Hartshorne, University of North Carolina at Charlotte, USA Richard E. Ferdig, Kent State University, USA Chapter 34 Social Issues and Web 2.0: A Closer Look at Culture in E-Learning ................................................. 613 Bolanle A. Olaniran, Texas Tech University, USA Hansel Burley, Texas Tech University, USA Maiga Chang, Athabasca University, Canada

Section 8 Enterprise 2.0, Healthcare, Finance, and Other Applications Chapter 35 Enterprise 2.0: Leveraging Prosumerism 2.0 Using Web 2.0 and Web 3.0 ........................................ 630 Chaka Chaka, Walter Sisulu University, South Africa

Chapter 36 Capturing Online Collaboration in the Design Elements Model for Web 2.0 and Beyond ................ 647 T. Andrew Yang, University of Houston-Clear Lake, USA Dan J. Kim, University of Houston-Clear Lake, USA Tri Vu, University of Houston-Clear Lake, USA Vishal Dhalwani, University of Houston-Clear Lake, USA Chapter 37 A Comparative Analysis of Online Social Networking Sites and Their Business Models ................. 662 T. Andrew Yang, University of Houston-Clear Lake, USA Dan J. Kim, University of Houston-Clear Lake, USA Chapter 38 Healthcare 2.0: The Use of Web 2.0 in Healthcare ............................................................................. 673 Shakib Manouchehri, University of Kassel, Germany Udo Winand, University of Kassel, Germany Chapter 39 Using a Web-Based Collaboration Portal and Wiki for Making Health Information Technology Decisions ......................................................................................................................... 682 R. Crowell, University of Connecticut Health Center, USA T. Agresta, University of Connecticut Health Center, USA M. J. Cook, University of Connecticut Health Center, USA J. Fiield, University of Connecticut Health Center, USA S. Demurjian, University of Connecticut, USA S. Carter, Community Health Centers, Inc., USA I. Becerra-Ortiz, Fair Haven Community Health Center, USA L. Tracey, StayWell Health Care, Inc., USA S. Vegad, Serebrum Corporation, USA K. Polineni, Serebrum Corporation, USA Chapter 40 Assessing the Total Cost of Ownership of Virtual Communities: The Case of the Berlin Stock Exchange................................................................................................................................... 699 Jan vom Brocke, University of Liechtenstein, Principality of Liechtenstein Christian Sonnenberg, University of Liechtenstein, Principality of Liechtenstein Christoph Lattemann, University of Potsdam, Germany Stefan Stieglitz, University of Potsdam, Germany

Chapter 41 Connecting the Real World with the Virtual World: The SmartRFLib RFID-Supported Library System on Second Life .......................................................................................................... 720 Katinka Kromwijk, ETH Zurich, Switzerland Çağrı Balkesen, ETH Zurich, Switzerland Gautier Boder, ETH Zurich, Switzerland Nihal Dindar, ETH Zurich, Switzerland Florian Keusch, ETH Zurich, Switzerland Ali Şengül, ETH Zurich, Switzerland Nesime Tatbul, ETH Zurich, Switzerland Chapter 42 Embracing the Social Web for Managing Patterns ............................................................................. 733 Pankaj Kamthan, Concordia University, Canada Chapter 43 Extending and Applying Web 2.0 and Beyond for Environmental Intelligence ................................. 748 Bhuvan Unhelkar, University of Western Sydney & MethodScience.com, Australia Bharti Trivedi, DD University, India

Section 9 Social Web: Foundations, Analysis, and Visualisation Chapter 44 Social Software and Web 2.0: Their Sociological Foundations and Implications .............................. 764 Christian Fuchs, University of Salzburg, Austria Chapter 45 Sociology of Virtual Communities and Social Software Design ........................................................ 790 Daniel Memmi, University of Quebec in Montreal, Canada Chapter 46 Online Human Activity Networks (OnHANs): An Analysis Based on Activity Theory .................... 804 Dan J. Kim, University of Houston-Clear Lake, USA T. Andrew Yang, University of Houston-Clear Lake, USA Ninad Naik, University of Houston-Clear Lake, USA Chapter 47 Visualising Social Networks in Collaborative Environments ............................................................. 817 Stephen T. O’Rourke, The University of Sydney, Australia Rafael A. Calvo, The University of Sydney, Australia

Chapter 48 The Discourses of Empowerment and Web 2.0: The Dilemmas of User-Generated Content ............ 828 Yasmin Ibrahim, University of Brighton, UK Chapter 49 How Employees Can Leverage Web 2.0 in New Ways to Relect on Employment and Employers .................................................................................................................................... 846 James Richards, Heriot-Watt University, UK Chapter 50 Privacy Implications and Protection in the New Ubiquitous Web Environment ................................ 863 Charalampos Z. Patrikakis, National Technical University of Athens, Greece Ioannis G. Nikolakopoulos, National Technical University of Athens, Greece Athanasios S. Voulodimos, National Technical University of Athens, Greece

Epilogue ............................................................................................................................................. 878 Compilation of References ............................................................................................................... 880

Detailed Table of Contents

Preface .................................................................................................................................................. xl Acknowledgment ............................................................................................................................... xliv

Volume I Section 1 Overview Chapter 1 Web X.0: A Road Map ............................................................................................................................ 1 San Murugesan, Multimedia University, Malaysia & University of Western Sydney, Australia The Web has evolved from its humble beginnings merely as a publishing medium intended for a small group of scientists to a medium of interaction, participation, and collaboration. It has dramatically inluenced almost every sphere of our activity and has created paradigm shifts. Encompassing new technologies, business strategies, and social trends, the Web continues to forge many new applications that we had never imagined before or were not previously feasible. It has created new paradigms in business, social interaction, governance, and education. In this chapter, we trace the Web’s continuing evolution and phenomenal strides, outline the features and characteristics of Web 2.0, 3.0, and X.0, and examine their prospects and potential. The ability to recognize new Web technologies for their potential in business, social and educational applications, and the ability to develop and deploy creative applications based on these technologies are the keys to continued success of the Web and our progress and well being. Chapter 2 An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement: The Web Before and Beyond 2.0 .......................................................................................................... 12 Sebastian Weber, Fraunhofer Institute for Experimental Software Engineering (IESE), Germany Jörg Rech, Fraunhofer Institute for Experimental Software Engineering (IESE), Germany Web 2.0 is a popular term used to describe a class of Web applications that offers mostly free services to its users. However, an exact deinition of the concepts, features, and technologies that argue for a Web 2.0 service is still missing. Similarly, terms such as Web 3.0, Web 4.0, or Web 2.5 also have no clear and unambiguous deinitions. This chapter reports the results of a Web and literature survey about Web

X.Y concepts. Based on several deintions, we synthesized new deinitions for Web X.Y, which provide an overview and can be used for differentia-tion, and we classiied contemporary Web services (e.g., Flickr) according to these deinitions. Section 2 Web Modeling and Design Chapter 3 A Model-Driven Engineering Approach for Deining Rich Internet Applications: A Web 2.0 Case Study........................................................................................................................... 40 Francisco Valverde, Universidad Politécnica de Valencia, Spain Oscar Pastor, Universidad Politécnica de Valencia, Spain Pedro Valderas, Universidad Politécnica de Valencia, Spain Vicente Pelechano, Universidad Politécnica de Valencia, Spain Web 2.0 applications emphasize the end-user involvement to provide the content. In this new scenario, an easy to use and a highly interactive user interface (UI) is a key requirement in order to appeal the end-user. The main objective of this chapter is to introduce a model-driven engineering process to create rich Internet applications (RIA) that address the requirements that a Web 2.0 application must fulill. To achieve this goal, an interaction model made up of two complementary models is proposed: On the one hand, an abstract interaction model, which clearly deines the interactions between the user and the system and on the other hand, a concrete RIA interaction model that speciies the semantics needed to accurately deine RIA for the Web 2.0 domain. Both models are introduced inside a model-driven code generation process with the aim of producing a fully functional Web 2.0 application. To illustrate the contribution of this chapter, the approach is applied in a case study related to the Web 2.0 domain. Chapter 4 Modular and Systematic Interface Design for Rich Internet Applications ........................................... 59 Gustavo Rossi, UNLP and Conicet, Argentina Matias Urbieta, UNLP and Conicet, Argentina Jeronimo Ginzburg, FCEyN, UBA, Argentina In this chapter, we present a design approach for the interface of rich Internet applications, that is, those Web applications in which the conventional hypermedia paradigm has been improved with rich interaction styles. Our approach combines well-known techniques for advanced separation of concerns such as aspect-oriented software design, with the object oriented hypermedia design method (OOHDM) design model allowing to express in a high level way the structure and behaviours of the user interface as oblivious compositions of simpler interface atoms. Using simple illustrative examples we present the rationale of our approach, its core stages and the way it is integrated into the OOHDM. Some implementation issues are inally analyzed.

Chapter 5 Towards Web 2.0 Applications: A Conceptual Model for Rich Internet Applications ......................... 75 Alessandro Bozzon, Politecnico di Milano, Italy Sara Comai, Politecnico di Milano, Italy Piero Fraternali, Politecnico di Milano, Italy Giovanni Toffetti Carughi, Università della Svizzera Italiana, Switzerland This chapter introduces a conceptual model for the design of Web 2.0 applications relying on rich Internet application (RIA) technologies. RIAs extend Web application features by allowing computation to be partitioned between the client and the server and support core Web 2.0 requirements, like real-time collaboration among users, sophisticated presentation and manipulation of multimedia content, and lexible human-machine interaction (synchronous and asynchronous, connected and disconnected). The proposed approach for the design of Web 2.0 applications extends a conceptual platform-independent model conceived for Web 1.0 applications with novel primitives capturing RIA features; the conceptual model can be automatically converted into implementations in all the most popular RIA technologies and frameworks like AJAX, OpenLaszlo, FLEX, AIR, Google Gears, Google Web toolkit, and Silverlight. Chapter 6 A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX ............................... 96 Marco Brambilla, Politecnico di Milano, Italy Piero Fraternali, Politecnico di Milano, Italy Emanuele Molteni, Web Models S.r.l., Italy This chapter describes how the design tool WebRatio (and its companion conceptual model WebML) have been extended to support the new requirements imposed by rich Internet applications (RIAs), that are recognized to be one of the main innovations that lead to the Web 2.0 revolution. Complex interactions such as drag and drop, dynamic resizing of visual components, graphical editing of objects, and partial page refresh are addressed by the RIA extensions of WebRatio. The chapter discusses what kinds of modelling primitives are required for specifying such patterns and how these primitives can be integrated in a CASE tool. Finally, a real industrial case is presented in which the novel RIA features are successfully applied. Chapter 7 Web 2.0: Self-Managing System Based on SOA Model and Grid Computing Overlay .................... 119 Wail M. Omar, Sohar University, Sultanate of Oman Web 2.0 is expected to be the next technology in the interaction between the enterprise applications and end users. Such interaction will be utilized in producing self-governance applications that are able to re-adjacent and reconigure the operation framework based on users’ feedback. To achieve this, huge numbers of underneath resources (infrastructures and services) are required. Therefore, this work proposes the merge of Web 2.0 technology and grid computing overlay to support Web 2.0 framework. Such merge between technologies is expected to offer mutual beneits for both communities. Through this work, a model for managing the interaction between the two technologies is developed based on the adapting of service oriented architecture (SOA) model, this model is known as SOAW2G. This model

manages the interaction between the users at the top level and resources at the bottom layer. As a case study, managing health information based on users’ (doctors, medicine companies, and others) experiences is explored through this chapter.

Section 3 Web Architecture Chapter 8 An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures ............... 135 Marcel Linnenfelser, Synlag Web Engineering, Germany Sebastian Weber, Fraunhofer Institute for Experimental Software Engineering (IESE), Germany Jörg Rech, Fraunhofer Institute for Experimental Software Engineering (IESE), Germany An important aspect of Web 2.0 mentioned by Tim O’Reilly is the rich user experience. Web 2.0 applications offer the user a desktop-like interface to bring back eficiency and productivity. The click-wait-andrefresh-cycle of normal Web applications leads to a less responsive, and thus less eficient, user interface. To serve the needs of these so-called rich Internet applications (RIA), many different approaches have emerged, based either on Web standards or on proprietary approaches. This chapter aims at deining a qualiied criterion system for comparing RIA platforms. Thereafter, those RIA platforms are selected and analyzed in terms of the criterion system that is most likely to become widely accepted. Chapter 9 The Layered Virtual Reality Commerce System (LaVRCS): Proposing an Immersive Web X.0 Framework for E-Commerce ............................................................................................... 159 Alan Rea, Western Michigan University, USA In this chapter, the author argues that virtual reality does have a place in e-commerce as a Web 2.0 application. However, VR is not ready to supplant standard e-commerce Web interfaces with a completely immersive VR environment. Rather, VRCommerce must rely on a mixed platform presentation to accommodate diverse levels of usability, technical feasibility, and user trust. The author proposes that ecommerce sites that want to implement VRCommerce offer at least three layers of interaction: a standard Web interface, embedded VR objects in a Web interface, and semi-immersive VR within an existing Web interface. This system is termed the layered virtual reality commerce system, or LaVRCS. This proposed LaVRCS framework can work in conjunction with rich Internet applications, Webtops, and other Web 2.0 applications to offer another avenue of interaction within the e-commerce realm. With adoption and development, LaVRCS will help propel e-commerce into the Web 3.0 realm and beyond. Chapter 10 Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era .......................... 178 Ming-Chien (Mindy) Wu, University of Western Sydney, Australia Bhuvan Unhelkar, University of Western Sydney & MethodScience.com, Australia

This chapter describes an approach to extending service oriented architecture (SOA) with mobile technologies (MT) resulting in what can be called mobile service oriented architecture (MSOA). Web aervices (WS) is a popular approach to business applications in the second Web generation (Web 2.0). Mobile technologies (MT) help people reach out and interact with each other anytime and anywhere, transcending time and location boundaries. MSOA brings together MT and WS to create opportunities for offering and consuming services over the wireless networks in Web 2.0 era-and beyond. Furthermore, the intelligent convergence of mobile connectivity, network computing, open technology, open identity, and several such emerging technologies pave the way for newer and wider range of service-oriented business opportunities. The authors describe this MSOA model and an approach to its validation through an implementation framework in this chapter. Chapter 11 Towards Web 3.0: A Unifying Architecture for Next Generation Web Applications ......................... 192 Tzanetos Pomonis, University of Patras, Greece Dimitrios A. Koutsomitropoulos, University of Patras, Greece Sotiris P. Christodoulou, University of Patras, Greece Theodore S. Papatheodorou, University of Patras, Greece While the term Web 2.0 is used to describe the current trend in the use of Web technologies, the term Web 3.0 is used to describe the next generation Web, which will combine Semantic Web technologies, Web 2.0 principles, and artiicial intelligence. Towards this perspective, in this work we introduce a 3-tier architecture for Web applications that will it into the Web 3.0 deinition. We present the fundamental features of this architecture, its components, and their interaction, as well as the current technological limitations. Furthermore, some indicative application scenarios are outlined in order to illustrate the features of the proposed architecture. The aim of this architecture is to be a step towards supporting the development of intelligent Semantic Web applications of the near future as well as supporting the user collaboration and community-driven evolution of these applications.

Section 4 Information Search, Bookmarking, and Tagging Chapter 12 Web 2.0—Social Bookmarking: An Overview of Folksonomies ....................................................... 206 Richard Derham, University of Canterbury, New Zealand Annette Mills, University of Canterbury, New Zealand Folksonomies is a relatively new concept and, as yet, it has not been widely studied in academic circles. In practice, folksonomies have therefore outpaced academic research in inding solutions to the problems facing them. The goal of this chapter is to bring together the current literature on folksonomies and explore avenues for future work. Hence, this chapter will examine what are folksonomies, what they are/ can be used for, and explore their beneits and challenges using real world examples from systems such as Delicious and Flickr. The chapter also overviews some of the current research and suggests avenues for further work.

Chapter 13 Social Semantic Bookmarking with SOBOLEO ................................................................................ 225 Valentin Zacharias, FZI Research Center for Information Technology, Germany Simone Braun, FZI Research Center for Information Technology, Germany Andreas Schmidt, FZI Research Center for Information Technology, Germany The novel paradigm of social semantic bookmarking combines the positive aspects of semantic annotation with those of social bookmarking and tagging while avoiding their respective drawbacks; drawbacks such as the lacking semantic precision of tags or the cumbersome maintenance of ontologies. Social semantic bookmarking tools allow for the annotation of Internet resources based on an ontology and the integrated maintenance of the ontology by the same people that use it. This chapter motivates social semantic bookmarking by examining the respective problems of tag based bookmarking and semantic annotation. Social semantic bookmarking is then introduced and explained using the SOBOLEO application as an example. It also gives an overview of existing applications implementing this new paradigm and makes predictions about its movement into the mainstream and remaining research challenges. Chapter 14 Social Bookmarking and Web Search ................................................................................................. 242 Yusuke Yanbe, Kyoto University, Japan Adam Jatowt, Kyoto University, Japan Satoshi Nakamura, Kyoto University, Japan Katsumi Tanaka, Kyoto University, Japan Social bookmarking is an emerging type of a Web service for reusing, sharing, and discovering resources. By bookmarking users preserve access points to encountered documents for their future access. On the other hand, the social aspect of bookmarking results from the visibility of bookmarks to other users helping them to discover new, potentially interesting resources. In addition, social bookmarking systems allow for better estimation of the popularity and relevance of documents. In this chapter, we provide an overview of major aspects involved with social bookmarking and investigate their potential for enhancing Web search and for building novel applications. We make a comparative analysis of two popularity measures of Web pages, PageRank and SBRank, where SBRank is deined as an aggregate number of bookmarks that a given page accumulates in a selected social bookmarking system. The results of this analysis reveal the advantages of SBRank when compared to PageRank measure and provide the foundations for utilizing social bookmarking information in order to enhance and improve search in the Web. In the second part of the chapter, we describe an application that combines SBRank and PageRank measures in order to re-rank results delivered by Web search engines and that offers several complimentary functions for realizing more effective search. Chapter 15 Social Tagging: Properties and Applications ...................................................................................... 260 Yong Yu, Shanghai Jiao Tong University, China Rui Li, Shanghai Jiao Tong University, China Shenghua Bao, Shanghai Jiao Tong University, China Ben Fei, IBM China Research Lab, China Zhong Su, IBM China Research Lab, China

Recently, collaborative tagging Web sites such as Del.icio.us and Flickr have achieved great success. This chapter is concerned with the problem of social tagging analysis and mining. More speciically, we discuss ive properties of social tagging and their applications: 1) keyword property, which means social annotations serve as human selected keywords for Web resources; 2) semantic property, which indicates semantic relations among tags and Web resources; 3) hierarchical property, which means that hierarchical structure can be derived from the lat social tagging space; 4) quality property, which means that Web resources’ qualities are varied and can be quantiied using social tagging; 5) distribution property, which indicates the distribution of frequencies of social tags usually converges to a power-law distribution. These properties are the most principle characteristics, which have been popularly discussed and explored in many applications. As a case study, we show how to improve the social resource browsing by applying the ive properties of social tags. Chapter 16 Improving Cross-Language Information Retrieval by Harnessing the Social Web............................ 277 Diana Irina Tanase, University of Westminster, UK Epaminondas Kapetanios, University of Westminster, UK Combining existing advancements in cross-language information retrieval (CLIR), with the new usercentered Web paradigm could allow tapping into Web-based multilingual clusters of language information that are rich, up-to-date in terms of language usage, that increase in size, and have the potential to cater for all languages. In this chapter, we set out to explore existing CLIR systems and their limitations, and we argue that in the current context of a widely adopted social Web, the future of large-scale CLIR and iCLIR systems is linked to the use of the Web as a lexical resource, as a distribution infrastructure, and as a channel of communication between users. Such a synergy will lead to systems that grow organically as more users with different linguistic skills join the network, and that improve in terms of language translations disambiguation and coverage. Chapter 17 Leveraging User-Speciied Metadata to Personalize Image Search ................................................... 296 Kristina Lerman, USC Information Sciences Institute, USA Anon Plangprasopchok, USC Information Sciences Institute, USA The social media sites, such as Flickr and del.icio.us, allow users to upload content and annotate it with descriptive labels known as tags, join special-interest groups, and so forth. We believe user-generated metadata expresses user’s tastes and interests and can be used to personalize information to an individual user. Speciically, we describe a machine learning method that analyzes a corpus of tagged content to ind hidden topics. We then these learned topics to select content that matches user’s interests. We empirically validated this approach on the social photo-sharing site Flickr, which allows users to annotate images with freely chosen tags and to search for images labeled with a certain tag. We use metadata associated with images tagged with an ambiguous query term to identify topics corresponding to different senses of the term, and then personalize results of image search by displaying to the user only those images that are of interest to her.

Section 5 Semantic Analysis and Semantic Web Chapter 18 Accessing, Analyzing, and Extracting Information from User Generated Contents .......................... 312 Paolo Casoto, University of Udine, Italy Antonina Dattolo, University of Udine, Italy Paolo Omero, University of Udine, Italy Nirmala Pudota, University of Udine, Italy Carlo Tasso, University of Udine, Italy The concepts of the participative Web, mass collaboration, and collective intelligence grow out of a set of Web methodologies and technologies which improve interaction with users in the development, rating, and distribution of user-generated content. UGC is one of the cornerstones of Web 2.0 and is the core concept of several different kinds of applications. UGC suggests new value chains and business models; it proposes innovative social, cultural, and economic opportunities and impacts. However several open issues concerning semantic understanding and managing of digital information available on the Web, like information overload, heterogeneity of the available content, and effectiveness of retrieval are still unsolved. The research experiences we present in this chapter, described in literature or achieved in our research laboratory, are aimed at reducing the gap between users and information understanding, by means of collaborative and cognitive iltering, sentiment analysis, information extraction, and knowledge conceptual modeling. Chapter 19 Wiki Semantics via Wiki Templating.................................................................................................. 329 Angelo Di Iorio, University of Bologna, Italy Fabio Vitali, University of Bologna, Italy Stefano Zacchiroli, Universitè Paris Diderot, France A foreseeable incarnation of Web 3.0 could inherit machine understandability from the Semantic Web and collaborative editing from Web 2.0 applications. We review the research and development trends which are getting today Web nearer to such an incarnation. We present semantic wikis, microformats, and the so-called “lowercase semantic web”: they are the main approaches at closing the technological gap between content authors and Semantic Web technologies. We discuss a too often neglected aspect of the associated technologies, namely how much they adhere to the wiki philosophy of open editing: is there an intrinsic incompatibility between semantic rich content and unconstrained editing? We argue that the answer to this question can be “no,” provided that a few yet relevant shortcomings of current Web technologies will be ixed soon.

Chapter 20 Towards Disambiguating Social Tagging Systems ............................................................................. 349 Antonina Dattolo, University of Udine, Italy Silvia Duca, University of Bologna, Italy Francesca Tomasi, University of Bologna, Italy Fabio Vitali, University of Bologna, Italy Social tagging to annotate resources represents one of the innovative aspects introduced with Web 2.0 and the new challenges of the (semantic) Web 3.0. Social tagging, also known as user-generated keywords or folksonomies, implies that keywords, from an arbitrarily large and uncontrolled vocabulary, are used by a large community of readers to describe resources. Despite undeniable success and usefulness of social tagging systems, they also suffer from some drawbacks: the proliferation of social tags, coming as they are from an unrestricted vocabulary leads to ambiguity when determining their intended meaning; the lack of predeined schemas or structures for inserting metadata leads to confusions as to their roles and justiication; and the latness of the structure of the keywords and lack of relationships among them imply dificulties in relating different keywords when they describe the same or similar concepts. So in order to increase precision, in the searches and classiications made possible by folksonomies, some experiences and results from formal classiication and subjecting systems are considered, in order to help solve, if not to prevent altogether, the ambiguities that are intrinsic in such systems. Some successful and not so successful approaches as proposed in the scientiic literature are discussed, and a few more are introduced here to further help dealing with special cases. In particular, we believe that adding depth and structure to the terms used in folksonomies could help in word sense disambiguation, as well as correctly identifying and classifying proper names, metaphors and slang words when used as social tags.

Section 6 Web Quality, Trust, Security, and Effort Estimation Chapter 21 Modeling Content Quality for the Web 2.0 and Follow-on Applications ........................................... 371 Roberto Sassano, University of Trento, Italy Luis Olsina, National University of La Pampa, Argentina Luisa Mich, University of Trento, Italy The consistent modeling of quality requirements for Web sites and applications at different stages of the life cycle is still a challenge to most Web engineering researchers and practitioners. In the present chapter, we propose an integrated approach to specify quality requirements to Web sites and applications. By extending the ISO 9126-1 quality views’ characteristics, we discuss how to model internal, external quality and quality in use views taking into account not only the software features but also the own characteristics of Web applications. Particularly, we thoroughly analyze the modeling of the content characteristic for evaluating the quality of information–so critical for the whole Web application eras. The resulting model represents a irst step towards a multi-dimensional integrated approach to evaluate Web sites at different lifecycle stages.

Chapter 22 A New Web Site Quality Assessment Model for the Web 2.0 Era...................................................... 387 Minseok Pang, University of Michigan at Ann Arbor, USA Woojong Suh, Inha University, South Korea Jinwon Hong, Inha University, South Korea Jongho Kim, Hyundai Research Institute, South Korea Heeseok Lee, Korea Advanced Institute of Science and Technology, South Korea To ind a strategy for improving the competitiveness of Web sites, it is necessary to use comprehensive, integrated Web site quality dimensions that effectively discover which improvements are needed. Previous studies on Web site quality, however, seem to have inconsistent and confusing scopes, creating a need of reconciliation among the quality dimensions. Therefore, this chapter attempts to provide a Web site quality model that can comprise all the quality scopes provided by previous studies. The relationship between the speciic dimensions of the quality model and the characteristics or merits of Web 2.0 was discussed in this chapter with actual Web site examples. It is expected that this study can help Web sites improve their competitiveness in the Web 2.0 environment. Chapter 23 Electronic Reputation Systems ........................................................................................................... 411 Mario Paolucci, LABSS-ISTC-CNR, Italy Stefano Picascia, LABSS-ISTC-CNR, Italy Samuele Marmo, LABSS-ISTC-CNR, Italy Reputation is a social control artefact developed by human communities to encourage socially desirable behaviour in absence of a central authority. It is widely employed in online contexts to address a number of dilemmas that the interaction among strangers can raise. This chapter presents a social-cognitive theory as a framework to describe the dynamics of reputation formation and spreading. In section 2 we examine the technology of reputation as implemented in some popular Web platforms, testing theory predictions about the tendency towards either a rule of courtesy or a rule of prudence in evaluation reporting, and thus trying to better understand the outcomes that each system promotes and inhibits. Chapter 24 Improving the Information Security of Collaborative Web Portals via Fine-Grained Role-Based Access Control................................................................................................................. 430 S. Demurjian, University of Connecticut, USA H. Ren, University of Connecticut, USA S. Berhe, University of Connecticut, USA M. Devineni, Serebrum Cooperation, USA Sushil Vegad, Serebrum Cooperation, USA K. Polineni, Serebrum Cooperation, USA Collaborative portals are emerging as a viable technology to allow groups of individuals to easily author, create, update, and share content via easy-to-use Web-based interfaces, for example, MediaWiki, Microsoft’s Sharepoint, and so forth. From a security perspective, these products are often limited and

coarse grained in their authorization and authentication. For example, in a Wiki, the security model is often at two ends of the spectrum: anonymous users with no authorization and limited access via readonly browsing vs. registered users with full-range of access and limited oversight in content creation and modiication. However, in practice, such full and unfettered access may not be appropriate for all users and for all applications, particularly as the collaborative technology moves into commercial usage (where copyright and intellectual property are vital) or sensitive domains such as healthcare (which have stringent HIPAA requirements). In this chapter, we report on our research and development effort of a role-based access control for collaborative Web portals that encompasses and realizes security at the application level, the document level (authoring and viewing), and the look-and-feel of the portal itself. Chapter 25 Web 2.0 Effort Estimation .................................................................................................................. 449 Emilia Mendes, The University of Auckland, New Zealand Web effort models and techniques provide the means for Web companies to formalise the way they estimate effort for their projects, and potentially help in obtaining more accurate estimates. Accurate estimates are fundamental to help project managers allocate resources more adequately, thus supporting projects to be inished on time and within budget. The aim of this chapter is to introduce the concepts related to Web effort estimation and effort forecasting techniques, and to discuss effort prediction within the context of Web 2.0 applications.

Volume II Section 7 Educational Applications Chapter 26 A Social Web Perspective of Software Engineering Education .......................................................... 472 Pankaj Kamthan, Concordia University, Canada The discipline of software engineering has been gaining increasing signiicance in computer science and engineering education. A technological revitalization of software engineering education requires a considerate examination from both human and social perspectives. The goal of this chapter is to adopt a systematic approach towards integrating social Web technologies/applications in software engineering education, both inside and outside the classroom. To that regard, a pedagogical patterns-assisted methodology for incorporating social Web technologies/applications in software engineering education is proposed and explored. The potential prospects of such integration and related concerns are illustrated by practical examples. The directions for future research are briely outlined. Chapter 27 University 2.0: Embracing Social Networking to Better Engage the Facebook-Generation in University Life ................................................................................................................................ 496 David Grifin, Leeds Metropolitan University, UK

The social networking Web site is one type of Web 2.0 innovation that has been embraced by universityaged young people. The success of Facebook and similar Web sites has prompted universities to explore how they might use social networking Web sites to engage with their students. In this chapter, I argue that universities are misguided in their attempts to use social networking groups to attempt to engage with students registered with the Web sites. I present empirical evidence from a case study university to substantiate this claim. A framework is developed to categorise the university-related Facebook groups and competing theoretical perspectives on diffusion of innovation are employed to analyse the participation in these groups by students. Recommendations are made for universities, and other organisations, intending to use social networking Web sites to engage with students. Chapter 28 On Using Wiki as a Tool for Collaborative Online Blended Learning ............................................... 511 Steve Wheeler, University of Plymouth, UK This chapter explores the use of the wiki, and its role as a cognitive tool to promote interaction and collaborative learning in higher education. The importance of the software to enable student created content, storage, and sharing of knowledge is reviewed. This chapter provides an evaluation of some of the affordances and constraints of wikis to promote critical thinking within a blended learning context. It assesses their potential to facilitate collaborative learning through community focused enquiry for geographically separated students and nomadic learners. One particular focus of the chapter is the development of new digital literacies and how students present their written work in wikis. The chapter also examines group dynamics within collaborative learning environments drawing on the data from a study conducted at the University of Plymouth in 2007, using wikis in teacher education. Finally, the chapter highlights some recent key contributions to the developing discourse on social software in what has been termed ‘the architecture of participation. Chapter 29 Integration of Web 2.0 Collaboration Tools into Education: Lessons Learned .................................. 522 Phillip Olla, Madonna University, USA Elena Qureshi, Madonna University, USA Web 2.0 is opening new capabilities for human interaction. It also broadens the way technology is used to collaborate more effectively. This chapter discusses instructional strategies and techniques used to successfully utilize Web 2.0 tools for classroom collaboration. It will also shed light on pedagogical issues that arise with the implementation of Web 2.0 into the educational setting. The chapter will present case studies describing how various Web 2.0 applications can be incorporated into a variety of courses in the areas of nursing, education, and computer information systems. Finally, recommendations for teachers and students on how to effectively use Web 2.0 tools to improve collaboration will be outlined. Chapter 30 ECHO: A Layered Model for the Design of a Context-Aware Learning Experience ......................... 539 Hadas Weinberger, HIT – Holon Institute of Technology, Israel

In this chapter, we suggest Echo, a model for utilizing Web technologies for the design of context-aware learning on the Web. Web technologies are continuously evolving to enhance information retrieval, semantic annotation, social interactions, and interactive experiences. However, these technologies do not offer a methodological approach to learning. In this chapter, we offer a new approach to Web-based learning, which considers the role of the user in shaping the learning experience. The key feature in Echo is the analysis and modeling of content for the design of a Web-based learning experience in context. There are three elements in Echo: 1) a methodology to guide the learning process, 2) techniques to support content analysis and modeling activities, and 3) a three-layered framework of social-semantic software. Incorporating this framework facilitates knowledge organization and representation. We describe our model, the methodology, and the three-layered framework. We then present preliminary results from ongoing empirical research that demonstrates the feasibility of Echo and its usefulness for the design of a context-aware learning experience. Finally, we discuss the usefulness of Echo and its contribution to further research in the ield of Web technology. Chapter 31 Advancing Learning Through Virtual Worlds .................................................................................... 556 Steve Mahaley, Duke Corporate Education, USA Robin Teigland, Stockholm School of Economics, Sweden Higher education institutions and corporations are increasingly exploring new pedagogical methods to align with learning styles of incoming students and employees, who are amazingly adept at using Web 2.0 applications. This chapter explores the use of virtual worlds, in particular that of Second Life, in educational activities by organizations such as higher education institutions or corporations. We begin by introducing virtual worlds with a particular focus on Second Life. We then provide an overview of the beneits of this environment for learning activities before presenting a set of potential learning activities that can be conducted within Second Life. We then discuss an in-depth example of 3D teaming-one learning activity within Second Life conducted by the authors. After a discussion of implementation challenges, we then present areas for future research. Chapter 32 Virtual Reality 2.0 and Its Application in Knowledge Building ......................................................... 573 Johannes Moskaliuk, University of Tuebingen, Germany Joachim Kimmerle, University of Tuebingen, Germany Ulrike Cress, Knowledge Media Research Center, Germany In this chapter, we will point out the impact of user-generated online virtual realities on individual learning and knowledge building. For this purpose, we will irst explain some of the central categories of virtual realities (VRs) such as presence and immersion. We will also introduce the term virtual reality 2.0 (VR 2.0), which refers to those new types of VRs that are characterized by typical features of the Web 2.0, such as the opportunity that exists for users to create content and objects themselves. We will explain why we think the term VR 2.0–as a combination of Web 2.0 and VR–is a good label for currently existing user-generated online VRs. This chapter will also explain the concept of knowledge

building, both in general terms and in the Web 2.0 context. The main emphasis of the chapter is on the signiicance of knowledge building for online VRs. In this context, we will describe the visualization of educational content, learner-object interaction, as well as personal, social, and environmental presence as its main features. We will also describe online VRs as a toolbox for user-generated content, and explain why the integration of different tools and seeing “living and learning” in context are relevant for applying user-generated online VRs in educational contexts. In conclusion, we will look at future trends for VR 2.0 environments. Chapter 33 Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education ................. 593 Haya Ajjan, University of North Carolina at Charlotte, USA Richard Hartshorne, University of North Carolina at Charlotte, USA Richard E. Ferdig, Kent State University, USA In this chapter, the authors provide evidence for the potential of Web 2.0 applications in higher education through a review of relevant literature on educational technology and social networking. Additionally, the authors report the results and implications of a study exploring student and faculty awareness of the potential of Web 2.0 technologies to support and supplement classroom instruction in higher education. Also, using the decomposed theory of planned behavior as the theoretical foundation, the authors discuss factors that inluence student and faculty decisions to adopt Web 2.0 technologies. The chapter concludes with a list of recommendations for classroom use of Web 2.0 applications, as well as implications for policy changes and future research. Chapter 34 Social Issues and Web 2.0: A Closer Look at Culture in E-Learning ................................................. 613 Bolanle A. Olaniran, Texas Tech University, USA Hansel Burley, Texas Tech University, USA Maiga Chang, Athabasca University, Canada Developing the foundations for intelligent applications that eficiently manage information is one goal of Web 2.0 technologies and the Semantic Web. As a result, the organization of Web 2.0 and other Semantic Web approaches to learning hold signiicant implications for learning, especially when one considers the role of cultures in learning and e-learning. Exploring how these technologies impact learning, this chapter focuses on social and cultural issues from potential users’ and learners’ standpoints. Furthermore, the chapter offers dimensions of cultural variability as a framework for its arguments. The chapter draws from existing literature and research to present implications of Semantic Web and Web 2.0, along with the issue of digital divide which is critical when exploring access to Web 2.0 technology platforms. The chapter ends by addressing key implications for Web 2.0 and the Semantic Web regarding usage and general effectiveness in the learning context.

Section 8 Enterprise 2.0, Healthcare, Finance, and Other Applications Chapter 35 Enterprise 2.0: Leveraging Prosumerism 2.0 Using Web 2.0 and Web 3.0 ........................................ 630 Chaka Chaka, Walter Sisulu University, South Africa This chapter explores the possibility of synergising Enterprise 2.0 and Web 3.0 through Enterprise 2.0 participation technologies such as blogs, social networking sites (SNSs), media sharing sites (MSSs), and mashups. In short, Enterprise 2.0 is Web 2.0 as applied to the business or commercial domain, and Web 3.0 is a much reined and sleeker Web, extending and improving the offerings of Web 2.0. In addition, the chapter investigates the notion of Prosumerism 2.0 in the context of Enterprise 2.0 and Web 3.0. Against this backdrop, the chapter provides, irstly, a short overview of Enterprise 2.0 and Web 3.0. Secondly, it delineates and discusses the idea of Prosumerism 2.0 in relation to Enterprise 2.0 and Web 3.0. Thirdly, it outlines how Enterprise 2.0 and prosumer-generated content (PGC) can be monetised through harnessing the hybrid participation technologies such as SNSs and MSSs. Chapter 36 Capturing Online Collaboration in the Design Elements Model for Web 2.0 and Beyond ................ 647 T. Andrew Yang, University of Houston-Clear Lake, USA Dan J. Kim, University of Houston-Clear Lake, USA Tri Vu, University of Houston-Clear Lake, USA Vishal Dhalwani, University of Houston-Clear Lake, USA When analyzing the design elements of Web 1.0 applications, Rayport and Jaworski’s 7C Framework (2001) is a model commonly used by researchers. With the advancement of the Web into the Web 2.0 generation, the 7C Framework is insuficient in addressing a critical feature ubiquitously present in Web 2.0 applications, that is, collaboration. In our previous work, we had extended the 7C Framework into the 8C Framework by incorporating the collaboration element in order to capture the collaboration element in Web 2.0 applications (Yang, Kim, Dhalwani, & Vu, 2008). In this chapter, we present the 8C framework as a reference model for analyzing collaborative Web 2.0 applications, including online social networking Web sites and online collaborative sites such as Wikipedia. Chapter 37 A Comparative Analysis of Online Social Networking Sites and Their Business Models ................. 662 T. Andrew Yang, University of Houston-Clear Lake, USA Dan J. Kim, University of Houston-Clear Lake, USA In the world of e-marketing, new business models are introduced to accommodate changes caused by various factors, including the markets, the services, the customers, among others. One latest trend of emarketing is social networking Web sites, many of which have attracted not only large number of users and visitors, but also business companies to place their online ads on the sites. As an important example of Web 2.0 applications, online social networks deserve comprehensive studying and analysis; they are not only employed as an effective vehicle of e-marketing, but may impact how future Web-based ap-

plications would be developed. In this chapter, we explore online social networking as a new trend of e-marketing, by conducting a comparative analysis of online social networking sites. We irst discuss the various types of online social networks, based on the classiication by Laudon & Traver (2008), and then analyze online social networks from a business strategy point of view, by discussing the primary revenue models for online social networking sites. The primary contribution of this chapter is a comparative analysis and discussions of representative online social networking sites and their respective revenue model(s). This chapter aims to provide the reader with a basic understanding of the emerging online social networking Web sites and their primary revenue models. Chapter 38 Healthcare 2.0: The Use of Web 2.0 in Healthcare ............................................................................. 673 Shakib Manouchehri, University of Kassel, Germany Udo Winand, University of Kassel, Germany From an economic, as well as a social point of view, healthcare is a signiicant part of our society and forms a major, ever-growing market. Therefore, this sector has the constant challenge of improving and reducing the cost of services. With respect to interaction, communication, and collaboration between patients and doctors, as well as among each other, the Internet provides new possibilities. Therefore a massive potential for innovation, by so called Web 2.0 applications, is offered. They are also increasingly used via mobile devices. The present article attends to this research with the aim to discuss potentials and restrictions of the use of Web 2.0 applications in healthcare as well as the mobile use of it. Chapter 39 Using a Web-Based Collaboration Portal and Wiki for Making Health Information Technology Decisions ......................................................................................................................... 682 R. Crowell, University of Connecticut Health Center, USA T. Agresta, University of Connecticut Health Center, USA M. J. Cook, University of Connecticut Health Center, USA J. Fiield, University of Connecticut Health Center, USA S. Demurjian, University of Connecticut, USA S. Carter, Community Health Centers, Inc., USA I. Becerra-Ortiz, Fair Haven Community Health Center, USA L. Tracey, StayWell Health Care, Inc., USA S. Vegad, Serebrum Corporation, USA K. Polineni, Serebrum Corporation, USA This chapter presents a case study highlighting development of a Web-based wiki-driven collaboration portal that is being used by a distributed group of community health organizations engaged in developing a strategic implementation plan for health information technology (HIT) at the point of care. The transdisciplinary approach to software development incorporates the perspectives, skill-set, and interests of a diverse group of stakeholders, including staff from the community health organizations, academic researchers, and software developers. The case study describes a select set of the challenges and strategies that have emerged in the planning and development process, including issues surrounding communication, training and development, and infrastructure. Prospects for future development are also explored.

Chapter 40 Assessing the Total Cost of Ownership of Virtual Communities: The Case of the Berlin Stock Exchange................................................................................................................................... 699 Jan vom Brocke, University of Liechtenstein, Principality of Liechtenstein Christian Sonnenberg, University of Liechtenstein, Principality of Liechtenstein Christoph Lattemann, University of Potsdam, Germany Stefan Stieglitz, University of Potsdam, Germany The usage of social software and virtual community platforms discloses opportunities to bridge the natural gap between customers and companies and thus serves as a tool for customer integration. Ideas generated by members of a virtual community can be utilized to innovate and improve the company’s value adding activities. However, the implementation and operation of virtual communities may have a signiicant impact on the inancial performance of a company. Hence, to measure the proitability of a virtual community appropriately, means of eficiency calculations have to be employed. The objective of this chapter is, therefore, to develop a measurement framework to evaluate the inancial performance of a virtual community. The focus is on calculating the total cost of ownership. After introducing a general measurement framework, a particular measurement system is derived from the framework and is subsequently applied to a real life example of the Berlin Stock Exchange. Chapter 41 Connecting the Real World with the Virtual World: The SmartRFLib RFID-Supported Library System on Second Life .......................................................................................................... 720 Katinka Kromwijk, ETH Zurich, Switzerland Çağrı Balkesen, ETH Zurich, Switzerland Gautier Boder, ETH Zurich, Switzerland Nihal Dindar, ETH Zurich, Switzerland Florian Keusch, ETH Zurich, Switzerland Ali Şengül, ETH Zurich, Switzerland Nesime Tatbul, ETH Zurich, Switzerland With recent developments in Web technologies enabling interaction in virtual environments, as well as the ones in sensor network technologies enabling interaction with the real world, we see an emerging trend towards bringing these two worlds together. In this chapter, we share our experiences in building an RFID-supported library system on Second Life called SmartRFLib, which successfully achieves this integration. Although SmartRFLib focuses on a library system as an application scenario, it has been designed as a general-purpose RFID data management and complex event detection system, and can also be used as a basis to build other RFID-based event monitoring applications. Chapter 42 Embracing the Social Web for Managing Patterns ............................................................................. 733 Pankaj Kamthan, Concordia University, Canada In this chapter, the affordances of the social Web in managing patterns are explored. For that, a classiication of stakeholders of patterns and a process for producing patterns are proposed. The role of the

stakeholders in carrying out the different worklows of the process is elaborated and, in doing so, the prospects presented by the technologies/applications underlying the social Web are highlighted. The directions for future research, including the potential of the convergence of the social Web and the Semantic Web, are briely explored. Chapter 43 Extending and Applying Web 2.0 and Beyond for Environmental Intelligence ................................. 748 Bhuvan Unhelkar, University of Western Sydney & MethodScience.com, Australia Bharti Trivedi, DD University, India This chapter aims to apply the intelligence used in businesses decision making to an organization’s environmental management strategy so as to support its green credentials. While the World Wide Web (WWW or Web for short) has had an impact on every aspect of human life, its current and upcoming versions, dubbed Web 2.0 and beyond, need to be considered in the context of environmental management. The use of decision making technologies and processes in this area of an organization is what we call “environmental intelligence” (EI). This EI can be used by businesses in order to discharge one of their signiicant corporate responsibilities–that of managing their activities that affect the environment including waste reduction, green house gas reduction, recycling, minimizing unnecessary human and material movements, and so on. Furthermore, the use of EI, it is envisaged, will also help organizations create local and industrial benchmarks, standards, audits, and grading that will help a large cross section of businesses to comply with the environmental requirements. The architecture of such enterprise intelligent systems needs to incorporate technologies like executable services, blogs, and wikis in addition to the standard communication and execution requirements of the Web. This chapter describes the literature review and the initial output of the research being carried out by the authors which, we hope, will eventually result in an environmentally intelligent Web-based business strategic system (EIWBSS).

Section 9 Social Web: Foundations, Analysis, and Visualisation Chapter 44 Social Software and Web 2.0: Their Sociological Foundations and Implications .............................. 764 Christian Fuchs, University of Salzburg, Austria Currently, there is much talk of Web 2.0 and social software. A common understanding of these notions is not yet in existence. Also the question of what makes social software social has thus far remained unacknowledged. In this chapter, a theoretical understanding of these notions is given. The Web is seen in the context of social theories by thinkers like Emile Durkheim, Max Weber, Ferdinand Tönnies, and Karl Marx. I identify three levels in the development of the Web, namely Web 1.0 as a web of cognition, Web 2.0 as a web of human communication, and Web 3.0 as a web of cooperation. Also, the myths relating to Web 2.0 and its actual economic and ideological role in contemporary society are discussed.

Chapter 45 Sociology of Virtual Communities and Social Software Design ........................................................ 790 Daniel Memmi, University of Quebec in Montreal, Canada The Web 2.0 movement is the latest development in a general trend toward computer-mediated social communication. Electronic communication techniques have thus given rise to virtual communities. The nature of this new type of social group raises many questions: are virtual communities simply ordinary social groups in electronic form, or are they fundamentally different? And what is really new about recent Web-based communities? These questions must irst be addressed in order to design practical social communication software. To clarify the issue, we will resort to a classical sociological distinction between traditional communities based on personal relations and modern social groups bound by functional, more impersonal links. We will argue that virtual communities frequently present speciic features and should not be assimilated with traditional communities. Virtual communities are often bound by reference to common interests or goals, rather than by strong personal relations, and this is still true with Web 2.0 communities. The impersonal and instrumental nature of virtual communities suggests practical design recommendations, both positive and negative, for networking software to answer the real needs of human users. Chapter 46 Online Human Activity Networks (OnHANs): An Analysis Based on Activity Theory .................... 804 Dan J. Kim, University of Houston-Clear Lake, USA T. Andrew Yang, University of Houston-Clear Lake, USA Ninad Naik, University of Houston-Clear Lake, USA Recently, Web 2.0 applications such as blogs, wikis (e.g., Wikipedia), social networks (e.g., MySpace), 3-D virtual worlds (e.g., Second Life), and so forth, have created fresh interest in the Internet as a new medium of social interactions and human collaborative activities. Since the emergence of Web 2.0 applications, Web services that support online human activities have gained an unprecedented boost. There have been conceptual studies on and overviews of individual Web 2.0 applications like blogs, online social networks, and so forth, but there has not been a study to date which provides a theoretical perspective on the online human activity networks (OnHANs) formed by these Web 2.0 applications. In this chapter, we classify various forms of OnHANs focusing on their social and business purposes, analyzing the core components of representative OnHANs from the angle of the activity theory, and inally providing a theoretical discussion concerning how OnHANs provide values to the individuals and the organizations involved in those activities. Chapter 47 Visualising Social Networks in Collaborative Environments ............................................................. 817 Stephen T. O’Rourke, The University of Sydney, Australia Rafael A. Calvo, The University of Sydney, Australia Social networking and other Web 2.0 applications are becoming ever more popular, with a staggering growth in the number of users and the amount of data they produce. This trend brings new challenges to the Web engineering community, particularly with regard to how we can help users make sense of all

this new data. The success of collaborative work and learning environments will increasingly depend on how well they support users in integrating the data that describes the social aspects of the task and its context. This chapter explores the concept of social networking in a collaboration environment, and presents a simple strategy for developers who wish to provide visualisation functionalities as part of their own application. As an explanatory case study, we describe the development of a social network visualisation (SNV) tool, using software components and data publicly available. The SNV tool is designed to support users of a collaborative application by facilitating the exploration of interactions from a network perspective. Since social networks can be large and complex, graph theory is commonly used as a mathematical framework. Our SNV tool integrates techniques from social networking and graph theory, including the iltering and clustering of data, in this case, from a large email dataset. These functions help to facilitate the analysis of the social network and reveal the embedded patterns of user behaviour in the underlying data. Chapter 48 The Discourses of Empowerment and Web 2.0: The Dilemmas of User-Generated Content ............ 828 Yasmin Ibrahim, University of Brighton, UK Consumer content generation in the Web 2.0 environment from a libertarian perspective is about the democratization of mediated knowledge where it creates the possibilities to produce new knowledge and media economies in a post modern world. This chapter examines the notions of empowerment afforded by multimedia technologies on the Internet where new forms of knowledge, politics, identity, and community can be fostered through the Web 2.0’s architecture of participation, collaboration, and openness. It also discusses how these unlimited possibilities to produce content present new social and ethical dilemmas. They not only challenge conventional ways in which knowledge and expertise have been constructed in modern and postmodern societies but also require more rigorous methods to identity what can constitute expert knowledge. The production of user-led taxonomies and data repositories has raised the need to re-examine user-generated content and its function and coexistence within the existing systems and archives of knowledge. Chapter 49 How Employees Can Leverage Web 2.0 in New Ways to Relect on Employment and Employers .................................................................................................................................... 846 James Richards, Heriot-Watt University, UK How and why businesses can and should exploit Web 2.0 communication technologies for competitive advantage has recently become the focus of scholarly attention. Yet at the same time, one key organizational actor in the business equation–the employee as an individual and collective actor with distinct interests from that of the employer, has been given scant attention. Using media accounts, questionnaire and interview data, this chapter seeks to map out early trends in employee interests in Web 2.0. The indings point towards three distinct, yet interconnected employee uses for Web 2.0–collaborative practices that extend employee abilities to exchange a wide-range of ‘insider information,’ express conlict, and ‘take action’ against employers. Due to the nature and size of cyberspace, however, more research is required to gauge the popularity and effect of these emergent trends.

Chapter 50 Privacy Implications and Protection in the New Ubiquitous Web Environment ................................ 863 Charalampos Z. Patrikakis, National Technical University of Athens, Greece Ioannis G. Nikolakopoulos, National Technical University of Athens, Greece Athanasios S. Voulodimos, National Technical University of Athens, Greece In this chapter, we are addressing the issue of privacy in our modern world of Internet, Web 2.0, personalization, location based services, and ubiquitous computing. The issue is initially viewed from the perspective of user proiles, starting from existing approaches used in social networking and mobile computing applications. Emphasis is given on the separation of personal and public information and the way it can be used in Web and mobile applications. Furthermore, identifying the importance and the actual meaning of privacy in an online world is a crucial and dificult task, which has to be carried out before trying to propose ways to protect the users’ privacy.

Epilogue ............................................................................................................................................. 878 Compilation of References ............................................................................................................... 880

xl

Preface

I look to the future because that’s where I’m going to spend the rest of my life. -- George Burns (1896-1996)

The World Wide Web has just turned 20! Within this short span of time, it has caused one of the most signiicant and inluential revolution of modern times; its inluence has impacted almost every aspect of our life and activities and almost all ields, irrevocably. And, in the past few years, it has evolved quite rapidly into Web 2.0, Web 3.0, and so on, forging many new applications that were not previously feasible. The Web has also caused paradigm shifts and transformational changes in business, social interaction, governance, and education, among others. The Web’s evolution continues, and there is no sign of it stopping. And, we are yet to discover and exploit the Web’s full potential. Perhaps we might not realize its full potential soon, as we don’t yet know what its full potential is. And, its potential is expanding in unanticipated directions. But what we can say is the Web’s future is very bright, and its inluence on us would be much greater than what it has been. This book is a humble attempt to present Web’s evolution in the recent years and portray its major inluences in different areas, and to look at the phenomenal evolution of Web from different perspectives – technological, business, and social, comprehensively and holistically. The book outlines new generation Web – Web 2.0, 3.0, and X.0 – and its applications, both existing and emerging, and how they are transforming our lives, work, education, and research. The book also presents some interesting new research that helps us in creating new kinds of applications that were unimaginable before. This Handbook of Research on Web 2.0, 3.0, and X.0: Technologies, Business, and Social Applications is a comprehensive reference that explores the opportunities and challenges the new generation of Web technologies and applications present, illustrated with real world examples and case studies, and examines the technical, social, cultural, and ethical issues these applications present. We believe the handbook provides valuable insights for further research on new generation Web technologies, applications, and social issues. We hope this book fulills its major objective of being an excellent resource for researchers, academics, and professionals seeking to explore the issues and emerging trends in Web and Web-based applications. The book also serves as reference for senior graduate students who want to get a glimpse of emerging new applications and garner some new ideas that they might want to pursue further. To help you easily navigate this volume, next, let me give you a peek into the handbook.

xli

A Preview of whAt’s inside In this handbook of research, we feature 50 carefully selected chapters that together present a big picture of the new generation Web and its applications and recent research work in this area. For easy identiication and comprehension, we present these chapters under nine themes: 1. Overview; 2. Web Modeling and Design; 3. Web Architecture; 4. Information Search, Bookmarking, and Tagging; 5. Semantic Analysis and Semantic Web; 6. Web Quality, Trust, Security, and Effort Estimation; 7. Educational Applications; 8. Enterprise 2.0, Healthcare, Finance, and Other Applications; and 9. Social Web: Foundations, Analysis, and Visualization. Overview. We begin the journey by providing you an overview of the Web’s evolution in the irst two chapters. By presenting a comprehensive overview of new generation Web, these chapters refresh or prepare you for gaining a better understanding and appreciation of technologies, applications, and issues discussed in the rest of the chapters. The irst chapter traces the Web’s continuing evolution and phenomenal strides, outlines the features and characteristics of Web 2.0, 3.0, and X.0 and examines their prospects and potential. The second chapter presents interesting perspectives on the Web X.Y movement, synthesizes new deinitions for the Web X.Y, and classiies well-known Web applications according to these deinitions. Web Modeling and Design. In this section, we present some of the technological aspects that lay the foundation for new generation Web applications. Easy-to-use, interactive user interface is a hallmark of Web 2.0 applications that appeals to users. First, we introduce a model-driven approach that incorporates interaction models to design of rich Internet applications (RIAs) and illustrate it with a case study, followed by modular interface design for RIAs, and a conceptual model that captures novel features of RIA features and that can be automatically converted into implementations in popular RIA technologies and frameworks. We also outline how the design tool WebRatio and its companion conceptual model based on WebML can be extended to support the new requirements imposed by RIAs. We also explore how to merge Web 2.0 technology with grid computing overlay to support the Web 2.0 framework and illustrate this idea with a case study--managing health information based on users’ experiences. Web Architecture. Then focusing your attention on Web architecture, we present criteria for evaluation of RIA architectures; an immersive Web X.0 framework for e-commerce, a mobile service oriented architecture for businesses, and a unifying architecture for next generation Web applications. Information Search, Bookmarking, and Tagging. Then turning your attention to the Web application arena, in six chapters, we outline how Web’s evolution is inluencing and improving information search, bookmarking, and tagging, all major activities of Web users. We present an overview on folksonomies, which is a relatively new concept that hasn’t been widely studied, and on social semantic bookmarking, a novel paradigm that combines the positive aspects of semantic annotation with those of social bookmarking and tagging while avoiding their respective drawbacks. We also outline the promises of social bookmarking for enhancing Web search and for building novel applications. Next, we present a comparative analysis of two popularity measures of Web pages, PageRank and SBRank, which is deined as an aggregate number of bookmarks that a given page accumulates in a selected social bookmarking system. For realizing a more effective search, we illustrate how SBRank and PageRank measures could be combined to re-rank results delivered by Web search engines Collaborative tagging, popularized by Web sites such as Del.icio.us and Flickr, has now become quite popular. We present a study on social tagging and their applications and on social tagging analysis and mining. We also outline how cross-language information retrieval could be improved by effectively harnessing advances in social Web and how user-speciied metadata could be used to personalize image search.

xlii

Semantic Analysis and Semantic Web. Web 3.0, which encompasses the Semantic Web, is on the rise. Hence, in our coverage, we look at the Semantic Web and semantic analysis, focusing on a couple of key aspects. Effectively harnessing blogs, wikis, social networks, online content sharing, and online collaboration, the Web has been swamped with user generated content (USG). USG is one of the key features of new generation Web and has created new value chains and business models. In this section, we deal with topics such as accessing, analyzing, and extracting information from USG, wiki semantics, and means of disambiguating social tags, also known as folksonomies. Web Quality, Trust, Security, and Effort Estimation. Today, major issues confronting the Web, particularly many Web 2.0 applications, are quality of contents and applications, trust and security. In this section, we discuss how to model content quality for the Web 2.0 applications, and then present a Web site quality assessment model. Next, we present an electronic reputation system to encourage socially desirable online behavior in absence of a central authority, as well as the dynamics of reputation formation and spreading, and a role-based access control for collaborative Web portals that realizes security at different levels of the portal. We also present effort estimation concepts for new generation Web applications. Educational Applications: Education and training has been an early and a major adopter of Web 2.0 and there have been several applications based on Web 2.0, transforming signiicantly how students gather and contribute information, interact, collaborate, and learn. In this section, we examine several key aspects of learning in the networked age, covering a range of topics, including: integrating social Web technologies and applications in software engineering education, both inside and outside the classroom; a pedagogical patterns-assisted methodology for incorporating social Web technologies/applications in software engineering education; embracing social networking to better engage the Facebook-generation in their university life; use of the wiki and its role as a cognitive tool to promote interaction and collaborative learning in higher education; and instructional strategies and techniques for successfully harnessing Web 2.0 tools for classroom collaboration and pedagogical issues that arise in these settings. In addition, in this section, we describe a system that facilitates context-aware learning on the Web, present a study on learning in virtual worlds, discuss the role of virtual reality 2.0 that characterizes typical features of the Web 2.0 and its application in knowledge building by enabling users create content and objects themselves; report the indings of a study on student and faculty use and perceptions of Web 2.0 technologies in higher education; and social and cultural issues in Web 2.0-based learning environments from potential users’ and learners’ perspectives and key implications for Web 2.0 and the Semantic Web on general effectiveness in the learning context. Enterprise 2.0, Healthcare, Finance, and Other Applications. Under this theme, we cover a range of topics of growing signiicance: Prosumerism 2.0 in the context of Enterprise 2.0 and Web 3.0; an 8C framework for analyzing collaborative Web 2.0 applications; comparative analysis of popular online social networks and their business models; healthcare 2.0 - the use of Web 2.0 in healthcare; a case study on a collaboration portal and Wiki that supports health information technology decisions; examination of impact of virtual communities on the inancial performance of a company, highlighting the Berlin Stock Exchange as an example; an RFID-supported library system on Second Life called SmartRFLib; embracing the social Web for managing patterns; and the use of Web 2.0 in environmental decision making - environmental intelligence (EI). Social Web: Foundations, Analysis, and Visualization. On our concluding theme, social Web, we tackle some interesting problems and issues. Though the terms social Web and social software have been widely used and talked about, to many, what makes social software social remains unclear. In the chapter, “Social Software and Web 2.0: Their Sociological Foundations and Implications,” we answer this question by examining Web in the context of social theories by thinkers like Emile Durkheim, Max

xliii

Weber, Ferdinand Tönnies, and Karl Marx, and view Web 1.0 as a web of cognition, Web 2.0 as a web of human communication, and Web 3.0 as a web of cooperation. Then, we examine the sociology of virtual communities and social software design and attempt to answer the question: Are virtual communities simply ordinary social groups in electronic form, or are they fundamentally different, and what is really new about recent Web-based communities? Then we classify various forms of online human activity networks (OnHANs) formed by Web 2.0 applications based on their social and business objectives, and provide a theoretical discussion on how these networks provide values to the individuals and the organizations involved in those activities. We present a simple strategy for developers to provide visualization functionalities to social networks, illustrating it with a case study. Then, focusing your attention on USG, we discuss how the unlimited possibilities that Web users now have to produce and widely share their content on the Web present new social and ethical dilemmas. We also report on a study on employee uses for Web 2.0 that came up with interesting indings: Employees use of Web 2.0 applications to share a wide-range of ‘insider information,’ express conlict, and ‘take action’ against employers. In our last chapter, we address the issue of privacy in our modern networked world supported by the Internet, wireless communications, Web 2.0, personalization, location based services, and ubiquitous computing.

in Closing I take pleasure in presenting you this comprehensive handbook that covers a range of areas and issues of current interest in the context of the Web’s evolution. I believe this handbook of research presents useful insights and ideas about the new generation Web and how you can embrace its potential. I also believe, whether you are a researcher, an academic, or a practicing professional seeking to explore the prospects and potential of new generation Web or a senior graduate student who wish to get a glimpse of emerging Web applications and some new ideas, you will ind the book a very helpful guide and a comprehensive informative resource. If all this sounds promising, read on! As Francis Bacon said, “Some books are to be tasted, others to be swallowed, and some others to be chewed and digested.” I hope, depending on your interest and need, you ind some things in this handbook to chew and digest and some other things to taste. If you think this book might be useful to someone you know, please recommend it to them. And, I welcome your comments and feedback on the handbook at [email protected]. Now, I am delighted to hand over the handbook to you.

San Murugesan October 2009 [email protected] www.webhandbook.info

xliv

Acknowledgment

As it is obvious, publication of this handbook wouldn’t have been possible without the contribution, support, and cooperation of several people. I would like to acknowledge with thanks their contribution. First, I would to thank each one of the authors for enthusiastically contributing to the handbook, and thereby sharing his/her research work and insights with the readers of this book. I also gratefully acknowledge their patience, support, and cooperation. Reviewers play a signiicant role in ensuring the quality and relevance of a publication, and this publication is no exception. I thankfully acknowledge valuable contributions of our reviewers (see page … ) in improving the quality of the chapters. Next, I would also like to thank the members of the Handbook Advisory Board (see page … ) for their advice and suggestions. The editorial team at IGI Global deserves my commendation for their key roles in publishing this volume and in ensuring its quality. In particular, I would like to thank Ms. Christine Bufton, Editorial Communications Coordinator, for her excellent enthusiasm, support, and cooperation. I also thank Prof. In Lee, Editor-in-Chief, Advances in E-Business Research Series (AEBR) Book Series, for his continued support and encouragement. It is not out of place to thank the marketing team at IGI Global for widely promoting the book to those who might beneit from it. Finally, I like would to thank my wife, Vijayakumari, who has been a constant source of inspiration and encouragement to me in making this book a reality and for providing the beautiful “OM” (also known as “AUM”) image that appears on the dedication page. I also thank my other family members, Nithya, Ravi Kumar, Suresh, and Sangeetha, for their support and well wishes. San Murugesan

A New Web Site Quality Assessment Model for the Web 2.0 Era

Tag Bowers as shown in Figure 2, you get highly relevant tags, as well as search results limited to Web 2.0 (the image in the center). ‘Appearance’ means the degree to which color, graphics, images, font, style, and animations are properly and consistently used. Some other studies mention this dimension as aesthetics or ‘look and feel’. A website should display visually appealing design (Kim et al., 2002). Selecting right colors with consideration of brightness or contrast makes users visually comfortable, while using inconsistent styles throughout a website makes users confused and lose interest in. In the Web 1.0 environment, when a website user requests certain information, the server transmits the entire information to the client, so that it is a very complex task to render dynamic graphics as shown in applications installed in PC. In the Web

2.0, however, that limitation has been overcome by technological progress. Specifically, RIA (Rich Internet Application), providing richer user interface, has been realized in Web 2.0 environment with the use of Ajax, Adobe Flex, Microsoft Silverlight, etc. (Moroney, 2007; Ogawa & Goto, 2006). As shown in Figure 3, the website using Flex can realize rich user interface more dynamically and elegantly. ‘Layout’ implies the degree to which visual elements such as texts, forms, frames, or tables are well placed and organized in a page to be easily recognizable and usable to user. For example, a table too wide to be showed in a screen without a scrollbar is inconvenient for users’ to browse. Brinck et al. (2002) point out that the goals of proper layout are simplicity, consistency, and focus. Nonetheless, layout needs to be designed

Figure 2. Fliker Related Tag Browers provides relevant subjects with links including search results (from http://www.airtightinteractive.com/projects/related_tag_browser/app/) © 2009 Airtight Interactive. Used with permission.

395

1

Chapter 1

Web X.0: A Road Map San Murugesan Multimedia University, Malaysia & University of Western Sydney, Australia

AbstrACt The Web has evolved from its humble beginnings merely as a publishing medium intended for a small group of scientists to a medium of interaction, participation, and collaboration. It has dramatically inluenced almost every sphere of our activity and has created paradigm shifts. Encompassing new technologies, business strategies, and social trends, the Web continues to forge many new applications that we had never imagined before or were not previously feasible. It has created new paradigms in business, social interaction, governance, and education. In this chapter, we trace the Web’s continuing evolution and phenomenal strides, outline the features and characteristics of Web 2.0, 3.0, and X.0, and examine their prospects and potential. The ability to recognize new Web technologies for their potential in business, social and educational applications, and the ability to develop and deploy creative applications based on these technologies are the keys to continued success of the Web and our progress and well being.

introdUCtion The Web has become the most significant technology of the 21st century. In its rapid rise, it has caused many welcome disruptions. For instance, it has made people change how they gather information, do their work, buy goods and services, connect with friends and family, spend their leisure time, and even find their partner and lost friends and acquaintances. It DOI: 10.4018/978-1-60566-384-5.ch001

has also forced businesses to rethink and change how they conduct business, connect with their customers and suppliers, innovate, and collaborate. Furthermore, the Web has changed even the face of politics and governance. Since its inception 20 years ago, the Web has evolved steadily and significantly and still continues to evolve along multiple directions. The nature and structure of the Web, as well as the way we use it, have been continuously changing. The Web evolution is huge that we have started to place the

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Web X.0

evolution—past, current, and anticipated—into different stages as Web 1.0 (the traditional Web), Web 2.0, Web 3.0, and so on. While the use of the terms Web 2.0 and Web 3.0 have become quite common now, they however, defy a widely agreed-upon, concise definition, perhaps because “the underlying phenomenon is so huge and important that it resists any attempt to pin it down.” These terms can be described from different viewpoints and in different ways depending on intended application; each of them is considered a collective term. The Web’s evolution, which we call Web X.0, or Web X.Y, movement, is aimed at harnessing the potential of the Web in a more interactive and collaborative manner with an emphasis on social interaction. It is also aimed at facilitating collaboration and leveraging the collective intelligence of peers, as well as of collective information available on the Web by judicious use of old and new Web technologies in new ways. Web 2.0 has become a mainstream technology now. Motivated by some highly successful social and business applications based on Web 2.0, such as MySpace, Linked-in, SecondLife, Flickr, and YouTube, Web 2.0 technologies and concepts are now widely used in several different domains. And within five years, as you can recognize, Web 2.0 has changed the face of the society and business significantly and has forged into enterprises in ways that were previously unimaginable. Web 3.0 has begun to make its headway and its promises are even more significant, and we are yet to experience its influence and impact. Given these scenarios, can you afford to simply ignore Web 2.0 and 3.0 – and the future incarnations of the Web - considering them simply as hype or a passing fad, as some educated skeptics do? Certainly not! In fact, you should harness and embrace them. And researchers in all areas – not just information and communication technology - must identify and address the problems and challenges the new generation Web pose and devise new ways of using them harnessing their potential.

2

In this chapter, setting the background for the chapters that follow, we outline the features and characteristics Web 2.0, 3.0 and X.0 and examine their prospects and potential.

web X.0: whAt does it rePresent As we pointed earlier, we can set the ongoing Web’s evolution into stages: Web 1.0, Web 2.0, Web 3.0, and Web 4.0 (see Figure 1). One way of identifying them based on what they do and who or what is at the core of their action. The first stage, Web 1.0, is about connecting information; Web 2.0 is about connecting people; Web 3.0 is about integrating data, knowledge, and applications on the Web and putting them to work in ways that make the Web more meaningful and about making Web as a collaborative platform; and Web 4.0 is about harnessing the power of human and machine intelligence on a ubiquitous Web, where both people and computers not only interact, but also reason and assist each other in smart ways (Murugesan, 2007c).

web 1.0 The traditional Web—now called Web 1.0 -- is primarily a one-way publishing medium. The primary objective has been to publish information for easy access by anyone using a standard Web browser through the Internet. Subsequently, it was put to use for commercial applications and online transactions giving birth to the emergence of electronic commerce, or e-commerce. Foundations for the Web were set in this phase. The major developments and advancements were protocols such as HTTP, markup languages such as HTML and XML, Web-centric languages such as Java and JavaScript, Web browsers, Web development platforms and tools, the creation of Web sites academic activities, the use of the Web for commercial purposes for the first time, emergence of

Web X.0

Figure 1. The evolution of the Web, source Murugesan (2007c)

some new innovative Web business models, and the growth of Web portals. Web 1.0 has been, and is, information-centric.

web 2.0 In 2004, Tim O’Reiley of O’Reilly Media coined the term Web 2.0. Web 2.0 allows – and encourages – all the users to create, share, and distribute information and images. In fact, Web 2.0 has caused a social revolution in the use of Web, and caused a paradigm shift from being a publishing medium to a participative medium. In other words, Web 2.0 technologies and applications have democratized the Web. Hence, it can be called as democratic Web. Web 2.0 encompasses Web technologies and services, such as blogs, social networking sites, wikis, communication tools, and folksonomies that emphasize sharing of content among users and online collaboration. It is also a highly interactive, dynamic application platform for fielding new kinds of applications. As Lin (2007) has noted, “Web 2.0 represents a paradigm shift in how people use Web. While most users were once limited to passively viewing

Web sites created by a small number of providers with markup and programming skills, now nearly everyone can actively contribute content online. Technologies are important tools, but they are secondary to achieving the greater goal of promoting free and open access to knowledge.” Thus, Web 2.0 is people-centric. Although Web 2.0 began simply as a consumer phenomenon, attracting numerous users for, and contributors to, blogs, social networks, and online information resources like Wikipedia, it has significantly impacted other application areas as well. In the last five years, a wide array of Web 2.0 applications was deployed for business and societal use, and many innovative online services have emerged - some of them are offered free to users. Many enterprises are reaping significant benefits from Web 2.0 by harnessing it for product development, market research, competitive intelligence gathering, and revenue generation.

web 3.0 In 2006, John Markoff, in an article published in The New York Times in 2006, called the next phase in the Web’s evolution, Web 3.0. Web 3.0

3

Web X.0

Table 1. What’s in the name Web 1.0 • Information-centric Web • Read only Web • Web of cognition

Web 2.0 • People Centric Web • Read-write Web

refers to a third generation of Web technologies and services that emphasize a machine-facilitated understanding of information on the Web in order to facilitate information aggregation and to offer a more productive and intuitive user experience. Web 3.0 is also called Semantic Web or meaningful Web. Under the umbrella of Semantic Web and Web 3.0, currently, significant developments are taking place and new Web 3.0 applications have begun to emerge.

web 4.0 While Web 3.0 is advancing and is marching toward main stream adoption soon, we name the next phase in Web’s evolution Web 4.0, or “Web X.0.” The objective of Web 4.0 is to add it further sophistication and higher levels of intelligence. For instance, in a Web 4.0 application, your software agent(s) roaming on the Internet or simply residing on your computer could reason and communicate with other such agents and systems and work collaboratively to accomplish things on your behalf. Web 4.0 is also known as “intelligent Web” or “smart Web.”

what’s in the name While many people support the idea of using version-like numbers to represent each phase of the Web’s evolution, others – a diminishing minority - are not for categorizing the Web and naming it like versions of software since, in their view, there is nothing significantly new to warrant a new name. There is, however, merit in naming them as

4

Web 3.0 • Machine-Centric Web • Semantic Web • Meaningful Web • Web of cooperation

Web 4.0 • Agent-centric Web • Smart Web • Intelligent Web • Web of Collaboration

proposed. First, these new names give researchers, application developers and the general public the notion that the Web is advancing and evolving to the next stage, and we are making progress. Secondly, perhaps more importantly, these names also encourage us to look into the advancements and their promises and potential and explore how we can exploit them to our advantage, making the Web experience even better. Table 1 lists various names in use for Web X.0. While some of us might not like the terms Web 2.0, 3.0, or X.0, all of us, of course, like to enthusiastically embrace the technologies, concepts, and applications they offer to us.

Coexistence of web X.0 The use of version-like numbers to represent each phase of the Web’s evolution, however, doesn’t mean that Web 1.0 is superseded by Web 2.0, or Web 2.0 by Web 3.0 – like in software revision or update. In this context, each stage has different objectives, as we had outlined, and specifically addresses different problems and offers different features -- all of which we need. Web 1.0, 2.0, 3.0, and 4.0 will continue to coexist, one supporting or forming the foundation for the others, as depicted in Figure 1. In Web applications, we still need the basic foundations of Web 1.0, and depending on the application, we can draw on other required features offered by Web 2.0, 3.0, and 4.0. Having looked at a big picture of the Web’s evolution, let’s now further examine Web 2.0, 3.0, and X.0 in some detail and their promises and its current status.

Web X.0

eXPloring And eMbrACing web 2.o Web 2.0, which is also known by different names such as Wisdom Web, People-centric Web, Participative Web, and Read/Write Web, is both a new usage paradigm and a new technology paradigm. It is an umbrella term or a collective term; it represents a collection of important collection of technologies, business strategies, and social trends (Murugesan, 2007a, 2007b). Web 2.0 is more dynamic and interactive than its predecessor, Web 1.0, letting users both access content from a Web site and contribute to it. Web 2.0 lets users keep up with a site’s latest content even without visiting the actual Web page. It also lets developers easily and quickly create new Web applications that draw on data, information, or services available on the Internet (Murugesan, 2007b). Web 2.0 is an umbrella term encompassing several new Web technologies: blog, wiki, mashup, social networks, RSS, tags, and syndication. For a brief overview of these technologies and support tools available for development of Web 2.0 applications refer to Murugesan (2007b). The architecture and technologies that make up Web 2.0 offer several key features, including: • • • • • •

Facilitating lexible Web design, creative reuse, and easier updates Providing a rich, responsive user interface Supporting collaboration and assisting in gathering collective intelligence Facilitating collaborative content creation and modiication by users Establishing social networks of people having common interests Enabling the creation of new attractive applications by reusing, combining, and/or merging different applications on the Web or by combining data and information from different sources

web 2.0 framework The Web 2.0 framework, shown in Figure 2, presents all the key elements of Web 2.0. There are three key parts to the Web 2.0 Framework, as outlined below: 1.

2. 3.

Web 2.0 is founded on seven key characteristics: participation, standards, decentralization, openness, modularity, user control, and identity. Web 2.0 is expressed in two key domains: the open Web and the enterprise. The heart of Web 2.0 is how it converts inputs such as user generated content (USG), opinions, applications, through a series of processing activities that involve recombination, collaborative filtering, structures, syndication to create emergent outcomes that are of value to the user and the entire community.

Because of many welcome features that supports and embraces user involvement and interaction, within five years it has become mainstream technology and application. As Dion Hinchcliffe (2009) notes, “Web 2.0 became vitally important -- even central in some cases -- to the very future of global culture and business. ... The concepts identified as Web 2.0 have proved to be highly insightful, even prescient, and are used around the world daily to guide everything from product development to the future of government.” The concepts identified as Web 2.0 have proved to be highly insightful, even prescient, and are used around the world daily to guide everything from product development to the future of government.

5

Web X.0

Figure 2. A holistic picture of Web 2.0, source www.futureexploration.net

eXPloring And eMbrACing web 3.o Web 3.0 represents the evolution of Web usage and interaction along several separate paths. For instance, Web 3.0 is about “transforming the Web into a database, a move towards making content accessible by multiple non-browser applications, the leveraging of artificial intelligence technologies, the Semantic Web, the Geospatial Web, or the 3D (three-dimensional) Web.” According to another similar perspective, Web 3.0 is “the Semantic Web; a 3D Web; a mediacentric Web; a pervasive Web; a large database presented as Web pages; or a combination of all of these (Metz, 2007; Murugesan, 2007)”.

6







Semantic Web. Providing better connections between blocks of information, the Semantic Web facilitates software applications that can anticipate what you really want to know or do. For example, when you read about a ilm on the Web, you can immediately get links to a lot of related content and services. 3D Web. This enables you to present threedimensional images on the Web and thereby to create virtual worlds. The Web as a 3D space presents several opportunities for new services, including new ways of living virtually online. Media-centric Web. This refers to an advanced, media-rich Web. Among other things, it might allow you to ind

Web X.0





media from media, also known as a “visual search.” For example, by presenting a photo of a building or your favorite painting to a search engine, you can get several photos that are similar to the one you presented to the search engine. Similarly, you could retrieve an entire song from a search engine when you present a small section of the song. Pervasive Web. The Internet and Web will become more pervasive as many gadgets and household items such as TVs, refrigerators, microwaves, and heaters are connected to the Internet and have a built-in Web browser for Web access. Database as Web pages. We can access and manage a database as Web pages openly and easily. We can also have control over our data through the Web pages.

Web 3.0 is also seen as contextual Web that is “increasingly verticalized by context, and the relevant content, community, and commerce elements are successfully mashed up ‘in context’” augmented by vertical or contextual search and personalization (Mitra, 2007). Rolling up these six elements as Web 3.0 open ups a whole set of new, personalized applications such as smart, virtual, personal shopping assistant, as outlined by Mitra (2007). As this author summarized in his report (Murugesan, 2007), “Web 3.0 is not just the Semantic Web. Neither is Web 3.0 just a collection of virtual worlds, nor is it the mobile Web. It is possibly an entry-level Semantic Web that can be visualized by virtual worlds and accessed through desktops, as well as handheld devices such as mobile phones, PDAs, and pocket PCs. For this vision to be realized, however, several new developments must take place. These developments include embedding of semantic specification into virtual worlds and the interpretation and specification of semantics through mobile devices, cross-site ID

recognition, and cross-site identification about authority of information.

web 3.0 is gaining Momentum Web 3.0 has begun to gain momentum and will eventually succeed, as it holds many benefits. The lesson of Web 2.0 technologies is that developers and users can now apply new technologies and applications in surprisingly new ways. We are already seeing businesses using Semantic Web technologies in interesting and unexpected ways. As Spivack (2007) observes, Web 3.0 “will manifest in several ways. In many cases, it will improve applications and services we already use. So for example, we will see semantic social networks, semantic search, semantic groupware, semantic CMS, semantic CRM, semantic e-mail, and many other semantic versions of apps we use today.” We will see major advances in the personalization of Web applications and the use of smart software agents to help users manage the complexity of their digital lives. In the search arena, search engines will get smarter; among other things, they will start to not only answer questions, but they will also accept commands. We will also see big improvements in integration and data and account portability between different Web applications. Web 3.0, if it emerges as it promises, does represent paradigm shift. It will usher new era in integrating and aggregating information. The way that information is found, data is analyzed, and Web applications are built is going to change radically because of these new technologies. Researchers and businesses should start investigating these technologies and figuring out how to best leverage them to their advantage. As we move on to embrace Web 3.0, we might encounter a new kind of security threat, known as semantic attack. Semantic attacks target the way we assign meaning to content and can become seri-

7

Web X.0

ous. For instance, falsifying input into a computer process can be much more devastating. Imagine the effects of, for instance: airplanes delayed or rerouted by the feeding of bad information into the air traffic control system; process control computers being fooled by falsified inputs; or a successful semantic attack on a financial database. We need to develop and implement safeguards against semantic attacks in addition to what we currently do to protect against physical and syntactic attacks.

hArnessing the web To better harness the ongoing and future developments in the Web arena, we offer the following recommendations (adopted from Murugesan, 2006a): 1.

2.

3.

4.

8

Examine the promise and potential of Web 2.0, 3.0, and X.0 in applications areas of your interest. The key to success in this hypercompetitive global environment is your ability to recognize new technologies for their potential and impact. Create a winning case for the application you have in mind and make an informed decision. You may need to make some strategic changes to what you currently do and how you do it, and you may also need to engage your customers and employees as we enter into a user-driven, participative, new world of opportunities. To derive better benefits, look beyond a rich user interface into more substantive ways of engaging users and integrating and aggregating data from different sources. You need to look at your application and the world in provocative new ways. Lay a roadmap with a big picture in the foreground. You should carefully choose appropriate Web technologies, tools, and strategies.

5.

Proceed incrementally and steadily in your journey to harness the different incarnation of the Web.

iMPliCAtions for it Web’s evolution has a significant impact on business computing by enabling better, faster, richer applications while reducing costs and offering tangible and measurable ROI. The emergence of “situational applications” is likely to impact IT services in organizations. By leveraging heterogeneous data and content, as well as the collective intelligence via mashup tools, business users, who traditionally have to rely on enterprise IT teams, now have more power at hand than ever. We are also now getting into development of what some call disposable solutions, built to use once in an emergent, adaptive fashion and thrown away when finished. These applications make use of AJAX, Flash, lightweight programming models, wikis, mashable assets, APIs, and feeds. New applications create a new design and development dilemma: fast and easy versus well designed and well engineered. We now have tools built to bring applications together very quickly in contrast with traditional development platforms. We need to rethink Web application development methods in light of Web X.0, as well as catalog design patterns. Addressing the issues of scalability, performance, and security of new generation applications is a key challenge to researchers and IT professionals. We need significant improvements in these areas.

ConClUsion It is the innovation of the researchers, the developers, developers, and the applications -- not just the technology -- that will drive new generation Web applications to new heights. As regards their

Web X.0

widespread adoption, once the prospects and value of new generation Web are realized, people and enterprises will see they cannot afford to live without it. The Web is a fertile area for research and development, as new generation Web pose new technical, business, and social problems which need to tackled comprehensively and holistically successfully addressing the trade-offs and limitations.

referenCes Hinchcliffe, D. (2009). The evolving Web in 2009: Web squared emerges to refine Web 2.0. Web 2.0 blog. Retrieved on June 26, 2009, from http:// web2.socialcomputingjournal.com/the_evolving_web_in_2009_web_squared_emerges_as_ web_20_mai.htm Lin, K.-J. (2007, March). Building Web 2.0. Computer, 101–102. doi:10.1109/MC.2007.159 Metz, C. (2007, March 14). Web 3.0. PC Magazine. (www.pcmag.com/article2/0,2704,2102852,00. asp). Mitra, S. (2007, February 14). Web 3.0 = (4C + P + VS). Sramana Mitra blog. (http://sramanamitra. com/2007/02/14/web-30-4c-p-vs). Murugesan, S. (2007a). Business uses of Web 2.0: Potential and prospects. Cutter Consortium Business-IT Strategies Executive Report, 10(1). Murugesan, S. (2007b). Understanding Web 2.0. IT Professional. Retrieved from http://www.computer.org/portal/web/buildyourcareer/fa009 Murugesan, S. (2007c). Get ready to embrace Web 3.0. Cutter Consortium Business Intelligence Executive Report, 7(8). Spivack, N. (2007, September 24). Gartner is wrong about Web 3.0. Minding the Planet blog. (http://novaspivack.typepad.com/nova_spivacks_weblog/2007/09/gartner-is-wron.html).

AdditionAl reAding Dion Hinchcliffe’s Enterprise Web2.0 (http:// blogs.zdnet.com/Hinchcliffe) reviews Web 2.0’s progress and explores Web 2.0’s enterprise applications. Mashable (www.mashable.com) presents research into social networks, particularly widgets and other social networking add-ons. ProgrammableWeb. (www.programmableweb. com) presents the latest mashups, and new and interesting developments in Web 2.0 APIs and in the Web as a platform. It includes a blog and three dashboards—home, mashups, and APIs—which are updated daily.

Key terMs And definitions Mashups: A Web mashup is a Web page or Web site that combines information and services from multiple sources on the Web. Similar to music mashups, where artists combine, for example, vocals from one song with the music from another, Web mashups combine information and/or complementary functionality from multiple Web sites or Web applications. A Web mashup server lets you connect, collect, and mash up anything on the Web as well as data on some backend systems. HousingMaps (http://www.housingmaps.com) is a typical mashup application. It pulls sales and rental information from the classified advertisement Web site Craigslist (http://www.craigslist. com) and displays the listings on interactive maps pulled from Google Maps. Users can drag the map to see what is available for sale or rent in a given region. Really Simple Syndication (RSS): It is a family of Web feed formats used for syndicating content from blogs or Web pages. RSS is an XML file that summarizes information items and links to the information sources. It informs users of updates to blogs or Web sites they’re interested in.

9

Web X.0

Web or blog RSS feeds are typically linked with the word “subscribe,” an orange rectangle, or with the letters XML or RSS in an orange box. Social Network: A virtual place where people create their own space on which they write blogs, post pictures, videos, or music, share ideas, and link to other locations they find interesting, and open up this space for access by their friends and their friends’ friends. Social networks are places to network with like-minded people and businesses. They are powerful and very popular medium for human communication and interaction. They have, indeed, become the one-stop forum for sharing information on anything and everything in a variety of formats. The power and influence of online social networks are truly remarkable. Enterprises, marketers, politicians, and application developers are harnessing this medium in ways that were unimaginable just a few years ago. Virtual World: A virtual world is a Web-based 3D interactive environment much richer than the traditional Web and looks like a “place” -- a real place or a fanciful one. Most are designed to be created or populated by their users. Users are represented by their avatars, and avatars can navigate and move around the world and communicate with other avatars by text or by voice. A virtual world is also a platform for socializing and community-building. Some virtual worlds, like the real world, have their own functional economy -- a money market for in-world virtual goods and services. Thus, a virtual world, as its name implies, is a world of its own in cyberspace. Virtual worlds have emerged as an online 3D space for a wide range of activities, including gaming, social networking, education and training, marketing, e-business, and so on. Web 1.0: The traditional Web is now called Web 1.0. It is primarily one-way publishing medium. It supports online transactions and offers only minimal users interaction. It is also called read-only Web. Web 2.0: It represents the second phase in the evolution of the Web, and it’s about harnessing the potential of the Web in a more interactive and

10

collaborative manner with an emphasis on social interaction. It is both a new usage paradigm and a new technology paradigm. It is also a collection of technologies, business strategies, and social trends. As an umbrella term, it encompasses technologies such as AJAX, Ruby, blogs, wikis, mashups, tagging, and social bookmarking, as well as Web feed standards such as RSS and Atom. As an application deployment platform, it makes use of APIs and Web services. Web 3.0: It represents the third phase in the evolution of the Web. It, among other things, supports a machine-facilitated understanding of information on the Web. Web 3.0 is a Semantic Web, a 3D Web, a pervasive Web, a large database presented as Web pages, or a combination of these. Web 3.0 is aimed at addressing the needs of a user in context by rolling up elements such as content, context, community, commerce, vertical or contextual search, and personalization. Web 4.0: This represents the forth phase in Web’s evolution. The objective of Web 4.0 is to add it further sophistication and higher levels of intelligence. Your software agent(s) roaming on the Internet or simply residing on your computer could reason and communicate with other such agents and systems and work collaboratively to accomplish things on your behalf. It is also known as “intelligent Web” or “smart Web.” Web Squared: It refers to the notion of using Web to address real-world problems. In 2009, Tim O’Reilly and John Battelle, coined this term in order to promote the idea that if we are going to solve the world’s most pressing problems, we must put the power of the Web to work—its technologies, its business models, and perhaps most importantly, its philosophies of openness, collective intelligence, and transparency. They said,” It’s time for the Web to engage the real world. Web meets World—that’s Web Squared.” Web X.0: It is a generic word to represent the Xth phase in the evolution of the Web. Wiki: A wiki is a simple yet powerful Webbased collaborative- authoring (or content-

Web X.0

management) system for creating and editing content. It lets anyone add a new article or revise an existing article through a Web browser. Users can also track changes made to an article. The term wiki is derived from the Hawaiian word wikiwiki,

which means fast or quick. The user-generated online encyclopedia Wikipedia is a wiki. Wiki offers an elegant collaboration platform for collaborative authoring, project management, new product development, and more.

11

12

Chapter 2

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement: The Web Before and Beyond 2.0 Sebastian Weber Fraunhofer Institute for Experimental Software Engineering (IESE), Germany Jörg Rech Fraunhofer Institute for Experimental Software Engineering (IESE), Germany

AbstrACt Web 2.0 is a popular term used to describe a class of Web applications that offers mostly free services to its users. However, an exact deinition of the concepts, features, and technologies that argue for a Web 2.0 service is still missing. Similarly, terms such as Web 3.0, Web 4.0, or Web 2.5 also have no clear and unambiguous deinitions. This chapter reports the results of a Web and literature survey about Web X.Y concepts. Based on several deinitions, we synthesized new deinitions for Web X.Y, which provide an overview and can be used for differentiation, and we classiied contemporary Web services (e.g., Flickr) according to these deinitions.

introdUCtion The World Wide Web (WWW) has been through many changes since its beginnings and has become the largest information platform worldwide. When Tim Berners-Lee published his ideas for hypertext in 1989, he could not have guessed how he would change our lives. Due to technical progress made since then, its use has become more and more intuitive and users can provide their own content DOI: 10.4018/978-1-60566-384-5.ch002

for public use more and more easily. Similarly, when O’Reilly Media coined the term “Web 2.0” in 2004, they combined a set of concepts under one notion. In addition, version numbers can be used to differentiate evolutionary steps of the Web, as it is common practice with software systems. The term “Web 2.0” – and it seems that the same will happen with “Web 3.0” – has often been abused as a marketing term over the years. Many people used it as a buzzword without knowing that it does not only constitute a particular technology, e.g., AJAX, but refers to other concepts and features.

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

Therefore, there exist a lot of different perceptions of Web 2.0 (or Web X.Y in general). This chapter aims at clarifying what Web 2.0 (Web X.Y) is, and what it is not. It goes into detail regarding the concepts (e.g., collaboration or mashups), features (e.g., tagging or microformats), technologies (e.g., AJAX or Flex), tools (e.g., Wikis or blogs) and services (e.g., Flickr1 or MySpace2) of Web 2.0. Based on a literature and Web survey, we present an overview of the evolution of the Web before and beyond it. We summarize existing Web X.Y definitions and derive new comprehensive definitions from these findings. However, the main focus lies on the classification of Web X.Y, including definitions with differentiating and common factors. In summary, this chapter provides a categorization of evolutionary Web steps that makes it possible to assign Web applications and services, as well as principles and concepts, to a particular Web step.

which concepts, definitions, technologies, and services are used.

research Method In order to systematically conduct the review, we roughly based the research method on the systematic literature review process synthesized by Kitchenham (2004). The following phases were conducted to realize this literature review. Besides identifying the need for a systematic literature review, the following steps were performed: •



design of the sUrvey • Today, the term Web 2.0 is omnipresent. In March 2008, Google Blog Search3 delivered over 10 million blog entries, Del.icio.us4 listed over 400,000 tagged bookmarks, and Amazon5 stocked over 1,700 related books. However, the ACM Digital Library6 returned only 337 scientific publications dealing with Web 2.0, which indicates that there exists only little research in this area. Furthermore, because many user groups have gotten in touch with Web 2.0 in many different ways, there exist many diverse perceptions of what Web 2.0 is all about. The disagreement is even greater regarding the meaning of Web 2.5, 3.0, 3.5, or 4.0. Thus, our main research objective was to identify the commonalities and variabilities of definitions for Web X.Y. Based on the available body of knowledge in the form of blog entries, scientific publications, and books, we elicited









Background research: Initial scoping survey to identify search terms for Web X.Y. While this is not a step deined by Kitchenham (2004), we performed it to retrieve as many search terms as possible within a short period of time (approx. 2 weeks). Review planning: Speciication of the research question, required data, and search terms, as well as identiication of search engines (i.e., data sources). Identiication of literature: Search for literature in the search engines and retrieval of titles, abstracts, and reference material. Selection of literature: Reading of literature abstracts, including (i.e., selecting) and excluding literature, and obtaining full-text versions of the selected literature. Analysis of the references in the obtained literature in order to identify further literature (i.e., repeating this phase with the new list of literature). Quality assessment: Reading the full papers or Web resources, evaluating their appropriateness, and identifying bias. Data extraction: Extraction of relevant data (e.g., deinitions, keywords, etc.) from the literature. Data synthesis: Structuring and systematization (descriptive / non-quantitative) of

13

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

the quality defects and quality defect diagnosis techniques found. The systematic literature review was conducted between October 2007 and March 2008 using the techniques described in the following subsections.

data sources and search terms In order to get a relatively objective picture of Web X.Y, we utilized many different data sources found via the search engines of Google7, ACM Digital Library, IEEE Xplore8, and Del.icio.us. Search terms used included “Web 0.5 / 1.0 / 1.5 / 2.0 / 2.5 / 3.0 / 3.5 / 4.0” in conjunction with “definition”, “concepts”, “examples”, etc. Furthermore, we used data sources concerning Web X.Y topics, such as Read/Write Web9, O’Reilly Radar10, TechCrunch11, or Mashable12. They all had valuable information, which represented the broad spectrum of people’s opinions about Web X.Y steps. In order to identify representative services for the particular Web X.Y steps, we utilized websites providing rankings of successful Web applications. We selected different websites with diverse criteria, such as the most famous services associated with a particular Web X.Y step (e.g., top 100 of best Web 2.0 applications in 2007), services compiled by a jury, or a list of successful websites in terms of traffic rankings13. In addition, we utilized Google Trends14, McKinsley (2007a; 2007b), and Gartner (2007a; 2007b) to infer trends beyond Web 2.0.

literature selection and literature Quality Assessment Our goal was to get an insight into which kinds of definitions exist for Web X.Y steps and which concepts and features constitute a particular step. Thus, we collected documents, images, and video, which included descriptions, examples, and

14

definitions about Web X.Y. Search requests were limited to English-language queries only. In most cases, our Del.icio.us bookmarks refer to Englishlanguage documents, too. Concept descriptions or definitions in scientific papers or books are regarded as sources with higher quality and thus priority in contrast to, for example, blog entries. In our own definitions, we considered sources with higher priority to have a higher value.

data extraction We extracted relevant passages (e.g., Web 2.0 concept or Web 3.0 definition) from the retrieved resources. After that, we aggregated the extracted information into the following topics: • •

• •

Web 0.5 / 1.0 / 1.5 / 2.0 / 2.5 / 3.0 / 3.5 / 4.0 deinitions Web 0.5 / 1.0 / 1.5 / 2.0 / 2.5 / 3.0 / 3.5 / 4.0 descriptions of concepts, technologies, tools, and services Comparisons of Web X.Y evolutionary steps (e.g., Web 1.0 vs. Web 2.0) Web X.Y services and companies (if possible, we attached the context (i.e., the concept of which the mentioned service is an example))

data synthesis Activities For every group of collected definitions, we extracted statements, respectively concepts, features, and technologies, and summarized them into defined terms (e.g., collective intelligence, social networking, or sharing), and therewith created lists of concepts ordered by occurrences. We utilized the extracted lists of concepts to infer our own definitions. In order to classify services to a specific Web X.Y step, we extracted a list of concepts that constitute a Web X.Y step and assigned them to the selected Web services. Of course, the degree of uncertainty increases after Web 2.0.

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

search result documentation Relevant resources were stored as Del.icio.us bookmarks and relevant passages within these findings, such as concept descriptions, definitions, or examples, were annotated and commented by us using the Web service Diigo15. We used the same vocabulary (i.e., tags) for both services. The aggregated groups of definitions and concepts are documented in a spreadsheet to extract our list of concepts for every group.

the web X.y Before O’Reilly Media coined the term Web 2.0 in 2004 and thereby created a new way of thinking about the Web, the Web had experienced continuous development. The initial idea of the Web arose in the early 1980s, but it was a long journey until the Web kicked off a revolution in information distribution in the early 1990s. In the following, we describe the evolutionary steps of the Web, from its beginnings, when Tim BernersLee developed the technological fundamentals (i.e., Web 0.5 – establishing the architecture of the Web), via the rise of the Web (Web 1.0) and its rapid commercialization (Web 1.5), up to the current Web (Web 2.0). While new stages of the Web supersede older ones, concepts and technologies from new stages do not completely replace older ones but co-exist with them (e.g., email, FTP, blogs, etc.). As an example, the Web 1.0 era was characterized by Web services whose content was created by the carrier and not by the users. While the content of Web 2.0 services mainly comes from their users, even in this Web 2.0 era, there still exist services that only let their users consume pre-built content (e.g., news services, such as BBC). In order to better understand the definitions of Web X.Y services presented in the follow-

ing sections, it might be helpful to first read the dimensions of the synthesized classification in the next part.

web 0.5 – the rise of tim berners-lee’s vision In the late 1980s and early 1990s, Tim BernersLee cleared the way for one of the biggest and most influential inventions of humanity – the World Wide Web (WWW or, in short, Web), which owes its name to Berners-Lee’s first homonymous browser called WorldWideWeb. Very early, he had a vision of a barrier-free Web, where machines of all types are connected to the Internet and a universal information space is established where everything is based on hypertext. The Web should become the central medium where people all over the world would be connected with each other and where data would always be up to date (Berners-Lee, 2000). In this era, the technical infrastructure with its fundamental technologies, such as HTML, URI, HTTP, Web server, and the concept of linking Web pages, were developed. During this phase, the Web emerged as a winner against competitive products, such as Gopher. Definition 1 Web 0.5 services are distributed and content-offering precursors to Web pages using non-standard technologies, protocols, and tools. Examples are systems such as Gopher, FTP, or Usenet.

web 1.0 – growth of the web: the first Mainstream websites Web 1.0 (1990 – 2000) was the phase during which the general public embraced the Web. It was also the time when standardization of the underlying technologies began, e.g., HTML or the HTTP protocol. This initial phase peaked from about 1993 until 1996, and represents an information space designed to help people all over the world

15

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

exchange information. However, this was a oneway publishing medium, because only website authors exclusively provided the content – the “read-only Web”. Definition 2 Web 1.0 services are presentation-oriented content viewing services based on technologies supporting static Web pages (mainly hard-coded HTML pages) without much interaction, used to display information. Typical examples were simple homepages or directory services, such as Altavista16, Yahoo17, or Netscape18, as well as basic supportive tools such as Web development tools (e.g., HTML editors) and basic search engines, such as AliWeb19.

web 1.5 – the web of experts People often label the time from about 1996 onwards as Web 1.5, with a dramatic growth in users gaining access to the Web. The Web as a platform experienced increasing commercialization when big Internet players, such as eBay20, Amazon, or Microsoft with its Internet Explorer browser, emerged. This time brought many technical revolutions, such as more dynamic Web pages created on the fly from an ever-changing database and content management systems (CMS). In contrast to Web 1.0, Web developers needed a lot more skills to create business websites – not only HTML, but also client-side scripting (e.g., JavaScript or Java Applets) and server-side programming (i.e., Common Gateway Interface (CGI)). Definition 3 Web 1.5 services are commerceoriented content-viewing services based on technologies supporting dynamic pages (e.g., DHTML) and form-based interaction that often had closed APIs and closed IDs for presenting company-generated content. Typical examples are Google, Amazon, or eBay, as well as basic supportive tools such as Content Management Systems or WYSIWYG Web development tools.

16

web 2.0 – the social web In 2004, O’Reilly Media first recognized that services such as Del.icio.us, Wikipedia21, or MySpace are representatives of a new Web era, which constitutes a shift away from a one-way medium towards a bidirectional read/write Web. In O’Reilly’s (2005) famous essay, they described Web 2.0 (2000 – 2010) as a new stage in the evolution of the Web. In the spirit of Web 2.0, Webbased applications make the most of the intrinsic advantages of the Internet as a platform. They get better as more people use them by capturing network effects; they enable collaborative work; they deliver rich user experiences via desktop-like interfaces; and they combine data from multiple sources into new services. The power of consumers is a lot stronger than in the time of Web 1.0, because the amount of Web users has exploded dramatically in the last ten years. This has opened new possibilities for users as well as for website operators. Blogs have replaced ordinary homepages and enable users to reach many people in an easy way. Social networking is a phenomenon especially with younger Web users. Facebook22, a popular social networking platform for students, is gaining about 100,000 new users every day, and 45% of registered users come back to the site every day (in March 200823). The most successful Web 2.0 services all have social networking capabilities. The distribution of Flex or AJAX suddenly has enabled Web developers to create desktop-like user interfaces. Public Web APIs are an important component of so-called mashups. Mashups combine data from different sources to create a new service with more value. An RSS feed is a syndication concept that enables people to keep up to date with websites without the need to explicitly visit these websites. Web desktops, such as Netvibes24 or Pageflakes25, bring back the original idea of Web portals and reduce information overload. Tagging and folksonomies are two major concepts of Web 2.0 that go hand in hand

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

and are integral parts of many community-based services. Tagging is a quick and easy technique that enables people to describe resources to be discovered by others. Del.icio.us, for example, enables its users to share and find bookmarks by providing description tags for the resources, whereas Flickr users tag photos and YouTube26 users tag videos. The definition of this umbrella term has suddenly enabled the Internet community to talk about the concepts and technologies of a new evolutionary Web stage. To some extent, Web 2.0 has become a buzzword, because it is used by people for everything that gets popular on the Web. Definition 4 Web 2.0 services are user-oriented, content-sharing (upload, edit, and download), social networking (personal data), or static mashup services based on technologies supporting dynamic micropages that harness collective intelligence. They may support an open API with closed data and closed ID in order to use the Web as a distributed file system (user-generated content) or collaboration system (networking effects). Typical examples are YouTube, Flickr, Digg, Del. icio.us, LinkedIn, or MySpace, as well as basic supportive tools, such as Wikis or blogs.

web 2.5 – the Mobile web Users of Web 2.5 (2005 – 2015) will be “alwayson”, carrying along their mobile devices connected to the Internet. Services such as Twitter27 indicate the way people in Web 2.5 use the Web. There is a shift away from “desktops” as unique Internet access towards an increased usage of mobile devices – off-site reading (e.g., with RSS feeds and Web desktops) and publishing (e.g., Twitter as a microblogging service or Diigo as a social annotation service) will be integral parts of Web 2.5. Although Web 3.0 will be the first Web stage to have semantic technologies as an integral part, first Semantic Web applications will exist already in Web 2.5, e.g., Twine28 or Freebase29. In Web

2.5, social networks will go beyond “ego surfing”. Semantic annotations will be a key concept of Web 2.5 social networks, with people describing themselves and their input so that they can connect automatically. Currently, many start-ups and research institutes are working on so-called social search engines that will go beyond keyword-only approaches and leverage semantic information within social networks (Breslin and Decker, 2007), e.g., PeerSpective30, Eurekster31, Yahoo! Answers32, Google Co-op33, or Wikia Search34. Yihong Ding (2008) describes the data portability dilemma as “the next great frontier for the Web”. A DataPortability35 workgroup has already been founded to address the problem of supporting the portability of user identities, photos, videos, and other forms of personal data of social networks. Blogger Luke Gedeon (2006) sees 3D Web as a technology for creating virtual worlds (Second Life36 was one of the first services of this kind), rather a feature of Web 2.5 than Web 3.0. Web 2.5 is the first stage in the evolution of the Web that may bring the Internet infrastructure to its boundaries. In 2007, the video sharing service YouTube consumed as much bandwidth as the entire Internet did in 2000 (Lohr, 2008). The New York Times wrote in February 2008 that a research firm projected that user demand for the Internet could outpace network capacity by 2011. However, this rather implies challenges in terms of the modernization of the infrastructure than causing an Internet blackout in the future (Lohr, 2008). Definition 5 Web 2.5 services will be (mobile) device-oriented, user-, link-, or time-sensitive, cross-site, content-moving, virtual-reality-based, or dynamic mashup services based on technologies supporting rich user interfaces and user-sensitive interfaces that might support an Open ID and Open Data in order to support RUE (Rich User Experiences) and personal data portability. Examples are Second Life, Diigo, or Yahoo pipes.

17

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

web 3.0 – the semantic web A common opinion is that Web 3.0 (2010 – 2020) is equivalent to the Semantic Web (Lassila and Hendler, 2007; Ayers, 2006; Hendler, 2008). Market analyst Mills Davis (2007) expects that semantic technologies will embrace all semantic techniques and open standards that can be applied on top of Web 2.0 (e.g., knowledge representation, basic reasoning, pattern detection, or ontologyand model-based inferencing). Intelligent agents will be working hand in hand with Web users to connect knowledge in real time using automated and semi-automated methods (first applications in early states exist, e.g., Twine or Freebase). According to Davis, further trends of Web 3.0 are intelligent user interfaces (that know about the user and are able to tailor system behavior and communication) and end-user development. Jim Hendler, professor of computer science, sees Web 3.0 as a combination of Web 2.0 technologies plus a subset of the Semantic Web (Borland, 2007). AdaptiveBlue founder Alex Iskold believes that the initial idea of the Semantic Web is not realizable: “Its ultimate goal is to deliver perfect answers, which are unattainable. It is technologically impractical to achieve” (Zaino, 2007). Maria Azua, vice-president of technology and innovation at IBM, shares the opinion that not all facets of the Semantic Web are mainstream capable because using the Semantic Web in its entirety involves massive effort (Borland, 2007). However, Eric Miller, MIT, believes that Web 3.0 will indeed harness semantic technologies, but will be a hybrid, spun from a number of technological threads (Borland, 2007). Blogger Steve Rubel (2008) expects that websites will be obsolete by 2012. According to him, the future are Web Services and not websites. Leading Web 2.0 players, such as Amazon, still continue to expand their offers of Web Services and APIs. The trend is towards Software as a Service (SaaS), where third-party users can leverage APIs free or for a fee. Alex Iskold (2007) has the

18

same vision of Web 3.0, where the old perception of protecting one’s own data at all costs is displaced by a new way of thinking that open data is a competitive advantage. For San Murugesan (2007), Web 3.0 will be an entry-level Semantic Web that will be visualized by virtual worlds, accessed through diverse devices. Furthermore, cross-site ID recognition and cross-site identification of information will be an integral part. According to him, Web 3.0 will make use of already matured Web 2.0 features, such as RSS, tagging, folksonomies, and widgets, but also of technologies evolving in Web 2.5 (e.g., micro-blogging, 3D Web, SaaS, and mashups). Nova Spivack (2006) defines key emerging technology trends for Web 3.0, such as ubiquitous connectivity (everybody is online – everywhere), network computing (e.g., distributed computing, SaaS), open technologies (e.g., open APIs, protocols, or open data), open identity (e.g., OpenID), and the intelligent Web (e.g., Semantic Web technologies, natural language searching37, or machine learning). Definition 6 Web 3.0 services will be contentoriented, semantic-based, context-sensitive services based on technologies supporting semantically enriched websites that might support portable IDs in order to use the Web as a database and an operating system. Examples are Eurekster, AskWiki, Twine, or Freebase.

web 3.5 – the Ubiquitous web Web 3.5 (2015 – 2025) is the transition towards the “Intelligent Web” many people expect as Web 4.0. Summarizing the thoughts of blogger Harshal Hayatnagarkar (2007), in Web 3.5, we will see fully pervasive services based on matured and embraced semantic techniques from Web 3.0. We expect key technologies of Web 3.0, such as 3D Web or semantic technologies, to be upgraded to the next level of sophistication. Advancements in Web 3.0 technologies will evolve within Web 3.5 and will be fully matured within Web 4.0. As

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

an example, we believe that established virtual worlds of Web 3.0 will evolve into more advanced 3D worlds, where people will use upcoming technologies, such as holograms (Kanaracus, 2008) or augmented reality, which will bring the virtual world (e.g., 3D-enhanced social networks) and the real world closer together. Jack Domme, Hitachi Data Systems’ chief operation officer, believes that we will have ubiquity of the Web. For example, today’s RFID technology will be part of nearly everything (e.g., every paper or device) around us and will enable the environment to become interactive (Kanaracus, 2008). Definition 7 Web 3.5 services will be fully pervasive, interactive, and autonomous agents considering the personal context based on advanced semantic technologies supporting reasoning and basic AI that might bring the virtual and the real world closer together. Examples might be 3D-enhanced virtual social networks, natural language services, or fully interactive real-life environments (e.g., RFID, ambient sensors).

web 4.0 – the intelligent web Since the Web has not even reached the third stage, the Web community can only speculate as to what we can expect from Web 4.0 (2020 – 2030). San Murugesan (2007) believes that in Web 4.0, sophisticated artificial intelligence technologies will come into play. Intelligent proactive agents will interact with each other and work hand in hand with the users within a Ubiquitous Web (Davis, 2007; Murugesan, 2007). According to Nils Müller, the line between human beings and devices will blur and even disappear (Kanaracus, 2008). In Web 4.0, we will not only be the ones that input information into a device (e.g., a computer in Web 1.0, a mobile phone in Web 2.0, our intelligent house connected to the Internet in Web 3.0); rather, we ourselves will be the information sources directly connected to the Internet. Chips implanted for restoring the sight of blind people

or sensors on the motor cortex of the brain for controlling a computer with thoughts (possibly an avatar within a 3D world) are scenarios that could be a reality in the future, since research is in progress (Kanaracus, 2008). Dean Kamen’s “Luke Arm” is an artificial arm (advanced prosthesis) that has the same capabilities as a normal human arm and has been ready for clinical trial tests since the beginning of 2008 (Adee, 2008). Such a technological revolution foretells what we can expect from Web 4.0 – the Web will pervade all parts of our lives. Ambient Assistant Living is one application where elderly people will wear artificial legs and arms equipped with sensors connected to the Internet that proactively interact with relatives and caregivers. If we believe Nova Spivack, WebOS is the next logical step from Web 3.0 (Farber, 2007). According to him, the Internet will become the planetary computer, where all IP-capable devices (e.g., computers, mobile phones, or implanted sensors) will compose one unit, i.e., one big parallel world. Definition 8 Web 4.0 services will be autonomous, proactive, content-exploring, self-learning, collaborative, and content-generating agents based on fully matured semantic and reasoning technologies as well as AI. They will support adaptive content presentation that will use the Web database via an intelligent agent. Examples might be services interacting with sensors and implants, natural language services, or virtual reality services.

ClAssifiCAtion of web X.y serviCes In order to classify services according to our synthesized definitions, we collected diverse sources containing rankings of Web applications and websites – for example, the 100 most popular Web 2.0 services in 2007 or the top 100 most

19

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

popular websites on the Web. We counted the number of times Web services were mentioned within these rankings and extracted a new list ordered by frequency. If the source containing the ranking provides tags, categories, or descriptions for websites, we utilized this information to collect the concepts that people (individuals, juries) associated with them. Finally, we derived a table of the most popular websites with their associated features. The set of Web X.Y features is a result of an intense analysis of definitions and descriptions regarding the meaning of Web X.Y. We

collected special keywords about concepts (e.g., collaboration or mashups), features (e.g., tagging or microformats), technologies (e.g., AJAX or Flex), tools (e.g., Wikis or blogs) and services (e.g., Flickr or MySpace). In order to systematize the characterization of a Web X.Y era, we clustered similar keywords and identified the six different dimensions “Frontend”, “Content”, “Backend”, “Ads”, “Services”, and “Search”, which differentiate services in the Web X.Y versions. In addition, we synthesized a list of concepts and features ordered by number of times mentioned in Web X.0 definitions that indicates

Figure 1. Most often mentioned features of Web X.0 stages

20

Web 1.5

Web 1.0

Reactive Search

Cross-Site Search

Reactive Services

Services

Advanced Search

Search

Search

Form-based Interaction

Frontend

Search

Dynamic Pages

Frontend



Strict Content Classification

Content











On-Site Commenting

Content































































● ●

Animated Ads

Closed API

Ads

Backend





























● ●















Off-Site Ads

Ads

















YouTube





Services











Wikipedia







Inactive Services

Boolean Search

Search

Strict/Preset Services

Insensitive Search

Search









Services

Plain Search

Off-Site Search

Insensitive Interface

Frontend

Search

No Interaction

Frontend

Search

Static Pages

Frontend









Plain Ads

Content Viewing

Ads





On-site Ads

BBC

eBay

Craigslist

Amazon

Content

Ads

Dimension

Table 1. Classification of Web X.Y services































Flickr





































● ●











Digg











Del.icio.us





















LinkedIn





































Technorati





























MySpace



























Netvibes





























StumbleUpon





















Buzzword























Freebase

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

21

22

Web 2.0

● ●

Content Mashup

Static Mashup

Immobile-distant Services

Services

Services







Services





On-Site Search



Syndication



Search

Content

Content



Search

Content Download

Tagging

Content

RIA

Content Editing

Content



Frontend

Content Upload

Backend



Device-Sensitive Interface

Open API

Backend



Frontend

Closed ID

Backend

Frontend

Closed Data

Ads



Dynamic Micropages

Multi-Media Ads

Content-Sensitive Ads

Ads

Table 1. continued













































































































































































































































































An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

Web 2.5

● ●

Dynamic Mashup

User-Sensitive Services

Time-Sensitive Services

Rule-based Services

Services

Services

Services

Services



Software Mashup

Services

Content-Sensitive Services

User-Sensitive Search

Search

Services

Device-Sensitive Search





Search



User-Sensitive Interface

Frontend



Off-Site Commenting

Content ●



Content Moving

Content







Open Data

Backend





Open ID

Backend ●

User-Sensitive Ads

Ads



Device-Sensitive Ads

Ads

Table 1. continued















An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

23

24

Web 3.0

Natural-Language Search

Location-Sensitive Services

Experience-based Services

Search

Services

Services

Location-Sensitive Interface

Frontend

Time-Sensitive Search

Content-Sensitive Interface

Frontend

Search

Portable ID

Backend

Location-Sensitive Search

Location-Sensitive Ads

Ads

Search

Time-Sensitive Ads

Ads

Table 1. continued





An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

what people see as the main characteristics of Web stages (see Figure 1). The described evaluation resulted in Table 1, where we classified several existing services using this classification system.

Frontend Presentation ◦ Static Pages refers to HTML-based Web pages that do not change and are only rarely changed to relect new information or news about the site. ◦ Dynamic Pages refers to Web pages that can dynamically change their content based on the selection of a menu item or tab (e.g., using Dynamic HTML). All information is stored in the page (i.e., only parts are shown) or the whole page is (re-)loaded on demand. ◦ Dynamic Micropages refers to Web pages that (almost continuously) reload information (or ads) based on a triggering event (e.g., time or the user) from the server. In addition, the server can trigger a server push (e.g., BlazeDS, Comet). In contrast to dynamic pages, only parts of the whole page are reloaded or exchanged (e.g., by using AJAX, JSON, or Flex HTTPService). Interaction ◦ No Interaction refers to pages that do not support any interaction besides providing content, providing metadata, commenting, or searching. ◦ Form-based Interaction refers to pages that allow interaction in the form of forms that are used to deposit contact or shipping data. ◦ RIA refers to “Rich Internet Applications”, which are perceived as true desktop applications but do

not need to be installed. They offer (almost) all the functionality of local desktop applications (e.g., drag & drop, menus, etc.). RIA services go beyond the classic website paradigm (e.g., Buzzword). Interface ◦ Preset (Insensitive) interfaces refer to frontends that are ixed and do not change in any case. ◦ User-Sensitive (personalized, adaptive) interfaces refer to frontends that automatically or manually adapt to the preferences of the user or a group of users. ◦ Device-Sensitive interfaces refer to frontends that adapt to the device the user uses to view the service and content (e.g., a mobile phone, iPhone, PDA, tablet PC, etc.). ◦ Location-Sensitive interfaces refer to frontends that are sensitive to the location in the physical world (e.g., using GPS). ◦ Content-Sensitive interfaces refer to frontends that are sensitive to the content they present (e.g., by getting darker if a dark movie is shown). ◦ Time-Sensitive interfaces refer to frontends that are sensitive to the time they are used, such as darker street maps at night in a navigational system.

Content Commenting ◦ On-Site commenting refers to comments or recommendations made by the users to the actual content. The comments are stored on the same website where the content is stored. ◦ Off-Site commenting refers to comments or recommendations made

25

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

by people to the actual content (and potentially to on-site comments). However, the comments are stored on another (probably independent) website (e.g., Diigo as social annotation service). Classiication ◦ Tagging refers to free classiications of content by people using their own words that best describe the content in their eyes. ◦ Strict Classiication refers to grouping content into predeined classes that cannot be changed by the user. Flow ◦ Upload refers to content that can be uploaded and stored on a server. ◦ Download refers to content that can be downloaded from a server by users other than the original author(s). ◦ Move refers to content that can be moved between servers (i.e., crosssite). Handling ◦ Viewing refers to content that can be also viewed on a server/website by users other than the original author(s). ◦ Editing refers to content (and not metadata) that can be viewed and edited on a server by users other than the original author(s).

ID (Identity) ◦ Closed ID refers to a personal account on a server/website (or partner websites of the same operator/organization (e.g., Yahoo Pipes and Flickr)) that identiies a person and is used to store personal data. ◦ Open ID refers to a personal account that can be used by different servers (i.e., a decentralized single sign-on service) in order to identify people, eliminating the need for multiple usernames across different websites. ◦ Portable ID (avatar) refers to a personal account that can be used at servers to unambiguously identify oneself (e.g., like an electronic passport), including the identiication of one’s role in a network of people (e.g., if they are using different nicknames or email addresses). Data ◦ Closed Data refers to content that is stored on one server and cannot be exported to other services. ◦ Open Data refers to (one’s contributed) content that can be transferred to another service (e.g., transferring one’s images from Flickr to another service), deleted, or otherwise changed.

Backend

Ads

API

Type ◦



26

Closed API refers to an application programming Web interface that cannot be used freely (maybe via a fee). Open API refers to an application programming Web interface that can be used freely by external (third) parties.

◦ ◦



Plain Text / Pictures refers to ads purely based on text or images. Animated refers to animated ads (e.g., animated GIFs, Flash, etc.) without other media (i.e., sound). Multi-Media refers to animated ads (e.g., videos, animated slides, or

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

Table 2. Collection of catchphrases and metaphors of Web X.0 steps Web 1.0

Web 2.0

Web 3.0

Web 4.0

Connects information (Davis, 2007)

Connects people (Davis, 2007)

Connects knowledge (Davis, 2007)

Connects intelligence (Davis, 2007)

Info-centric (Murugesan, 2007)

People-centric (Murugesan, 2007)

Machine-centric (Murugesan, 2007)

Agent-centric (Murugesan, 2007)

Social Web (MacManus, 2007a)

Intelligent Web (MacManus, 2007a; Spivack, 2007)

AI Web (MacManus, 2007a)

The Semantic Web (Farber, 2007)

The Web OS (Farber, 2007) Intelligent Web (Davis, 2007; Murugesan, 2007) Smart Web (Murugesan, 2007)

Allows individuals to create and share ideas (Krupp, 2007)

Allows groups to create and share ideas (Krupp, 2007)

Allows societies to create and share ideas (Krupp, 2007)

Is the singularity (Krupp, 2007)

Gives the Internet itself a brain (Richards, 2007) Interaction (Kiss, 2008)

Recommendation and personalization (Kiss, 2008)

The document Web (MacManus, 2007b)

The Data Web (MacManus, 2007b)

Back-end (Richards, 2007)

Front-end (Richards, 2007)

Back-end (Richards, 2007)

First time to show the value of standards (MacManus, 2007b)

Teaches us how liberating standards can be (MacManus, 2007b)

Reflects on what worked in Web 2.0 (MacManus, 2007b)

Centralized “them” (O’Brien, 2007)

Distributed “us” (O’Brien, 2007)

Decentralized “me” (O’Brien, 2007)

interactive feedback ads) including sound and other media.



Source ◦

On-Site refers to ads from the website operator. ◦ Off-Site / Mixed-in refers to ads that are dynamically mixed into a website. Sensitivity ◦ Content-Sensitive refers to ads that are sensitive to the content presented to the user. ◦ User-Sensitive (personalized) refers to ads sensitive to an individual person (e.g., using information on the person and his (search) history). ◦ Device-Sensitive refers to ads tailored to the characteristics of the device the content (and ad) is viewed on.



Front-end (Richards, 2007)

Location-Sensitive refers to ads that depend on the physical location in the real world where the content is presented (e.g., advertisements for food on the parking lot of a mart). Time-Sensitive refers to ads that are sensitive to the time they are presented.

Services Sensitivity ◦ Preset (Insensitive) refers to services that are insensitive to external events. ◦ Content-Sensitive refers to services that are sensitive to the content processed by the service.

27

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement









User-Sensitive (personalized) refers to services sensitive to the user (e.g., using information on the person and his (search) history). Device-Sensitive refers to services tailored to the characteristics of the device the content is processed on (or for, by the service). Location-Sensitive refers to services that depend on the physical location in the real world where the service is executed or for which the content is processed (e.g., Semapedia38). Time-Sensitive refers to services that are sensitive to the time they are executed (e.g., darkening the display of navigational devices at night).

Activity ◦ Inactive refers to services that produce their service or content independently of their users. ◦ Reactive refers to services that have to be triggered and guided by the user and react by executing their process. Mobility ◦ Immobile-Distant / Hosted refers to services that are executed on one server (e.g., Buzzword39). ◦ Mobile refers to services that are executed on a server but can change the server and distribute themselves. Intelligence ◦ No refers to services with hard-coded processes. ◦ Rule-based refers to services based on (static/predeined) rules or processes. ◦ Experience-based (learning, adaptive) refers to services that can learn and adapt their rules (e.g., Amazon’s personalized suggestions). ◦ Exploring (looking for new information, services, etc.) refers to services that learn and optimize their service

28

(i.e., goal) and explore their environment to further optimize their service (e.g., by exploring new websites and collecting new information (news) relevant to the user). Mashup ◦









No Mashup (one source) refers to services that use only one source of content. Content Mashup refers to data mashups mixing different data streams / blocks from different providers into a completely new service. Software Mashup refers to mashups using multiple services (reusable functionality) that are applied on a single stream / block of content. Static Mashup refers to mashups that are programmed by developers and cannot be changed easily. Dynamic Mashup refers to mashups that are developed using a mashup development tool and that can be changed by the end-user (or a technically experienced user).

Search Power / Complexity ◦ Plain Search refers to simple search based on indexed words. ◦ Boolean Search refers to a simple querying language for tailoring a speciic search. ◦ Advanced Search refers to more complex search forms that can exploit speciic types or formats of content preset by the user (e.g., “iletype” in Google Search). ◦ Natural Language Search refers to a search language that is based on natural language (e.g., “when was Wikipedia founded?”).

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

Activity ◦ Reactive refers to search services that are triggered by the user. ◦ Proactive refers to search services that are executed pro-actively. ◦ Syndication (e.g., RSS or Atom feeds) refers to search services that can be subscribed to and consist of machinereadable data. Location ◦ Off-Site refers to search services that index information on other websites (e.g., Google). ◦ Cross-Site refers to meta-search services that query multiple websites (or other search engines) (e.g., Clusty. com). ◦ On-Site refers to search services that index information on one’s own website. Sensitivity ◦ Preset (insensitive) refers to search services that are insensitive to external information. ◦ User-Sensitive (personalized) refers to search services personalized to the user (e.g., using information on the person and his (search) history). ◦ Device-Sensitive refers to search services tailored to the characteristics of the device the index is searched on (or for/ by the service). ◦ Location-Sensitive refers to search services that depend on the physical location in the real world where the service is executed or for which the content is processed. ◦ Time-Sensitive refers to search services that are sensitive to the time they are executed.

ConClUsion Chronologically, the terms Web 0.5, Web 1.0, and Web 1.5 originated after the term Web 2.0 was coined by O’Reilly in 2004. According to the Sapir-Whorf Hypothesis, defining a term for a set of concepts enables people to talk about them (Hardman and Pemberton, 2008). Consequently, the naming of Web X.Y steps enabled people to talk about a specific Web era, which subsumes concepts, features, patterns, and technologies. However, even with versioning numbers to differentiate groups of Web services, many people still use them as buzzwords without knowing the core commonalities of these services. This chapter described prospective trends and visions for the “Web X.Y”, such as Web 2.5, 3.0, or 4.0. We collected concepts (e.g., collaboration or mashups), features (e.g., tagging or microformats), technologies (e.g., AJAX or Flex), tools (e.g., Wikis or blogs), and services (e.g., Flickr or MySpace) of Web X.Y in order to develop a new classification system for Web services. In addition, this helped us to synthesize new, more precise definitions for Web X.Y. Figure 2 depicts the steps of the Web on a timeline. We extracted the time spans from the discovered references. However, the steps are very fuzzy and no exact start or end times can be given (illustrated by the gradients). However, while new stages of the Web supersede older ones, concepts and technologies from older stages still exist in the newer stages. The horizontal lines indicate concepts or features (e.g., OWL) and the corresponding Web step (e.g., Web 3.0). For example, OWL emerged in 2004 but is the first integral part of a later Web step (Web 3.0). Additionally, while Web 2.0 was defined in 2004, the summarization of services under this version of the Web started earlier. Wikipedia was founded in 2001, blogs were coined in 1997, and the first Wiki was developed in 1995. Similarly, if Web 3.0 represents the “Semantic Web“, its rise started

29

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

Figure 2. Web X.Y timeline

30

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

in 1999 with the vision of Tim Berners-Lee, or in 1993 with Gruber’s definition of ontologies in computer science. This reveals that assigning concepts or features to one particular Web step is not always clear, because virtually every concept needs a very long preliminary lead-time, or else the concept may have existed in the “real world” for a long time (e.g., the concept of social networks). As depicted, we assume that every decade consumes one full version number (e.g., Web 2.0 encompasses 2000-2009) (as specified by Radar Networks (Farber, 2007)) and half versions arise between the full versions (e.g., Web 1.5 arose in 1995-2005). Nevertheless, some sporadic services do occur before these time spans and announce the Web step to come (e.g., research on the Semantic Web started around 2001 and first services for Web 3.0 do exist). Table 2 lists a collection of metaphors for Web X.0 steps, which emphasize the theme of a particular Web step. The table also emphasizes that there exist diverse point of views on Web X.Y steps, especially for future steps. As with the “Intelligent Web”, which is assigned to Web 3.0 and Web 4.0, some people define evolutionary Web steps earlier than others. As one can see, these catchphrases all capture different aspects of a Web X.0 step. However, they focus on a more or less similar theme. Web 1.0 focuses mainly on information presentation and could be named “Information Web”, whereas Web 2.0 focuses mainly on the participation and collaboration of users and therefore could be named “Users’Web” or “Social Web”. Consequently, Web 3.0 will probably focus on semantic technologies to unlock the wealth of information and could be named the “Semantic Web”, whereas Web 4.0 will probably focus on agents, intelligent assistance, as well as smart, proactive, and learning services that might be circumscribed as the “Intelligent Web”. However, for Web 3.0 and Web 4.0, we can only imagine what will become reality.

referenCes Adee, S. (2008). Dean Kamen’s “Luke arm” prosthesis readies for clinical trials. Retrieved on April 18, 2008, from http://spectrum.ieee.org/ feb08/5957 Anderson, C. (2004). The Long Tail. WIRED magazine, Issue 12.10, Retrieved September 13, 2008, from http://www.wired.com/wired/ archive/12.10/tail.html Ayers, D. (2006). The shortest path to the future Web. IEEE Internet Computing, 10(6), 76–79. doi:10.1109/MIC.2006.137 BakerS. (2006). Web 3.0. http://www.businessweek.com/the_thread/blogspotting/archives/ 2006/10/web_30.html Bala (2007). Kick Starting Web 3.0. Retrieved September 13, 2008, from http://balaarjunan.wordpress.com/2007/08/09/kick-starting-web-30/ Berlind, D. (2007). Jason Calacanis, Nick Carr, Microsoft, Google, and the most prized currency of Web 3.0: Trust. Retrieved September 13, 2008, from http://blogs.zdnet.com/ Berlind/?p=814&tag=btxcsim Berners-Lee, T. (2000). Weaving the Web. Harper Business. Berners-Lee, T. (2007). Giant Global Graph. Retrieved September 13, 2008, from http://dig. csail.mit.edu/breadcrumbs/node/215 Borland, J. (2007). Technology review part II: A smarter Web. Retrieved on April 18, 2008, from http://www.technologyreview.com/ Infotech/18396/?a=f Borland, J. (2007). Technology Review Part I: A Smarter Web. Retrieved September 13, 2008, from http://www.technologyreview.com/ Infotech/18395/?a=f

31

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

Breslin, J., & Decker, S. (2007). The future of social networks on the Internet: The need for semantics. IEEE Internet Computing, 11(6), 86–90. doi:10.1109/MIC.2007.138

Farber, D. (2007). Metaweb: Creating the world’s searchable global brain. Retrieved September 13, 2008, from http://blogs.zdnet.com/ BTL/?p=4625&tag=btxcsim

Catone, J. (2008). Web 3.0: Is It About Personalization? Retrieved September 13, 2008, from http://www.readwriteweb.com/archives/ web_30_is_it_about_personalization.php

Farber, D. (2008). Web 3.0 belongs to those who control personal profile infrastructure. Retrieved September 13, 2008, from http://news.cnet. com/8301-13953_3-9907300-80.html?tag=nefd. top

Clinton, D. (2005). Web 2.0. Retrieved September 13, 2008, from http://blog.unto.net/work/ on-web-20/ CRM editors (2008). 50 Social Sites That Every Business Needs a Presence on. Retrieved September 13, 2008, from http://www.insidecrm.com/ features/50-social-sites-012808/ Davis, M. (2007). Semantic wave report 2008 (executive summary). Retrieved on April 18, 2008, from http://www.project10x.com/dispatch. php?task=exsum&promo=sw2008rw01 Ding, Y. (2007). Implicit Web: A brief introduction. Retrieved September 13, 2008, from http:// blogs.zdnet.com/web2explorer/?p=413 Ding, Y. (2008). Data portability: The next great frontier for the Web. Retrieved on April 18, 2008, from http://www.internetevolution.com/author. asp?section_id=542&doc_id=145677& Farber, D. (2007). From Semantic Web (3.0) to the WebOS (4.0). Retrieved on April 18, 2008, from http:// blogs.zdnet.com/BTL/?p=4499

Fawzi, M. (2006). Wikipedia 3.0: The End of Google? Retrieved September 13, 2008, from http://evolvingtrends.wordpress.com/2006/06/26/ wikipedia-30-the-end-of-google/ Ferrell, K. (2008). Web 3.0: What’s Next After What’s Next. Retrieved September 13, 2008, from http://www.bmighty.com/network/showArticle. jhtml?articleID=206800301 Frey, C. (2004). How Amazon.com survived, thrived and turned a profit. Retrieved September 13, 2008, from http://seattlepi.nwsource.com/ business/158315_amazon28.html Gartner. (2007). Gartner identifies the top 10 strategic technologies for 2008. Retrieved on January 14, 2008, from http://www.gartner.com/ it/page.jsp?id=530109 Gartner. (2007). Hype cycle for emerging technologies. Retrieved on January 14, 2008, from http://www.gartner.com/ DisplayDocument?id=509710

Ellis, S. (2008). The Future is Interactive, Not Online. Retrieved September 13, 2008, from http://thenewmarketing.com/blogs/steve_ellis/ archive/2008/03/17/5467.aspx

Gedeon, L. (2006). Is it really Web 3.0 or just Web 2.5? Retrieved on April 18, 2008, from http://luke.gedeon.name/is-it-really-web-30-orjust-25.html

Farber, D. (2006). Web 2.0 isn’t dead, but Web 3.0 is bubbling up. Retrieved September 13, 2008, from http://blogs.zdnet.com/BTL/?p=3934http:// edgeperspectives.typepad.com/edge_perspectives/2005/12/ready_for_web_3.html

Hagel, J. (2005). Ready for Web 3.0? Retrieved September 13, 2008, from http://edgeperspectives.typepad.com/edge_perspectives/2005/12/ ready_for_web_3.html

32

Hardman, L., & Pemberton, S. (2008). The path to Web n+1. ERCIM News, 72, 15–17.

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

Hayatnagarkar, H. (2007). Web 3.0-the road to El Dorado. Retrieved on April 18, 2008, from http://evolutioneer.wordpress.com/2007/08/04/ web-30-the-road-to-el-dorado

Kanaracus, C. (2008). Is the ‘4.0’ era is upon us? Retrieved on April 18, 2008, from http://www.infoworld.com/article/08/03/04/Is-the-4.0-era-uponus_1.html?source=NLC-TB&cgd=2008-03-04

Hendler, J. (2008). Web 3.0: Chicken farms on the Semantic Web. IEEE Computing, 41(1), 106–108.

Khare, R. (2006). Microformats: the next (small) thing on the semantic Web? Internet Computing, IEEE, 10(Issue 1), 68–75. doi:10.1109/ MIC.2006.13

Hinchcliffe, D. (2006). The State of Web 2.0. Retrieved September 13, 2008, from http://web2. socialcomputingmagazine.com/the_state_of_ web_20.htm Hinchcliffe, D. (2006). Web 2.0 definition updated and Enterprise 2.0 emerges. Retrieved September 13, 2008, from http://blogs.zdnet.com/ Hinchcliffe/?p=71 Hinchcliffe, D. (2007). A bumper crop of new mashup platforms. Retrieved September 13, 2008, from http://blogs.zdnet.com/Hinchcliffe/?p=159 Hinchcliffe, D. (2008). Mashups turn into an industry as offerings mature. Retrieved September 13, 2008, from http://blogs.zdnet.com/ Hinchcliffe/?p=174 Hinchcliffe, D. (2008). openid: The once and future enterprise Single Sign-On? Retrieved September 13, 2008, from http://blogs.zdnet.com/ Hinchcliffe/?p=159 Iskold, A. (2007). Web 3.0: When Web sites become Web services. Retrieved on April 18, 2008, from http://www.readwriteweb.com/archives/web_30_ when_web_sites_become_web_services.php Iskold, A. (2008). Semantic Web: What Is The Killer App? Retrieved September 13, 2008, from http://www.readwriteweb.com/archives/semantic_web_what_is_the_killer_app.php Jeffery, K. G. (2008). The Future of the Web: An ERCIM View. ERCIM News, no. 72, page 7

Kiss, J. (2008). Web 3.0 is all about rank and recommendation. Retrieved on April 18, 2008, from http://www.guardian.co.uk/media/2008/ feb/04/web20?gusrc=rss&feed=media Kitchenham, B. (2004). Procedures for undertaking systematic reviews. (Tech. Rep. TR/SE-0401). Keele: Department of Computer Science, Keele University and National ICT, Australia Ltd. Kolbitsch, J., & Maurer, H. (2006). The Transformation of the Web: How Emerging Communities Shape the Information we Consume. Journal of Universal Computer Science, 12(2), 187–213. Krupp, A. (2007). The four Webs: Web 2.0, digital identity, and the future of human interaction. Retrieved on April 10, 2008, from http://www. alexkrupp.com/fourwebs.html Lassila, O., & Hendler, J. (2007). Embracing Web 3.0. IEEE Internet Computing, 11(3), 90–93. doi:10.1109/MIC.2007.52 Lin, K.-J. (2007). Building Web 2.0. Computer, IEEE, 40(Issue 5), 101–102. Lohr, S. (2008). Video road hogs stir fear of Internet traffic jam. Retrieved on April 18, 2008, from http://www.nytimes.com/2008/03/13/ technology/13net.html?_r=1&adxnnl=1&oref= slogin&pagewanted=1&adxnnlx=120656227208fjAyVxoFirqFXCzn1dhw MacManus, R. (2007). Define Web 3.0 contestwinners of Web 2.0 expo tickets. Retrieved on April 18, 2008, from http://www.readwriteweb.com/ archives/define_web_30_contest_winners.php

33

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

MacManus, R. (2007). On Web 3.0. Retrieved on April 18, 2008, from http://www.readwriteweb. com/archives/on_web_30.php

Mitra, S. (2007). Web 3.0 = (4C + P + VS). Retrieved September 13, 2008, from http://sramanamitra.com/2007/02/14/web-30-4c-p-vs

MacManus, R. (2007). 10 Future Web Trends. Retrieved September 13, 2008, from http://www. readwriteweb.com/archives/10_future_web_ trends.php

Moffitt, S. (2007). Web 4.0 (and possibly 5.0)- You Knew This Was Coming. Retrieved September 13, 2008, from http://buzzcanuck.typepad.com/ agentwildfire/2007/01/web_40_and_poss.html

MacManus, R. (2007). 10 Semantic Apps to Watch. Retrieved from http://www.readwriteweb.com/ archives/10_semantic_apps_to_watch.php

Morrison, C. (2007). What is Web 3.0? It’s Web 2.0 with a brain. Retrieved September 13, 2008, from http://venturebeat.com/2007/10/21/what-isweb-30-its-web-20-with-a-brain/

Markoff, J. (2006). Entrepreneurs See a Web Guided by Common Sense. Retrieved September 13, 2008, from http://www.nytimes.com/2006/11/12/ business/12web.html McKinsey. (2007). Eight business technology trends to watch. Retrieved on April 18, 2008, from http://www.mckinseyquarterly.com/Information_Technology/Eight_business_technology_trends_to_watch_2080 McKinsey. (2007). How businesses are using Web 2.0: A McKinsey global survey. Retrieved on January 14, 2008, from http://www.mckinseyquarterly. com/article_abstract_visitor.aspx?ar=1913&l2=1 3&l3=11&srid=9&gp=1 Metz, C. (2007). Web 3.0. Retrieved September 13, 2008, from http://www.pcmag.com/ article2/0,2817,2102852,00.asp Mikroyannidis, A. (2007). Toward a Social Semantic Web. Computer, IEEE, 40(Issue 11), 113–115. Miller, P. (2005). Thinking about this Web 2.0 thing. Retrieved September 13, 2008, from http:// paulmiller.typepad.com/thinking_about_the_future/2005/08/thinking_about_.html Miller, P. (2007). Anthony Lilley doesn’t seem keen on Web 3.0. Retrieved September 13, 2008, from http://blogs.zdnet.com/ web2explorer/?p=399&tag=btxcsim

34

Murugesan, S. (2007). Get ready to embrace Web 3.0. Cutter Consortium. (Tech. Rep. 7(8)). Murugesan, S. (2007). Business Uses of Web 2.0: Potential and Prospects. Cutter Consortium report, Vol. 10, No. 1 Murugesan, S. (2007). Understanding Web 2.0. IT Professional, 9(Issue 4), 34–41. doi:10.1109/ MITP.2007.78 Naone, E. (2008). Technology Review Special Reports: Social Search. Retrieved September 13, 2008, from http://www.technologyreview.com/ Infotech/20138/ O’Brien, R. (2007). The next thing after 2.0! Retrieved on April 18, 2008, from http://www. outofrhythm.com/category/web30 O’Reilly, T. (2005). What is Web 2.0. Retrieved on April 18, 2008, from http://www.oreillynet. com/pub/a/oreilly/tim/news/2005/09/30/whatis-web-20.html Paczkowski, J. (2007). Web 3.Oh- God- WillThis- Silly- Versioning- Never- Stop?!! Retrieved September 13, 2008, from http://digitaldaily. allthingsd.com/20071004/web30/

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

Richards, J. (2007). Web 3.0 and beyond: The next 20 years of the Internet. Retrieved on April 18, 2008, from http://technology.timesonline.co.uk/ tol/news/tech_and_web/the_web/article2726190. ece

Spivack, N. (2007). Web 3.0 Roundup: Radar Networks, Powerset, Metaweb and Others... Retrieved September 13, 2008, from http://novaspivack. typepad.com/nova_spivacks_weblog/2007/02/ web_30_roundup_.html

Rubel, S. (2008). The future is Web services, not Web sites. Retrieved on April 18, 2008, from http://www.micropersuasion.com/2008/03/thefuture-is-w.html

Spivack, N. (2007). How the WebOS Evolves? Retrieved September 13, 2008, from http:// novaspivack.typepad.com/nova_spivacks_weblog/2007/02/steps_towards_a.html

Schaurer, B. (2005). What put the ‘2’ in Web 2.0? Retrieved September 13, 2008, from http:// adaptivepath.com/images/publications/essays/ What_puts_the_2_in_Web_20.pdf

Spivack, N. Minding the planet: The meaning and future of the Semantic Web. Retrieved September 13, 2008, from http://lifeboat.com/ex/minding. the.planet

Segaran, T. (2007). Programming Collective Intelligence: Building Smart Web 2.0 Applications. O’Reilly Media.

Staab, S.; Domingos, P.; Mike, P.; Golbeck, J.; Li Ding; Finin, T.; Joshi, A.; Nowak, A.; & Vallacher, R.R. (2005). Social networks applied. Intelligent Systems. IEEE, vol 20, Issue 1, 80-93

Simmons, J. (2007). Where Do You Stand on Web 3.0? Retrieved September 13, 2008, from http://www.semanticfocus.com/blog/entry/title/ where-do-you-stand-on-web-30

Wainewright, P. (2005). Web 3.0: The API-driven application. Retrieved September 13, 2008, from http://blogs.zdnet.com/SAAS/?p=72

Solis, B. (2007). Web 3.0 is About Intelligence. Retrieved September 13, 2008, from http://bub. blicio.us/web-30-is-about-intelligence/

Wainewright, P. (2005). What to expect from Web 3.0. Retrieved September 13, 2008, from http:// blogs.zdnet.com/SAAS/?p=68

Spalding, S. (2007). How to Define Web 3.0. Retrieved September 13, 2008, from http://howtosplitanatom.com/news/how-to-define-web-30-2/

Watt, S. (2007). Mashups – The evolution of the SOA, Part 1: Web 2.0 and foundational concepts. Retrieved September 13, 2008, from http://www128.ibm.com/developerworks/webservices/ library/ws-soa-mashups/

Spivack, N. (2006). The third-generation Web is coming. Retrieved on April 18, 2008, from http:// www.kurzweilai.net/meme/frame.html?main=/ articles/art0689.html Spivack, N. (2006). Does the Semantic Web = Web 3.0? Retrieved September 13, 2008, from http:// novaspivack.typepad.com/nova_spivacks_weblog/2006/11/does_the_semant.html Spivack, N. (2007). Diagram: Beyond Keyword (and Natural Language) Search. Retrieved September 13, 2008, from http://novaspivack. typepad.com/nova_spivacks_weblog/2007/03/ beyond_keyword_.html

Watt, S. (2007). Mashups – The evolution of the SOA, Part 2: Situational applications and the mashup ecosystem. Retrieved September 13, 2008, from http://www-128.ibm.com/developerworks/ webservices/library/ws-soa-mashups2/ Wong, J., & Hong, J. (2008). What do we “mashup” when we make mashups? In Proceedings of the 4th International Workshop on End-User Software Engineering (WEUSE), pp. 35-39, Leipzig, Germany, May 12, 2008.

35

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

Xu, L. (2007). Evolution of the World Wide Web. A historical view and analogical study. Retrieved September 13, 2008, from http://www.deg.byu. edu/ding/WebEvolution/evolution-prelude.html Zaino, J. (2007). The semantic curmudgeon. Retrieved on April 18, 2008, from http://www. semanticweb.com/article.php/3703201 Zang, N., Rosson, M., & Nasser, V. (2008). Mashups: who? what? why? CHI ‘08 Extended Abstracts on Human Factors in Computing Systems, Florence, Italy, April 05 - 10, 2008).

Key terMs And definitions Web 0.5: Web 0.5 services are distributed and content-offering precursors to Web pages using non-standard technologies, protocols, and tools. Examples are systems such as Gopher, FTP, or Usenet. Web 1.0: Web 1.0 services are presentationoriented content viewing services based on technologies supporting static Web pages (mainly hard-coded HTML pages) without much interaction, used to display information. Typical examples were simple homepages or directory services, such as Altavista, Yahoo, or Netscape, as well as basic supportive tools such as Web development tools (e.g., HTML editors) and basic search engines, such as AliWeb. Web 1.5: Web 1.5 services are commerce-oriented content-viewing services based on technologies supporting dynamic pages (e.g., DHTML) and form-based interaction that often had closed APIs and closed IDs for presenting company-generated content. Typical examples are Google, Amazon, or eBay, as well as basic supportive tools such as Content Management Systems or WYSIWYG Web development tools. Web 2.0: Web 2.0 services are user-oriented, content-sharing (upload, edit, and download), social networking (personal data), or static

36

mashup services based on technologies supporting dynamic micropages that harness collective intelligence. They may support an open API with closed data and closed ID in order to use the Web as a distributed file system (user-generated content) or collaboration system (net-working effects). Typical examples are YouTube, Flickr, Digg, Del. icio.us, LinkedIn, or MySpace, as well as basic supportive tools, such as Wikis or blogs. Web 2.5: Web 2.5 services will be (mobile) device-oriented, user-, link-, or time-sensitive, cross-site, content-moving, virtual-reality-based, or dynamic mashup services based on technologies supporting rich user interfaces and user-sensitive interfaces that might support an Open ID and Open Data in order to support RUE (Rich User Experiences) and personal data portability. Examples are Second Life, Diigo, or Yahoo pipes. Web 3.0: Web 3.0 services will be content-oriented, semantic-based, context-sensitive services based on technologies supporting semantically enriched websites that might support portable IDs in order to use the Web as a database and an operating system. Examples are Eurekster, AskWiki, Twine, or Freebase. Web 3.5: Web 3.5 services will be fully pervasive, interactive, and autonomous agents considering the personal context based on advanced semantic technologies supporting reasoning and basic AI that might bring the virtual and the real world closer together. Examples might be 3D-enhanced virtual social networks, natural-language services, or fully interactive real-life environments (e.g., RFID, ambient sensors). Web 4.0: Web 4.0 services will be autonomous, proactive, content-exploring, self-learning, collaborative, and content-generating agents based on fully matured semantic and reasoning technologies as well as AI. They will support adaptive content presentation that will use the Web database via an intelligent agent. Examples might be services interacting with sensors and implants, naturallanguage services, or virtual reality services.

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

Mashup: Mashups refer to an ad-hoc composition of content and services coming from different sources to create entirely new services that were not originally provided by any integrated source.

15

16 17 18

endnotes 1

2

3

4

5

6

7

8

9

10

11

12

13

14

Photo sharing service acquired by Yahoo (http://www.flickr.com). It is considered to be the first social network. In 2008, it is the best-known one worldwide (http://www.myspace.com). Specialized Google search for finding and searching in blogs on the Web (http://blogsearch.google.com) Social bookmarking service for storing and describing links online (http://del.icio.us) Largest online store worldwide that sells many products, e.g., books or clothes (http:// www.amazon.com) Full-text search of ACM journals and conference proceedings (http://portal.acm.org) In 2008, the world’s best-known search engine (http://www.google.com) Full-text access to publications of IEEE and IEE (http://ieeexplore.ieee.org) Blog that provides daily Web technology news, reviews, and analysis (http://www. readwriteweb.com) O’Reilly Media Blog that watches and reports on interesting technology news (http:// radar.oreilly.com) Blog that profiles and reviews Internet products and companies (http://www.techcrunch. com) Blog concerned with social networking news and applications (http://www.mashable. com) Traffic rankings of websites (http://www. alexa.com) Comparison of worldwide interest in a topic by means of search queries over time (http:// www.google.de/trends)

19 20

21

22 23

24

25 26

27

28

29

30

31

32

Social bookmarking and annotation service, which can be used as a research tool (http:// www.diigo.com) http://www.altavista.com http://www.yahoo.com In the mid-1990s, Netscape was a leading computer service company best known for its browser. It was acquired by AOL in 1998 (http://www.netscape.aol.com) http://www.aliweb.com Is an online auction service launched in 1995 (http://www.ebay.com) Wikipedia is a multilingual, Web-based, free content encyclopedia project (http://www. wikipedia.org) http://www.facebook.com http://www.facebook.com/press/info. php?statistics Netvibes is a multi-lingual, AJAX-based personalized start page much like Pageflakes (http://www.netvibes.com) http://www.pageflakes.com Video sharing service acquired by Google. Is one of the most used services in 2008 (http://www.youtube.com) Twitter is a service for people to communicate and stay connected through the exchange of quick, frequent answers to one simple question (http://www.twitter.com) Semantic Web application by Radar Networks with aspects of social networking, wikis, blogging, and knowledge management systems(http://www.twine.com) Developed by Metaweb, which describes Freebase as “an open shared database of the world’s knowledge” (http://metaweb.com) Integrating social networks and Web search (http://peerspective.mpi-sws.mpg.de) Eurekster provides “Swickis”, which are configurable search engines (http://www. eurekster.com) A community-driven website, which enables users to ask questions that are answered by other users (http://answers.yahoo.com)

37

An Overview and Differentiation of the Evolutionary Steps of the Web X.Y Movement

33

34

35

36

38

A platform that allows Web developers to build their custom search engine (http:// www.google.com/coop) Open-source search engine where the users determine the relevance (http://alpha.search. wikia.com) http://dataportability.onconfluence.com/display/dpmain/ DataPortability+Project+Charter Internet-based virtual world where users interact through avatars (http://www.secondlife.com)

37

38

39

AskWiki is a natural search engine in a very early stage that uses semantic technologies and seeks to provide specific answers to questions using information from Wikipedia articles (http://www.askwiki.com) Connects Wikipedia knowledge with relevant places in physical space (http://www. semapedia.org) Web-based word processor (http://www. buzzword.com)

Section 2

Web Modeling and Design

40

Chapter 3

A Model-Driven Engineering Approach for Deining Rich Internet Applications: A Web 2.0 Case Study Francisco Valverde Universidad Politécnica de Valencia, Spain Oscar Pastor Universidad Politécnica de Valencia, Spain Pedro Valderas Universidad Politécnica de Valencia, Spain Vicente Pelechano Universidad Politécnica de Valencia, Spain

AbstrACt Web 2.0 applications emphasize the end-user involvement to provide the content. In this new scenario, an easy to use and a highly interactive user interface (UI) is a key requirement in order to appeal the end-user. The main objective of this chapter is to introduce a model-driven engineering process to create rich Internet applications (RIA) that address the requirements that a Web 2.0 application must fulill. To achieve this goal, an interaction model made up of two complementary models is proposed: On the one hand, an abstract interaction model, which clearly deines the interactions between the user and the system and on the other hand, a concrete RIA interaction model that speciies the semantics needed to accurately deine RIA for the Web 2.0 domain. Both models are introduced inside a model-driven code generation process with the aim of producing a fully functional Web 2.0 application. To illustrate the contribution of this chapter, the approach is applied in a case study related to the Web 2.0 domain. DOI: 10.4018/978-1-60566-384-5.ch003

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

A Model-Driven Engineering Approach for Deining Rich Internet Applications

introdUCtion The Web has evolved to a platform where the enduser has a key role. A few years ago, the end-user was a passive consumer of information that was provided by the traditional sources (press, editors, etc.) or by people with technological knowledge about the Web. Nowadays, with the emerging popularity of blogs, wikis and other Social Web applications, regular users are able to create and share every kind of content. This new paradigm of applications which emphasizes the end-user involvement as the principal resource is called Social Web or Web 2.0. Therefore, from Web applications where users only retrieve information, the current Web requires rich interfaces that provide users with a more intuitive interaction experience. If the most representative Web 2.0 applications are analyzed, we can notice that they are supported by UIs more related to desktop interfaces than the HTML-based traditional ones. This new type of application (Duhl, 2003) architecture is called Rich Internet Application (RIA).With this new paradigm the border between a desktop application and a web one has begun to be blurry. Some good examples are the eBay Desktop (eBay, 2008) and the Google Earth (Google, 2008) applications. Both desktop applications provide the same information, functionality and interaction mechanisms as the corresponding Web ones. In fact in the next years, the users will access the Web from a widely array of mobile devices that must provide richer interfaces (Jones & Mardsen, 2005). Around this new application paradigm different technologies such as AJAX, REST Services or JavaScript UI frameworks have arise to support RIA development (Noda & Helwig, 2005). In addition, all these technologies are playing an important role when a Web 2.0 application is developed. However, as the number of the technologies involved in the development increases, the cost and the maintenance problems also increase. In the past years, Web Engineering methods have improved Web applications development by ap-

plying the Model-driven Engineering principles (Murugesan, 2008). These methods have provided promising results to enhance the development of the so-called by now “traditional” Web applications or “Web 1.0” applications. However, their conceptual models (Comai & Carughi, 2007) and methodologies (Preciado et al., 2005) lack the expressiveness needed to face the development of RIAs. Firstly, the interaction between users and the system is not described with the same detail that the system functionality and navigation. In a Web 2.0 application the user interaction is a critical requirement since the end-user contribution is essential. And secondly, there is not a clear distinction between the interaction, which describes the set of actions that the user can perform together with the information system, the interface, which constitutes the graphic elements that support this interaction (buttons, grids, multimedia components), and the aesthetics characteristics (such as layout, fonts, size, color etc.). Therefore, is obvious that in order to develop successfully Web 2.0 application from a Web Engineering perspective, the past methods must be adapted and/or extended. The HCI community has proposed several approaches to specify the interaction between the user and the system without taking into account the target platform. A common agreement is to define two abstraction levels in order to model the interaction: an abstract level to describe the interaction without taking into account technological issues and a concrete level to deal with platform concrete requirements. This approach is more flexible than the definition of a single Presentation Model proposed traditionally by the Web Engineering methods. The main research goal of the work presented in this chapter is to define a model-driven approach to produce RIA interfaces that satisfy the Web 2.0 interaction requirements. It is important to remark that this chapter deals with the interaction between the user and the information system, but not takes into account the social interaction that Web 2.0

41

A Model-Driven Engineering Approach for Deining Rich Internet Applications

applications imply. In any case, the modeling and implementation of those “richer” interactions between the user and the information system, constitute a mandatory step to evolve the current Web Engineering methods. To achieve this goal an Interaction Model made up of two models is proposed by following the HCI approach. These two models are: 1) An Abstract Interaction Model that defines the interactions between the user and the system and 2) A Concrete RIA Interaction Model that introduces the semantics needed to produce RIA interfaces. In order to define the above mentioned models, the Interaction Pattern concept is introduced at the conceptual level. Additionally, both models are integrated inside a model-driven method with capabilities of automatic generation of code. As result, the final output of this work must be a Model-driven Engineering approach to produce fully functional RIAs. To better illustrate the approach, a case study based on a Web 2.0 application has been selected. The rest of the chapter is structured as follows: Section 2 describes the background of this work. Section 3 introduces the case study and describes the Interaction Model proposed to model RIA interfaces. Section 4 discusses the future research directions related to the Web 2.0 model-driven development. Finally, the conclusions and the lessons learned are stated.

bACKgroUnd To clearly define the background of this chapter, first several Web Engineering methods, which have been proposed in the past years, are introduced. Additionally this section introduces the new works that have addressed how to extend the methods in order to support RIA development. Finally, the OOWS Web Engineering Method, which has been developed in our research group, is briefly introduced. The OOWS models are the starting point from which the new Interaction Model is defined.

42

related work In order to develop traditional “Web 1.0” applications, works from the HCI and the Web Engineering community have been proposed in the past years. On the one hand, several Web Engineering Methods such as WebML (Ceri et al., 2003), UWE (Koch, 2006) or OOHDM (Schawbe et al., 1996) among others (Rossi et. al., 2008) propose a Presentation Model to deal with the Web UI specification. These models have provided interesting results to define traditional HMTL-like interfaces where navigation and population retrieval are the main interactions. But in the RIA domain they have to be extended (Preciado et al., 2005) in order to provide the required expressivity. On the other hand, the HCI approaches such as USIXML (Vanderdonckt et al., 2004) or TERESA (Mori et al., 2004), provide abstract UI models richer than the Web Engineering ones. However, these models are too generic because their purpose is to manage any kind of platform or interaction modality (voice, visual etc.). In addition, in both cases it is unclear how the generated UI is linked to the system functionality. The proposal of this chapter is to combine the best practices from both HCI and Web Engineering fields in order to provide a Model-driven Engineering solution. Currently, Web applications are being improved by means of richer UIs in order to achieve a more intuitive interaction experience. This trend has become evident in the Web 2.0 domain that provides simple but very effective UIs. To achieve this degree of interactivity, the UI is not based in HTML but also in powerful technological frameworks. Hence the Business Logic layer is processed at the server side whereas the interface is processed at the client side. This new architecture paradigm, which is called Rich Internet Application (RIA), is not only related with the new underlying technology but it also has lead to new research discussions as the possibility of develop RIAs from a model-driven perspective. Since several Web 2.0 applications have being

A Model-Driven Engineering Approach for Deining Rich Internet Applications

developed using RIA technologies, to provide methods, techniques and tools to deal with these complex technologies is a key requirement. Several works in the Web Engineering field have proposed methodological extensions to face the RIA development. Firstly, Bozzon & Comai (2006) extend the WebML method to support the RIA specification. The proposed extension allows the definition of which data, operations, hypertext links and page content must be processed by the client instead of the server. Additionally, this work proposes a code generation process to obtain the final application using as target a RIA technology. In the context of the OOHDM method, Urbieta et al. (2007) propose and approach to designing RIA interfaces, considering the separation of the interface structure from the behavior. That work proposes an aspect-oriented perspective, to combine different concerns related to the interface composition. Hence the UI is defined as a composition of several interface atoms. That strategy resembles the interface design that several social Webs have applied in where end-users, customize their application view using interface components. In contrast, several works have addressed the RIA development from a HCI perspective. A metamodel for defining RIA UIs from an Abstract Interface Model is proposed in Ruiz et al. (2007). In that work a clear relationship is established between the abstract level and the concrete UI that represents the interaction. Another interesting approach, the RUX-model (Linaje et al., 2007) proposes how to define at the concrete level the time-related behaviors (Temporal Presentation) and the event-response actions (Interaction Presentation) to define RIA interface components. It is true that several interface components of a Web 2.0 application can be developed using the traditional “Web 1.0” methods. However, the common agreement in the works mentioned above is that the current Web Engineering methods must be extended at the conceptual level in order to deal with RIA development. Since there is a close

relationship between Web 2.0 applications and RIA, a Model-Driven Web Engineering method must consider this new technological architecture. Therefore, this chapter focuses on providing interaction models to support the RIA development in the context of the Web 2.0.

oo-Method and the oows web engineering Method In the context of our Research Center, several methods related to Model-driven Engineering of Information Systems have been developed in the past years. Since the UI is a key issue in the Information Systems development, some ideas which can also be applied to the RIA domain have been proposed in these methods. As a consequence, these works are the basis from where the proposed Interaction Model arises. OO-Method (Pastor & Molina, 2007) is an automatic code generation method that produces an equivalent software product from a system conceptual specification. OO-Method provides an UML-based - PIM using the MDA conceptswhere the static and dynamic aspects of a system are captured by means of four complementary models, which capture the data, business logic and interface requirements of an information system. Specifically, the OO-Method Presentation Model is based on the JUST-UI (Molina, 2002) pattern language to build UIs in terms of the Problem Space. This model defines the UI as a composition of Interaction Units that represent the main interactive operations to be performed. Additionally the Presentation Model provides Elementary Patterns that constrain and detail the behaviour of each Interaction Unit using elementary interaction operations. It is worth to remark that the OO-Method Presentation Model corresponds to an abstract representation of a UI without any details of the visual appearance. In order to extend OO-Method with the principles proposed by the Web Engineering community, OOWS (Fons et al., 2003) was defined.

43

A Model-Driven Engineering Approach for Deining Rich Internet Applications

OOWS is a Web Engineering method that provides methodological support for Web application development. This method has been developed as an extension of OO-Method to provide a better support for Web-related concepts. In order to achieve that goal, OOWS introduces three new models into the OO-Method conceptual schema in order to support the particular navigational and presentation aspects of a Web application in an adequate way. These models are: •





44

User Model: A User Diagram to specify the types of users that can interact with the system. The types of users are organized in a hierarchical way by means of inheritance relationships in order to specify navigation specializations. Navigational Model: This model deines the system navigational structure. It describes the navigation allowed for each type of user by means of a Navigational Map. This map is depicted by means of a directed graph whose nodes represent navigational contexts and whose arcs represent navigational links that deine the valid navigational paths over the system. Navigational contexts are made up of a set of Abstract Information Units (AIU), which represent the requirement of retrieving a chunk of related information. AIU are made up of navigational classes, which represent views over the classes deined in the Object Model (Class Diagram). These views are represented graphically as classes that are stereotyped with the «view» keyword and that contain the set of attributes and operations that will be available. Web Presentation Model: Thanks to this model, we are able to specify the visual properties of the information to be shown. The main difference with the OO-Method Presentation Model is that this model is focused on Web UIs. As a result it provides some primitives more suitable to the Web

domain. To achieve this goal, a set of presentation patterns is proposed to be applied over the Navigational Model. Some patterns that can be deined with this model are information access mechanisms (ilters, indexes and search views), information layout (register, tabular, master-detail, etc), order criteria (ascendant/descendent) or pagination cardinality. The OO-Method code generation process, which is implemented by the OLIVANOVA tool (CARE, 2008), must also be extended in order to automatically incorporate the code obtained from these OOWS models. In order to provide this extension, a parallel translation process which generates code from the OOWS models, have been defined using a model compiler made up of a set of model-to-code transformation rules. These translation processes (OO-Method and OOWS) are integrated at both the conceptual and the implementation level. The parallel translation process is supported by a tool that generates Webbased interfaces from the OOWS models. This tool also provides a visual editor in order to create the OOWS models. Further details about the OOWS tool, can be found in Valverde et al. (2007b). The Abstract Interaction Model proposed in this work is defined as an extended revision of both the OO-Method Presentation Model and the OOWS models. Since a concrete model is not defined, each IU or Elementary Pattern has always the same UI implementation. Although the UI implemented by default provides a solution for a wide array of scenarios, more flexibility is usually demanded by both analysts and customers. The approach presented here emphasizes the specification of interactive tasks and the use of two level of abstraction as the HCI community proposes (Calvary et al., 2003). As several patterns defined in the OO-Method PM and OOWS are useful in order to define a RIA interface, these patterns are taken into account to define the new Abstract Interaction Model.

A Model-Driven Engineering Approach for Deining Rich Internet Applications

Model-driven engineering of User interfACes for the web 2.0 The main contribution of this chapter is to introduce a Model-driven Engineering process for the development of RIA. Specifically, this development process is introduced using a Web 2.0 case study in order to illustrate the suitability of the proposed process in this domain. The Figure 1 illustrates the approach and the relationship between the models and the final code. The definition of Web 2.0 business logic is fairly similar to the traditional Web applications; therefore the chapter focus is placed on the user interface and interaction issues. The proposed approach assumes that the underlying functionality is represented by means of a conceptual model. In the context of our work, that functionality (Figure 1.up-right) has been defined using the previously introduced OO-Method models (See Section 2.2). Hence

the key issue to address is which models should be defined in order to cover the new interface interaction expressivity. This main contribution can be divided into three sub-goals: 1. To define the elements of the Interaction Model which represent the user-system interaction from an abstract perspective. 2. The introduction of a Concrete Interaction Model that captures the specific semantics detected in the RIA interfaces that supports Web 2.0 applications. 3. To discuss how both models could be introduced inside a code generation process to obtain the final RIA interface. In the next sections, these goals are described in detail using the case study that is introduced in the next section.

Figure 1. Model-driven engineering process overview

45

A Model-Driven Engineering Approach for Deining Rich Internet Applications

A social web Case study: Paper finder • To illustrate the approach proposed and to provide a real-world example of the concepts that are going to be introduced, a Web application case study is firstly described. The domain selected for the case study is the Social Web domain, which is strongly related to the Web 2.0 concept. Web applications from this domain emphasize the user involvement in the content creation; whereas in traditional “Web 1.0” applications the user is a passive information consumer, in the Web 2.0 domain the user is an active information producer. Furthermore, the user is who establishes relationships between the data by using semantic annotations that are called tags. These annotations provide an easy mechanism to create networks made up of related data. Since the final user has a key role in the application functionality, powerful UIs are provided to encourage the content creation. These UIs are usually developed using advanced UI technologies, thus there is a close relationship between RIAs and Social Web applications. The main goal of the proposed case study is to facilitate finding academic papers uploaded by different authors. A usable UI must be provided for easily finding academic papers. The case study is based on the CiteUlike Web site, which is focused on the bibliographic references management. The specific functional requirements that the Web application must fulfill are: •





46

The authors must register into the system introducing information about its university afiliation, research interests, etc. Registered authors must be able to introduce information about their academic papers: title, abstract, publication date, etc. The papers must be annotated using tags or keywords which describe the paper domain, the topic discussed or the research area. These tags must provide a mechanism



to create relationships between similar papers. Each paper can be reviewed by registered authors in terms of the contribution relevance, originality, or technical content. From this review and external indexes -such as the number of cites or the journal/ conference impact factor- a value (Paper Rank) is calculated to classify the paper importance. Different search mechanisms should be provided to ind the most valued papers in a particular research area, the most read paper, etc.

The main difference between this application and traditional paper repositories as DBLP, is that authors create the references, evaluate the contributions and establish the links between papers. Hence, a virtual community defines both the content and the most popular items. As a consequence, a more “democratic” mechanism is provided to rate and find interesting academic papers.

the Abstract interaction Model In the context of this work, interaction is defined as the actions that take place between a human user and an interface, which acts as the communication link to the software system functionality, in order to perform a particular task. Therefore, in the interaction process there are three main actors: the user, the software system and the interface between them. The aim of the Abstract Interaction Model is to describe the interaction between the user and the system not including technological issues related to the final implementation. This model must be related to a Concrete Interaction Model which describes the interface to carry on the interaction. Thus, the same interaction can be implemented by means of different UIs related. The proposed Abstract Interaction Model

A Model-Driven Engineering Approach for Deining Rich Internet Applications

is made up of two main conceptual elements: the Interaction Map and the Abstract Interaction Patterns (AIP).

The Interaction Map Each type of user who accesses an application has a set of tasks which can be performed. The Interaction Map is a directed graph associated to a particular user, whose nodes represent the interactive tasks available and whose arcs represent the transitions between tasks. For instance, the Figure 2 shows the Interaction Map for the “Author” user type. The tasks directly linked to the author (“Search Information” and “Publish Paper”) can be accessed by the user in any point of the interface. However to perform the “Review Paper” and the “Define Tag” tasks, firstly the “Search Paper” task must be accomplished. Transitions represented in the Figure 2 as arrows can be modified using conditions which must be satisfied to trigger the task transition. In the example, to perform the transition to the “Review Paper” task, first an object id, which represents a Paper, must be selected in the previous task. After that, this object will be used to determine which paper the user is going to review or to tag. More complex transition conditions can be defined using OCL-like formulas.

Abstract Interaction Patterns An Abstract Interaction Pattern (AIP) models a generic solution for a user-system interaction

without taking into account the underlying technology. The main objective of AIPs is to describe each task defined in the Interaction Map in detail. From the analysis of several UIs in different domains and technological platforms, we have observed that there are widely used interactions that can be described as patterns. It is important to note that AIPs are not patterns from a design (Tidwell, 2005) or an implementation (Gamma et al., 1995) point of view. Instead of defining concepts related to the Solution Space, the AIPs specify widely accepted interactions from the Problem Space perspective. This approach has two main advantages: 1.

2.

An AIP represents a generic solution. As a consequence, these patterns are not strictly related to a UI development method or technological platform. Therefore, the AIPs can be reused in other Model-driven approaches. The AIP concept is consistent with the current OO-Method Presentation Model, in which the Interaction Units and the Elementary Patterns can be redefined as AIPs. Furthermore some OOWS conceptual primitives, such as the access mechanisms, can be also fitted inside the AIP concept. For that reason, these patterns can be easily introduced in the RIA development process proposed in this chapter.

If we analyze the flow of information between the interface and the information system, two main

Figure 2. Interaction map for authors

47

A Model-Driven Engineering Approach for Deining Rich Internet Applications

AIPs can be defined: Population and Service. To illustrate the definition of these patterns Figure 3 shows an example using UML stereotyped elements. The Population AIP defines an interaction to accomplish a system information retrieval. In other words, this pattern can be defined as a query interaction to return a set of instances to the user. The Population AIP is modeled as a view over the model that represents the involved system data entities; for example an UML Class Diagram (see Figure 3 right). This view is made up of a Main Population and a set of its attributes that describes what information will be retrieved. For example, the “Search Information” interactive task is defined using a Population AIP over the class Paper (Figure 3. left) and the attributes we want to show to the user: the paper title, the abstract and the publication date. This information can be extended by means of several Complementary Populations, which have a structural relationship with the Main Population. For instance the set of Authors who have written the paper, or in other words, the instances from Author related to Paper through the “has written” relationship. Through the Complementary Population, only the instances related to the Main Population are shown. The Service AIP abstracts an operation (for example a class service) that modifies the state of Figure 3. Abstract interaction patterns example

48

the system objects. The Service AIP is defined as a view over the arguments of a service which is offered by the information system. For each argument of the service signature, an Input Argument AIP is created to insert the corresponding value. Therefore, the Service AIP abstracts two “basic” interactions: the input of the argument values and the execution of the service. The interactive task “Publish Paper” is a clear example for using this AIP. The system must provide a method “New” over the class Paper (Figure left) in order to execute that functionality. This pattern provides a mechanism to the user to fill in the arguments of that method (for example the paper title and so on) and send the functionality request to the system. In OO-Method, these main AIPs are defined using a Class Diagram because it represents the interface with the system functionality. However, since all the interaction behavior with a system cannot be defined using only the Population and the Service AIPs, Auxiliary IPs are introduce to constraint the behavior and/or to refine more accurately the interaction. For instance, a Population AIP defined over the class Author should be extended with a filter condition in order to retrieve the authors from a specific country. These patterns are named as Auxiliary AIPs, because they always must be related to a Population or to a Service

A Model-Driven Engineering Approach for Deining Rich Internet Applications

AIP. This set of Auxiliary AIPs is defined using the Elementary Patterns from JUST-UI and the access mechanisms from OOWS. Some examples of Auxiliary AIPs are: •





Filter: a ilter is always related to a Population AIP. By default, a Population AIP retrieves all the instances that compose the view. A ilter deines a well-formed formula that restricts the population to be retrieved; only the instances that satisfy the formula are shown to the user. Optionally, the user can introduce a value to complete a formula. In our case study, to return all the papers in the “Search Papers” task is not a good approach. By using a ilter deined over the attribute “Publish Year” from the class “Paper”, the user can introduce a year value to constraint the papers to show. Index: This pattern is deined over an attribute from a Population AIP showing to the user the attribute different values. When a user chooses a value from the index, only the instances that comply with the value selected are shown. In our case study, the user should be able to perform a search over the deined tags. For this purpose an index must be deined. Hence a list of different tags will be provided and the user will select which type of papers must be retrieved. Selection List: this pattern deines a set of values associated to an Input Argument from a Service AIP. Applying this pattern, the user can only choose one value from the provided list to ill the input. The set of values can be a static list of values or a dynamic list of values linked to a Population AIP. For instance, when the user wants to publish a paper, the paper must be related to an author. A Selection List over the Argument “author” in the Service “Publish”, will allow the user to select an author previously created.



Validation Rule: this pattern is related to an Input Argument from a Service AIP. It deines a rule based on a well-formed logic formula that must be satisied by the introduced value. If the value is not correct, an error message is shown to the user. This pattern is useful in the case study to validate whether the user has introduced a correct value (for instance to check if the publish date introduced meets the date format).

Currently eight auxiliary patterns which can be consulted in Valverde et al. (2007a) have been detected. These patterns define an abstract language pattern to model the interaction. From this Abstract Interaction Model a preliminary and functional UI can be generated. Figure shows a HTML-based interface obtained using the AIPs presented here and the development process introduced in Valverde et al. (2007b). Nevertheless, to take advantage of the rich capabilities that RIA frameworks provide, a Concrete Model is needed.

the Concrete riA interaction Model Each interactive task must have an UI that the user can use to perform it. Although the interaction with the system can be defined abstractly, the interfaces are usually highly coupled and/or constrained by the target technology. For instance, the UI to send an e-mail is clearly different in a Web mail application than in a mobile phone e-mail client. However the interaction with the system functionality “send an e-mail” is the same in both devices: to connect to a SMTP server, to log the e-mail account and to upload a text message. Therefore, it is an interesting approach to define separately the interaction specification and its interface. Web 2.0 applications have been clearly influenced by the new interactions introduced by RIA technologies. These new interactions cannot accurately be described with traditional approaches because they

49

A Model-Driven Engineering Approach for Deining Rich Internet Applications

Figure 4. UI generated from an abstract interaction model

are coupled to a particular technology. Therefore a less abstract modeling level must be introduced to deal with these new requirements. Several approaches mentioned above (section 2.1) address the problem introduced here. However these approaches have two main disadvantages: 1.

2.

50

Although it is clear that RIA have changed the traditional request-response Web application paradigm, these changes are mainly related to the technologies involved. Therefore, from a conceptual point of view, the fact that the behavior is defined either at the client side or at the server side is an issue which should be hidden to specify the interaction. The UI developed using RIA technologies are made up of widgets with a rich behavior. To define this kind of UIs in terms of generic or basic components (buttons, windows,

links etc.) is a difficult task due to there are several elements involved. Our approach proposes the use of RIA Interaction Patterns at the concrete level. Around the Web (Yahoo, 2008) and in the UI interface development literature (Tidwell, 2005), can be found several patterns to aid the RIA development. These patterns are usually been applied in real-world examples and in many cases, the pattern provides not only the solution description but also the implementation. In addition, several of these patterns have been applied in popular Web 2.0 applications. The main objective of our proposal is to abstract these patterns in terms of a Concrete Model that can be linked to the abstract one. To define the conceptual elements of this model, several RIA technologies (Noda & Helwig, 2005) and pattern languages have been analyzed in

A Model-Driven Engineering Approach for Deining Rich Internet Applications

order to abstract the concepts needed. It is worth saying that these patterns are usually described as Interaction Design Patterns (Borchers, 2000). Instead of using directly a design pattern, we use the concepts and the solution proposed to define Concrete Interaction Patterns (CIP) for RIA development. The use of Interaction Patterns at Concrete level has two main advantages: 1.

2.

This approach reuses the concepts and solutions which are being applied in industrial RIA development. The main goal is to create a Concrete Interaction Model which is based on concepts and solutions accepted by the Web development community. Since they are defined as generic solutions in the RIA domain, these patterns can be re-used to produce code for several technological platforms.

CIPs extend the traditional design pattern definition (Problem, Context, Solution and Scenario) to include: 1) Metamodel: it defines the concepts that the pattern abstracts using a metamodelling language. This representation must be used to create specific instances of the pattern. Additionally, it includes a textual description about the different entities, relationships and constraints defined in the metamodel. 2) Interaction semantics: it specifies precisely the interaction expected when the pattern is applied. Therefore, it describes the interface components, the interface events and the communication with the business logic. Additional models, such as Concur Task Trees (Paterno, 2004) or UML Activity Diagrams, are used to explain the semantics. In the approach presented, the AIPs describe the interface-system interaction behavior whereas the CIPs specify the user-RIA interface interaction. Therefore, to establish a link between both levels each CIP must reference one or more AIPs to define a complete interaction. In order to illustrate this approach two CIPs used in the case study are presented next. A simplified version

of the metamodel and the interaction semantics represented as a CTT are included to formalize the patterns. To our knowledge the complex behavior and the UI that these patterns represent, cannot be directly modeled using the traditional Web Engineering methods.

Quick Review Pattern Problem: The user wants to rate and add a little text review about an item (a book, a movie etc.) that has been retrieved in a Web page. Context: This pattern can be used to encourage user’s opinion in a Website because the user can add rate an item in a quick and simple way. Therefore ranking indexes according to user preferences can be easily calculated. Solution: Provide a link next to the object to display a little text box (about 500 characters), in which the review must be written, and a slider with the rating in a numeric scale (for instance from 1 to 10). The user must write the review, select the rate and send it. The set of reviews of the object are updated automatically with the new review. Figure 5 illustrates the UI expected. Scenario: This pattern is applied in the “Review Paper” task from the case study. When an “Author” user is reading the papers reviews, a quick review could be added. The user must introduce also a numeric value about the quality of the paper in order to calculate the “Paper Rank”. This review will be stored in the system together with the user information (user name and review date time). Metamodel: This pattern (Figure 6. Left) is linked to the Abstract Model through two entities: 1) a Class from the Object Model which defines which domain element (papers in our example) is associated with the reviews. This class is associated to a “Population AIP” in order to show the different items that can be reviewed and 2) a Service that is invoked to create the review with the rating that has been introduced by the user. To define the maximum length of the review the

51

A Model-Driven Engineering Approach for Deining Rich Internet Applications

Figure 5. UI for the Quick Review Pattern

Figure 6. Metamodel and CTT for the Quick Review Pattern

“Text Length” property can be specified. Additionally, the pattern has an additional Concrete entity named “Slider”. That entity represents the UI widget used by the end-user to input the review rating. This entity adds three properties to define the maximum and minimum value of the rating and the step value (in our example from 0 to 10, with a scale step of 0.5). Interaction Semantics: The interaction (Figure 6. Right) starts when the user selects an item to be reviewed. The identification of the item (for example, the paper title) is used by the “Retrieve Average Rating” system task to provide the current average rating. Next the user can rate the item using the slider (“Rate Item” task) or write the review (“Write Review” task). When both user tasks have been finished the new review is stored in the server while the current reviews list is updated at the client side.

52

Tag Cloud Pattern Problem: Users need a mechanism to browse the information by the most popular topics taking into account the content of the paper. Context: This pattern provides an easy way to classify items about similar topics. Since, this classification is defined by the end-user, the pattern should be applied when no strict taxonomy has been defined. Solution: A tag cloud is a list of keywords or tags where each tag is more or less highlighted (using a bigger font size or a darker color) according to the tag popularity. When the user selects a tag, only the items which were annotated with the selected tag are shown. Scenario: In our case study a Tag Cloud is useful in the “Search Papers” task. When a paper is uploaded to the system, users can define tags

A Model-Driven Engineering Approach for Deining Rich Internet Applications

to classify the papers and to improve the future searches. After that, these tags are used to define and index by the different papers keywords and ordered by their popularity Metamodel: A Tag Cloud is associated with two entities from the Abstract Model (See Figure 8 left). First, with a Population AIP that defines the population over which the pattern is applied, and secondly, an IndexAIP in order to represent the different tags to filter the population. Both AIPs are linked through an “Atributte” from a Class Diagram in which the tags are defined. The pattern includes four additional properties to refine the interaction: 1) Max Tags: specifies the maximum tags present in the index in order to avoid non-relevant ones 2) Show Instances Number: adds to the tag the number of population instances that are annotated with that tag, 3) and 4) Highlight By Size/Color: this properties define how the popularity of a tag is represented to the user by means of a bigger font size or a more intense color (In Figure 7 both properties are set to true).

Interaction Semantics: The two first system tasks of this pattern (See Figure 8 Right) are performed at the server side. Firstly the items defined by the Population AIP are retrieved and an Index AIP, which acts as the “Tag Cloud”, is built using a common attribute of the population. Next, the user can select a tag in order to filter the population items in the client side without retrieving them again. Those last two tasks can be performed several times.

from Models to the riA interface Code A goal of this chapter is to show how Model-driven Engineering is useful not only to model RIAs but also to produce the final UI. Around the Web 2.0 idea, there are several concepts and technologies that can be abstracted using conceptual models. Using the presented approach, the analyst can define a RIA without knowing the underlying details related to technology. In other words, the analyst defines the interaction related to UI and the model

Figure 7. UI for the Tag Cloud Pattern

Figure 8. Metamodel and CTT for the Tag Cloud Pattern

53

A Model-Driven Engineering Approach for Deining Rich Internet Applications

compiler generates the implementation using the more suitable technological concepts (AJAX calls, Javascript widgets, REST Services, etc.). For brevity reasons the code-generation process is briefly introduced in this chapter. Two main phases can be distinguished in this process 1) The specification of the Interaction Model (both Abstract and Concrete views) and 2) The Interaction Model transformation to the RIA interface. A previous step to the specification of the Interaction Model is to define the metamodels for both Abstract and Concrete Models. To achieve this task the Eclipse Modelling Framework (Budinsky et al., 2003) has been used. This framework provides the tool support needed for specifying the Interaction Models. The Interaction Model specification can be divided into four main steps: 1.

2.

3.

4.

The conceptual model that represents the system functionality must be created or imported as an UML Class Diagram to the Interaction Model tool. The different users of the application must be identified and associated with the allowed tasks. From this information, their Interaction Maps are defined. For each Interactive Task from an Interaction Map, the interaction is described in terms of AIPs. These patterns should reference classes, attributes, operations and relationships from the Class Diagram defined in the first step. When an Interactive Task has been described, for each AIP a Concrete RIA Interaction Pattern is selected and mapped. The mapping only could be defined using CIPs suitable to the interaction that the AIP provides. If no CIP is associated to an AIP, the model compiler must associate a default one.

It is worth to remark that in this approach the Abstract Interaction Model is not transformed into a Concrete Interface Model as several HCI approaches propose. The relationship between

54

both models, i.e. which Concrete RIA Interaction Pattern will implement an AIP, must be defined by the analyst but there is not a model-to-model transformation step. In other approaches the input to the final code generation phase is only the Concrete UI Model. The disadvantage of such an approach is the difficulty for defining a Concrete UI model which contains the expressivity of the Abstract UI Model. In our approach, the input to the model compiler is an Interaction Model which is comprised of the information from both levels of abstraction. As the Abstract Interaction Model is present in the generation process, the Concrete Interaction Model complements that model rather than replacing it. With the aim of defining a code generation process two main elements are needed: firstly a model-driven development framework as openArchitectureWare (OAW, 2008). This framework must be used to define a set of model-to-code rules that implements a RIA Interface Model Compiler. And secondly, a target technological platform (OpenLaszlo, Adobe Flex or a JavaScript Framework) that supports the patterns defined in the Concrete RIA Model. Previous successful experiences in the Desktop domain, with the industrial tool OLIVANOVA, and in the traditional Web development, with the OOWS Web Engineering method (Valverde, 2007b), have been used to apply the same principles to RIA Development. The Figure 1 (bottom) shows the resultant application architecture. Firstly, the Abstract Interaction Model is transformed into an Interaction Façade: a set of services (defined as REST Services) that abstracts the system functionality. Then, from the Concrete RIA Interaction an UI is generated according to the implementation defined in each Interaction Pattern. Finally, the Interaction Facade previously defined, acts as communication link between the functionality and the RIA interface.

A Model-Driven Engineering Approach for Deining Rich Internet Applications

fUtUre reseArCh direCtions Web 2.0 applications will gain more relevance in the next years. Therefore, new Web Engineering methods will appear to deal with this new paradigm. In this chapter, we have faced the Web 2.0 development from a technological point of view, providing a link between RIAs and the Social Web. RIA technologies are starting to become more usual in the Web Development and even, in Desktop development. There are clear indicators in the software industry that has started to emphasize the use of these new technologies. Currently, Web applications are evolving from HTML interfaces to more sophisticated ones, so that the line between a Desktop application and Web one will be blurry. However, Web 2.0 development has to take into account the social perspective too. The user relationships that appear in the Social Web are an interesting research topic, because they provide a great value to our applications. A future line of research is to detect which “Social Patterns” (Yahoo, 2008) are being successfully applied in Web 2.0 applications in order to include them in the development method. In this chapter the Interaction Model related to the proposed approach have been introduced. However, an environment to define the models and to produce the related code is needed. Further works will focus on the tool support for building the models presented as well as the model transformation rules. Previous tool development experiences to support Web Engineering methods can be applied to this new domain. Therefore, it is expected that several model-driven tools will be developed or extended to support the improved methods. These tools will aid to validate the different approaches by means of the modeling and implementation of several case studies. These case studies will provide an interesting feedback about the most suitable patterns to implement Web 2.0 applications. Finally, with the purpose of producing highquality Web 2.0 applications, the development

methods must include usability aspects. As future research, usability features must be included to guarantee that generated systems are quality systems. In order to define these usability requirements previous usability works (Panach et al., 2007) and standards as the w3C accessibility standard WAI-ARIA (Cooper et al., 2008) will be taken into account.

ConClUsion In the new Web 2.0 domain, a richer interface plays a key role to attract users. To improve the development of this type of application, a modeldriven approach to develop RIAs interfaces has been presented in this chapter. This approach introduces an Interaction Model made up of Interaction Patterns at two level of abstraction. The Interaction Pattern concept has been described using a Social Web case study. From the experiences using this approach three main lessons have been learned: 1.

2.

3.

The two level separation of the UI specification allows to the analyst to face better the interaction requirements expected in RIA. In the one hand, at Abstract level the interaction can be defined with modeling elements close to the Presentation Models proposed in the Web Engineering field. On the other hand, the Concrete Level extends the semantics for developing RIAs preserving the previously defined abstract interactions. The Concrete RIA Model has been defined using real world patterns applied in Web 2.0 applications. The main advantage is that the solution described by these patterns has been widely validated. Using these patterns, the analyst is dealing with concepts at the modeling level which can be easily related to the final UI. The use of a Model-driven Engineering approach abstracts the wide array of

55

A Model-Driven Engineering Approach for Deining Rich Internet Applications

technologies and architectural decisions involved in Web 2.0 development and specifically in RIA. The RIA domain can be classified as a very technological oriented one. The use of models is an interesting help to deal with the technological underlying complexity that the RIA development implies.

Calvary, G., Coutaz, J., Thevenin, D., Limbourg, Q., Bouillon, L., & Vanderdonckt, J. (2003). A unifying reference framework for multitarget user interfaces. Interacting with Computers, 15(3), 289–308. doi:10.1016/S0953-5438(03)00010-9

Once the Interaction Model has been validated after modeling several Web 2.0 applications, the final step of this research is to include this new model inside the OO-Method software generation process. As a consequence, it will be possible to automatically generate Web 2.0 applications that satisfy the interaction requirements expected.

Ceri, S., Fraternali, P., Bongio, A., Brambilla, M., Comai, S., & Matera, M. (2003). Designing dataintensive Web applications. Morgan Kaufmann.

ACKnowledgMent This work has been developed with the support of MEC under the project SESAMO TIN2007-62894 and the FPU grant AP2005-1590.

referenCes

CARE Technologies S. A. (2008). OLIVANOVA tool. Retrieved in September 2008, from www. care-t.com/products/index.asp

Comai, S., & Carughi, G. T. (2007). A behavioral model for rich Internet applications. Paper presented at the 7th International Conference in Web Engineering. Cooper, M., Schwerdtfeger, R., Seeman, L., & Pappasn, L. (2008). Accessible rich Internet applications (WAI-ARIA) version 1.0 [electronic version]. Retrieved in September 2008 from www. w3.org/TR/wai-aria/ Duhl, J. (2003). Rich Internet applications-IDC report [electronic version]. eBay Inc. (n.d.). EBay desktop application. Retrieved in September 2008, from desktop.ebay.com

Borchers, J. O. (2000). A pattern approach to interaction design. Paper presented at the ACM Conference on Designing Interactive SystemsDIS, New York.

Fons, J., Pelechano, V., Albert, M., & Pastor, O. (2003). Development of Web applications from Web enhanced conceptual schemas. Paper presented at the ER 2003.

Bozzon, A., & Comai, S. (2006). Conceptual modeling and code generation for rich Internet applications. Paper presented at the 6th International Conference on Web Engineering (ICWE).

Gamma, E., Helm, R., Johnson, R., & Vlissides, J. (1995). Design patterns: Elements of reusable object-oriented software. Addyson Wesley.

Budinsky, F., Merks, E., Steinberg, D., Ellersick, R., & Grose, T. J. (2003). Eclipse modeling framework. Addison-Wesley Professional.

Google. (n.d.). Google earth application. Retrieved in September 2008, from earth.google. com Jones, M., & Marsden, G. (2005). Mobile interaction design. John Wiley & Sons.

56

A Model-Driven Engineering Approach for Deining Rich Internet Applications

Koch, N. (2000). Software engineering for adaptive hypermedia applications. Unpublished doctoral dissertation, Ludwig-Maximilians-University of Munich, Munich.

Pastor, O., & Molina, J. C. (2007). Model-driven architecture in practice. A software production environment based on conceptual modelling. Springer.

Linaje, M., Preciado, J. C., & Sánchez-Figueroa, F. (2007). Engineering rich Internet application user interfaces over legacy Web models. IEEE Internet Computing, 53–59. doi:10.1109/MIC.2007.123

Paternò, F. (2004). ConcurTaskTrees: An engineered notation for task models. In D. Diaper, N. Stanton & N. A. Stanton (Eds.), The handbook of task analysis for human-computer interaction (pp. 483-501). London: Lawrence Erlbaum Associates.

Molina, P. J., Melia, S., & Pastor, O. (2002). JUST-UI: A user interface specification model. Paper presented at the Proceedings of Computer Aided Design of User Interfaces, CADUI’2002, Valenciennes, Francia. Mori, G., Paterno, F., & Santoro, C. (2004). Design and development of multidevice user interfaces through multiple logical descriptions. IEEE Transactions on Software Engineering. Murugesan, S. (2008). Web application development: Challenges and the role of Web engineering. In G. Rossi, O. Pastor, D. Schwabe & L. Olsina (Eds.), Web engineering: Modelling and implementing Web applications (pp. 7-32). Springer. Noda, T., & Helwig, S. (2005). Rich Internet applications-technical comparison and case studies of AJAX, Flash, and Java based RIA [electronic version]. Best Practice Reports University of Wisconsin-Madison. OAW. (2008). openArchitectureWare 4.3 framework. Retrieved in September 2008, from www. openarchitectureware.org/ Panach, J. I., Condori, N., Valverde, F., Aquino, N., & Pastor, O. (2007). Towards an early usability evaluation for Web applications. Paper presented at the International Conference on Software Process and Product Measurement-Mensura, Palma de Mallorca, Spain.

Preciado, J. C., Linaje, M., Sánchez, F., & Comai, S. (2005). Necessity of methodologies to model rich Internet applications. Paper presented at the 7th IEEE International Symposium on Web Site Evolution. Rossi, G., Pastor, O., Schwabe, D., & Olsina, L. (2008). Web application development: Challenges and the role of Web engineering. Springer. Ruiz, F. J. M. (2007). A development method for user interfaces of rich Internet applications. Diploma of Extended Studies in Management Sciences, Université Catholique de Louvain, Belgium. Schwabe, D., Rossi, G., & Barbosa, S. (1996). Systematic hypermedia design with OOHDM. Paper presented at the ACM Conference on Hypertext, Washington. Valverde, F., Panach, J. I., & Pastor, Ó. (2007a). An abstract interaction model for a MDA software production method. Paper presented at the 26th International Conference on Conceptual Modeling (ER 2007). Valverde, F., Valderas, P., Fons, J., & Pastor, O. (2007b). A MDA-based environment for Web applications development: From conceptual models to code. Paper presented at the 6th International Workshop on Web-Oriented Software Technologies (IWWOST), Como (Italy).

57

A Model-Driven Engineering Approach for Deining Rich Internet Applications

Vanderdonckt, J., Limbourg, Q., Michotte, B., Bouillon, L., Trevisan, D., & Florins, M. (2004). USIXML: a user interface description language for specifying multimodal user interfaces. Paper presented at the Proceedings of W3C Workshop on Multimodal Interaction WMI’2004, Sophia Antipolis, Greece. Tidwell, J. (2005). Designing interfaces. O’Reilly Media. Urbieta, M., Rossi, G., Ginzburg, J., & Schwabe, D. (2007). Designing the interface of rich Internet applications. Paper presented at the Fifth Latin American Web Congress (LA-WEB). Yahoo. (2008). Yahoo design pattern library. Retrieved in September 2008, from developer. yahoo.com/ypatterns/

Key terMs And definitions AJAX: Acronym of Asynchronous JavaScript and XML. Is a set of programming techniques applied in the client side of a Web browser in order to retrieve only the information which needs to be updated. CTT: Acronym of Concur Task Tree. A modeling notation to describe the different tasks involved in an interactive system and the relationships between them.

58

Interaction Model: A conceptual model that represents the communication between the user and the Information System by means of a user interface. Interaction Pattern: An Interaction Pattern describes a solution for a common user-system interaction in terms of the Problem Space by means of conceptual models. Model Driven Engineering: is a software development methodology in which models are the first artifact for describing, designing and implementing the final application REST Service: A Web Service invoked through the HTTP protocol in order to obtain the state representation (usually as a XML document) of an information resource. Rich Internet Application: A new paradigm of Web Application that transfers the processing of the user interaction to the client in order to produce richer UIs. Social Web: is a type of Web application which emphasizes the end-users involvement, the relationships between them and the shared interests of the community

59

Chapter 4

Modular and Systematic Interface Design for Rich Internet Applications Gustavo Rossi UNLP and Conicet, Argentina Matias Urbieta UNLP and Conicet, Argentina Jeronimo Ginzburg FCEyN, UBA, Argentina

AbstrACt In this chapter, we present a design approach for the interface of rich Internet applications, that is, those Web applications in which the conventional hypermedia paradigm has been improved with rich interaction styles. Our approach combines well-known techniques for advanced separation of concerns such as aspect-oriented software design, with the object oriented hypermedia design method (OOHDM) design model allowing to express in a high level way the structure and behaviours of the user interface as oblivious compositions of simpler interface atoms. Using simple illustrative examples we present the rationale of our approach, its core stages and the way it is integrated into the OOHDM. Some implementation issues are inally analyzed.

introdUCtion One of the key issues of the Web 2.0 is the emergence of new possibilities to improve the usability of Web software; in this chapter we focus on the interface of those Web applications which exhibit advanced interaction features. Designing the interface of these rich internet applications (RIAs from now on) is DOI: 10.4018/978-1-60566-384-5.ch004

difficult as they must cleverly combine hypermedialike interfaces as in “conventional” Web software (therefore using navigation as the main interaction style), with the interface functionality we find in desktop applications with drag and drop, information pop-up and other diverse interface effects. To make matters worse, these applications must also deal with a myriad of functional or non functional concerns which might be persistent or volatile (i.e. be active for short periods of time).

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Modular and Systematic Interface Design for Rich Internet Applications

RIAs evolve even faster than the “old” Web applications because designers are quickly learning how to improve the typical hypertext-like functionality with richer interface behaviours and new “patterns” of improvement arise everyday (“Ajax Patterns”, 2008). As a consequence this permanent “beta” state of RIA complicates things further: new interface widgets or interaction styles are constantly introduced, checked to assess users’ acceptance and further considered core components or eliminated. This fact can be faced with many different strategies which are just being evaluated in the community. The most relevant ones are the following: •





Design the new RIA from scratch and eventually make the “old” and the new applications co-exist. This is the case of Yahoo mail in which users can either use the conventional Web mail or switch to the new desktop-like mail. A variant of the previous alternative is to use the old design models and wrap it with RIA functionality by consuming the information of existing databases from the new interface. This is the approach proposed by the RUX model (Linaje, 2007). Incrementally improve the interface functionality with small design changes in such

Figure 1. a) a simple index, and b) a RIA-based index

60

a way that RIA functionality is seamlessly introduced without changing the overall application style. This strategy has been chosen for example by Amazon.com and recently formalized in (Rossi, 2006). Figure 1 shows an example of a product index in Amazon both in hypertext and in RIA styles. While the index in Figure 1 a follows the classical hypertext behaviour (scroll to find an item, click on the item and navigate to its page), the one in Figure 1 b has incorporated two classical RIA behaviours: first scrolling is emulated in a page area, and second it is possible to see the product information as a pop-up therefore improving the use of screen space. However, independently of the chosen approach, there is a need to specify the interface behaviour of the RIA in such a way that the stakeholders can easily understand the intended behaviour of each interface object and the impact it has on other objects. A good design approach should additionally recognize the need to support seamless evolution as discussed above. In this work we describe a novel approach for specifying the interface of RIA, which is based on an extension of the Abstract Data Views (ADV) design model (Cowan, 1995), used in the ObjectOriented Hypermedia Design Method (OOHDM).

Modular and Systematic Interface Design for Rich Internet Applications

After presenting the basic concepts of our approach, we show how interaction functionalities corresponding to different application concerns can be specified separately, and then put to work together either by using object-oriented composition or aspect-like weaving. We demonstrate how a wise separation of design concerns can help us to improve the application’s evolution. The chapter is structured as follows: We first characterize RIA and illustrate the typical RIA behaviours with a simple motivating example. Secondly we present related works in the same field. Then we introduce the OOHDM as our base framework emphasizing user interface design. We next present our approach and exemplify it showing the specification of recurrent interface patterns in RIA and conclude discussing some further work we are pursuing.

bACKgroUnd Characterization of riA interfaces Similar to the broader field of Web 2.0, it is not easy yet to precisely say what characterizes a RIA. However, there are many features which distinguish RIA from conventional Web applications such as: •





Rich interaction capabilities, which tend to mimic those already existing in desktop applications; as mentioned in the introduction some of them are pop-ups, interface fading, drag and drop of interface objects, modal windows and dialogs, keyboard shortcuts, etc. Complex client-side processing which moves some of the operations which were usually implemented in the server to “rich” clients. In some cases, even part of the business logic might reside in the client. Elimination of full page refreshing to provide navigation; in this way a link’s target

• •

can be seen in the same page which initiated navigation, allowing the implementation of sophisticated navigation behaviours such as transclusion (Kolbitsch, 2006; Nelson, 1995). Notice that these features change the usual Web navigation semantic, in which when we traverse a link, the original node is closed and the target is opened. Server to Client communication to implement automatic refresh of (part of) pages. Multimedia (videos, etc), geographic objects and processes (such as in Google Maps), animations, etc.

Though it is not an objective of this paper to discuss all these features, it is easy to see that most of them represent not only a breakthrough regarding the previously known Web style, but they also require further research to be incorporated in Web Engineering approaches. As an example, consider transclusion, the possibility of “inserting” the target of a link in the place of the anchor which triggered navigation. While transclusion was early defined in the hypertext community (Nelson, 1981), most Web design methods had ignored it so far, mainly because its implementation was not feasible1. We refer the reader to (Bozzon, 2006; Wright, 2008) for a complete characterization of RIA In the context of this chapter we will focus on those types of applications in which rich interface behaviours are implemented to improve their usability, such as e-commerce sites, advanced Web mail clients (like Gmail or Yahoo mail), internet radios (such as Pandora), etc. Though our approach can be used to implement most of the above mentioned features we will use examples of “conventional” Web applications which exhibit RIA interface features, to simplify the user’s task, improve his access to information, make navigation more dynamic, etc.

61

Modular and Systematic Interface Design for Rich Internet Applications

Motivating example Some E-commerce sites are evolving their Web sites from the conventional hypertext style to RIA. By introducing RIA features they are able to improve usability and therefore attract more potential buyers. RIA functionality allows showing relevant product’s data in novel ways. The well-known Amazon store is one example of an application which takes advantage of these features for presenting product information using videos, music sample, etc. It is also an interesting case study of incremental application improvement: RIA functionality is usually introduced seamlessly without breaking the whole application structure and look and feel. Even though in several occasions these features are later discarded, the overall structure stays stable. While this is clearly a business decision, its counterpart in software should assure that the design is also stable and not compromised each time there is a change. Our example is a narrow part of the Amazon site. In Figure 2 a, we show a watch with its picture, price, etc. When the mouse is over some small picture below (highlighted with an ellipsis in Figure 2 a) the main picture is updated with the small one maximizing it. Amazon aims at providing extra information by means of annotations on the picture. Each comment is shown inside of bubbles for highlighting some special watch features. In Figure 2 b when the mouse is over small annotated picture, the main

one is updated and it is also enriched with a mark over a specific picture coordinate (highlighting the back LCD) and its corresponding comment. With these simple RIA features, Amazon satisfies client curiosity or lack of knowledge about the product in a concise manner with just a quick look, avoiding the need to read long paragraphs of descriptions. We show later in this chapter how to specify this specific behaviour in such a way that the specification is modular, therefore admitting later variability. From a design point of view, this example comprises most interesting RIA features before mentioned. Additionally, these kind of extensions are usually a nightmare for architects and developers due to the fact that they don’t know if the new functionality will be part of the application core or it will just be available for a period of time (e.g. if users do not “accept” it). Later we will show how to specify the extension shown in Figure 2 b by using an incremental design improvement.

related work Even though the interface style of RIA is rather new, the need for methodological support for RIA has been already addressed (Preciado, 2005). An extension of the WebML approach to support RIA is presented in (Bozzon, 2006). In (Linaje, 2007) a complete model-based approach to build interactive interfaces for RIA has been presented; the authors mention that this approach has been

Figure 2. a) product information, and b) annotated picture

62

Modular and Systematic Interface Design for Rich Internet Applications

already implemented in the context of WebML though it is general enough to be “plugged” to other approaches. A component library concept is introduced in RuxModel (Preciado, 2008) which specifies components to be reused as patterns solutions. Although they are defined to run over different platforms, they only exist inside the RuxTool realm and its definition can’t be applied in non-RuxTool compliant platforms. In OOH4RIA (Meliá, 2008) the authors present a framework for designing RIA which uses OOH as design approach and Google Web Toolkit (GWT) as implementation technology. The approach uses concrete GWT widgets for designing the user interface; this eases the implementation process but does not allow reusing user interface designs, because the solution is tied to the underlying implementation technology. The work presented here is somewhat different; while being technology-independent, it is not yet completely elaborated as a model-driven approach, as we are just building tools to support mappings to implementation. Additionally we have addressed an important aspect of this kind of software: the need to support evolution when multiple (eventually crosscutting) concerns are present. By separating the interfaces corresponding to different concerns we are able to compose the specification of interface atoms in a transparent way. In the following section we contextualize our approach by briefly describing the OOHDM design method.

the oohdM design framework OOHDM, similarly to other Web development approaches such as UWA (UWA, 2002) OOWS (Pastor, 2001), UWE (Koch, 2001), or WebML (Ceri, 2000), organizes the process of development of a Web application into five activities: requirements gathering, conceptual (or content) design, navigational design, abstract interface (or presentation) design and implementation. Dur-

ing each activity a set of object-oriented models describing particular design concerns are built or enriched from previous iterations. We next describe conceptual and navigational aspects and in a separate sub-section interface design.

Conceptual and Navigational Aspects The first activity is intended to collect and analyze the stakeholders’ requirements. Use Cases (Jacobson, 1996) and User Interaction Diagrams (Vilain, 2000) can be used for this purpose. After gathering requirements, a set of guidelines is followed in order to derive the conceptual model of the application from the requirement specification. During this phase of conceptual design, a model of the application domain is built using well-known object-oriented modelling principles. The conceptual model should not reflect the fact that the application will be implemented in a Web environment, since the key Web functionality will be specified during navigational design. The navigational structure of the Web application is defined by a schema containing navigational classes. OOHDM offers a set of predefined types of navigational classes, i.e., nodes, links, anchors and access structures, many of which can be directly derived from conceptual relationships. By using a viewing mechanism, classes in the conceptual model are mapped to nodes in the navigational model while relationships are used to define links among nodes. Navigational Contexts (Schwabe, 1998) allow organizing the global navigational structure by describing sets of similar objects (products of some type, products recommended by a user, etc). In the context of RIA the conceptual and navigational models are built using the same heuristics. Therefore we will concentrate mainly on the details of user interface specification.

User Interface Design The last design phase previous to implementation is the abstract interface design. In this activity, the

63

Modular and Systematic Interface Design for Rich Internet Applications

user interface of the application is specified by means of Abstract Data Views (ADVs) (Cowan, 1995). ADVs are a formal model of interface objects; they allow specifying the structure and behaviour of nested interface objects and their relationships with other software components (in this case nodes and conceptual objects). In OOHDM we define an ADV for each node class, indicating how each node’s attribute or sub-node (if it is a composite node), will be perceived by the user. An ADV can be seen as an Observer (Gamma, 1995) of the node (which is call Abstract Data Object or ADO), expressing the node’s perception properties as nested ADVs or primitive types (e.g. buttons). ADVs can be classified either as “pure” interface objects, therefore acting as behavioural controllers of other objects (e.g. buttons which trigger applications behaviours), as interfaces of navigation objects thus providing interface support for navigation (e.g. showing information or anchors of a node), or as interfaces of application objects; in this latter case interface objects trigger application behaviours not directly related with navigation. A configuration diagram (Schwabe, 1998) is used to express how these properties relate with the node’s attributes. ADVs are also used to indicate how interaction will proceed and which interface effects take place as a result of user-generated

events. These behavioural aspects are specified using ADV-charts (Cowan, 1995). We next show how different aspects of a Web software interface are specified with ADVs.

Specifying Structural Aspects of Web Applications Interfaces An ADV has a structure (expressed with a set of attributes), behaviour (defined by the set of messages or external events it can handle) and can be recursively composed of other interface objects. Given their composite structure ADVs can be mapped in a rather straightforward way onto XML documents. In Figure 3 a we show the ADV corresponding to the ChangeablePicture component of the Web interface of Figure 2 a. This ADV is composed of other nested and primitive ADVs like Pictures or Text, showing how the component will be perceived by the user. In Figure 3 b the actual screen corresponding to the ADV is shown. Notice that the positions of nested objects in the ADV reflect the look and feel of the interface. ADVs “observe” ADOs (known as ADV “owners”) both as interface views and for triggering application or interface behaviours. The ADVs of Figure 3 get their contents from the corresponding ADO, in this case a node in the navigational model. As mentioned before, the

Figure 3. a) ADV for a product figure, and b) actual figure interface

64

Modular and Systematic Interface Design for Rich Internet Applications

relationships between the ADV and its ADO is described using configuration diagrams, a combination between UML classes and collaboration diagrams, showing which messages are interchanged between the ADV (acting as a client) and the ADO (in the role of a server). In Figure 4, the ADV ChangeablePicture gets information invoking the getImage() and getText() methods of Node ChangeablePicture. This information is used to render the concrete interface components: Image, Description and SmallImage (array). The “i” parameter references the actual selected image in the index. In “old” Web applications, we usually specify one nested ADV observing a single ADO. It might be the case that the same node has many different associated ADVs (for example to provide multiple interfaces), but it is not common that two nodes (ADOs) juxtapose in a single ADV. In RIA, we might need that one ADV consumes information from different ADOs (nodes) according to the interface characteristics described in Section 3. Notice that the behavioural aspects of a conventional Web application’s interface are fairly simple. When an anchor is selected the actual node’s ADV must be closed and the corresponding link’s target (another node) must be opened. Though there are other interesting but simple interface behaviours in Web applications, such as allowing form completion, we will directly explain the mechanisms to express ADVs’ behaviours

when we focus on the dynamics of RIA interfaces in the next Section. After the interface has been fully specified, the conceptual, navigation and interface models are mapped onto a particular runtime environment. In order to ease the adoption of the OOHDM approach, we have implemented a framework, named CAZON (Rossi, 2006) which supports the semi-automatic generation of code from OOHDM models, including the instantiation of Web pages from OOHDM navigational models.

expressing rich interface behaviours with Advs As previously discussed RIAs are characterized for having dynamic interfaces and rich behaviours which improve user experience; while using ADVs allow us to express the static features of a conventional application, we use ADV-charts to specify the dynamic aspects of this kind of applications. Following we present our solution to the problem with a brief explanation on ADV-charts, a set of examples and a discussion on how to tame evolution. Finally we describe a proposal to use this notation to specify RIA patterns, illustrating it with some well-known patterns.

Figure 4. Configuration diagram for pictures

65

Modular and Systematic Interface Design for Rich Internet Applications

describing the dynamics of a riA with Adv-Charts ADV-Charts are a variant of state machines that allow expressing interface transformations which occur as the result of user interaction. They are similar to StateCharts (UML, 2008), though they are more expressive in communicating the dynamics of interfaces, as it is possible to nest states in objects and objects in states as shown in the examples of this section. As shown in the example of Figure 5, a transition in an ADVchart is annotated with an ID, the event(s) that causes it, a precondition that must be satisfied in order for the transition to fire, and a post-condition which is obtained after processing it. This post-condition is expressed in term of object’s properties that are changed after the transition. We also use a function Focus which indicates the position of the cursor and a pseudo-variable PerCont (referring to the perception context) to indicate the objects which are perceivable; these objects are “added” or “subtracted” from the perception context. The keyword “owner” references the observed ADO which can be used as part of a transition definition for querying the owner’s state. In Figure 5 we see the ADV-Chart specifying the behaviour of the ChangeablePicture compo-

Figure 5. A simple transition

66

nent: when the mouse is over an icon, the current image and description objects must be updated with the data corresponding to the index of the icon. The owner is then asked for returning these objects, given an index as a parameter. The arrow back to itself points out that the SmallImage component will be at the initial state after the transition is accomplished. The composite nature of ADV-Charts allows (by nesting states into ADVs) indicating how different lower-level ADVs are affected when the user interacts with the system. They can be also used (in combination with configuration diagrams) to indicate the way in which conceptual or navigational operations are triggered by interface events. While the nesting of states in ADVs follows the Statecharts semantics, meaning that an ADV can be in some states (either AND-ed or XOR-ed), the nesting of ADVs inside states shows the ADVs that might be perceivable in that state.

Dealing with Interface Complexity and Evolution As discussed earlier, a critical design issue for complex RIA arises from the fact that they deal with different application concerns. Some concerns occasionally crosscut each other. In some cases the crosscutting concerns are volatile: they arise

Modular and Systematic Interface Design for Rich Internet Applications

during the application’s evolution and perhaps they will have to be removed later from the application, due to its temporal or beta characteristics. The key for dealing with these problematic concerns is to use well-known separation of concerns techniques. While the crosscutting or volatility (Rossi, 2006) problem may appear both at the conceptual, navigational or user interface layers, we will focus only on interface issues. The reader can find further information on our approach for the other design levels by referring to (Ginzburg, 2007; Gordillo, 2006; Rossi, 2006). As an example, in Figure 6 we show how the ChangeablePicture component described above has been improved by incorporating comments for specific areas of the pictures. While this new

functionality is in beta test (e.g. waiting for users’ feedback) the original interface component should remain oblivious from it. This RIA behaviour can be seamlessly incorporated to the original component by using a Decorator (Gamma, 1995) combined with a transformational approach. First a decorator for the original Picture conceptual class is created. This decorator enhances it with the getSquarePosition() and getComment() methods. Then the ADV ChangeablePicture is improved by replacing its static image field with a new decorated version; corresponding ADV and ADV-Chart are specified in Figure 7. The DecoratedPicture ADV-Chart is composed of two main states: if its owner (a decorated picture) has comments, then it is in the Commented

Figure 6. Incorporating a new interface feature

Figure 7. Improved ADV and ADV-chart

67

Modular and Systematic Interface Design for Rich Internet Applications

state; otherwise it is Uncommented and only the static image is perceivable. When the ADV-Chart is on the Commented state, the Square and the Comment ADVs may be perceivable. Transition 1 of this ADV-Chart specifies that the square has to disappear 5 seconds after the ADV was opened. The second transition indicates that when the mouse is over the DecoratedPicture instance, the square has to be perceivable again (until the mouse is not more over the component which is specified in transition 3). Transitions 4 and 5 respectively specify the dynamic of the comment bubble in function of the focus over the square ADV. Finally, we need to include the new specified structure into the original component. For expressing the integration, we have defined a simple specification language (Ginzburg, 2007) which allows indicating point-cuts and insertions at the abstract interface level. Using it, we can apply the following transformation over the ADV ChangeablePicture in order to dynamically replace its static image field with the new decorated version.

position field specifies that the referenced “relativeTo” component must be replaced.

Some Words on Implementation

Target ADV ChangeablePicture

ADVs can be systematically mapped into web pages implemented with technologies like JSP, JSF, XSL, etc. and rich behavioural aspects can be implemented using AJAX (Garrett, 1995) or OpenLaszlo (“Open Laszlo”, 2008) which are XML based technologies. At the same time, transformation specifications can be mapped into XSL Transformations (“XSL”, 2008). These transformations are capable of inserting, deleting or replacing fragments of code belonging to the user interface and implemented with XML-compliant languages like the ones mentioned before. Using XSL transformations, rich behaviours can be incorporated in the interface by inserting blocks of JavaScript functions. In some cases, when existing interface behaviours are overridden, we may profit from a JavaScript facility which allows to redefine functions at runtime and to wrap one function into another (“Aspect Oriented Programming and Javascript”, 2007).

Add Image DecoratedPicture

specifying riA interface Patterns

RelativeTo ADV ChangeablePicture.Image

To complete our presentation, we show in this section how different interface patterns can be specified using our approach. Patterns are a topic of research since the 90’s in software engineering (Gamma, 1995) and more recently in the Web Engineering arena (Van Duyne, 2003). Patterns are a good way to describe recurrent problems and their solutions such that these solutions can be reused each time the problem arises. Using patterns experienced designers can convey their wise strategies to novices, therefore leveraging the level of design projects. With this same goal, an impressive set of RIA patterns have emerged ; their aim is to describe

Position replace The field “Target” indicates the name of the ADV which will suffer the transformation. Also, inner ADVs may be specified using a “.” .The “Add” field indicates which elements must be inserted in the target, either an ADV or an immediate specification, which is used when the inserted field is simple enough to avoid the specification of another (auxiliary) ADV. Finally, we indicate the insertion position by using the “Relative” fields, which in this case is the inner ADV Image. The

68

Modular and Systematic Interface Design for Rich Internet Applications

solutions for recurrent design problems in Web software by using rich interface behaviours. While patterns in (“Yahoo! Patterns”, 2008) are informally specified, ADVs and ADV-Chart can be used as specification tools to formalize the structure of the solution of these RIA patterns therefore improving reuse of interface design aspects. Patterns are usually described using templates such as the one described in (Gamma, 1995); for the sake of simplicity we will use a simplified version of this template, containing:





Some times a user needs to fill a text box where the expected data is a large string, therefore with the risk of introducing typos. Figure 8 shows a typical text field in a Web e-mail client context. When sending mail to a not usual address (i.e. people who we don’t contact usually) we may not remember the address and to solve this problem, we have to navigate to an

• • •

Pattern name: a simple key reference for designers and coders. Motivation: a brief and concrete description of the real problem and its context. Intent: A summary of the pattern aim Structure: Using ADV and ADV-Charts, the solution is modelled describing its structural and behavioural aspects.



Participants: The most important elements in the pattern Known uses: Usage example/s.

To illustrate our approach we will describe two simple patterns from (“Yahoo! Patterns”, 2008): auto-complete and carrousel.

Auto-Complete Motivation

Figure 8. “To:” text field WITHOUT auto-completion

Figure 9. “To:” text field WITH auto-completion

69

Modular and Systematic Interface Design for Rich Internet Applications

out list, select target contacts and finally continue writing. In Figure 9, the text field is enriched with an auto-completion feature which also highlights the part of the address which matches with the input field.





Intent When the valid set of input data is known, we can prevent mistypes by showing compatible matches to the partial field’s value.



Structure The structure comprises three main diagrams and one specification for describing the whole pattern behaviour: •

Participants •

ComplexADV: a target ADV must match this structure where there is a search like

Figure 10. Target interface stereotype

Figure 11. Pattern structure for auto-complete pattern

70

functionality by means of a textField. The ADV is presented in Figure 10. A Search Result ADV-Chart (Figure 11 at left) describes ADV state transitions, their conditions and implied changes. SearchResult ADV (Figure 11 at center) describes the interface structure where matching results will be placed, and popped up to the user below the compleADV component. Integration Speciication (Figure 11 at left) speciies to the weaver that the Search result must be placed below compleADV component.

SomeTextField: Text ield which will be illed by the user. A simple text ield context is shown in Figure 10.

Modular and Systematic Interface Design for Rich Internet Applications



SearchResult ADV: Interface component which will show the resolved information (Figure 11 right). Mainly, it indicates that the result component comprises a set of strings.

Examples Google Mail: Uses auto-complete pattern on all of its address ields, i.e.: From, To, CC, etc. (See www.gmail.com) Yahoo search engine: Uses auto-complete for providing common searches witch matches with search ield input. (See www.yahoo. com)

Carrousel Motivation A Web page presenting a long linearly ordered index of items in an e-commerce site may require the user to repeatedly scroll the whole page. In order to improve navigability and highlight some index items, one possible improvement is to make the index scrollable in a restricted space. A pair of buttons (Left and Right) are introduced which trigger a shift (left or right) between items as shown in Figure 12 highlighting the main index component. Amazon has spread this pattern all over the site improving user navigability and experience.

Intent Instead of using the browser scroll-bar, we introduce two application scroll controls which help to reduce the space devoted to the index. Alternatively, the scroll could be vertical or circular, in a carrousel style; this latter style might be preferable for a small number of elements because we can keep them all visible.

Structure In this case, there are two diagrams for describing just one interface structure and behaviour that can be part of a most complex interface as shown in Figure 12 where it is part of a more complex page. In Figure 13, we see the specification of Carrousel using ADV-Charts. The variable OFFSET references the carrousel’s window offset with respect to the owner’s list, which points out the starting index from which the owner’s items are shown.

Participants In figure 13, we can identify the following participants: •

Carrousel ADV: It is the main user interface component which holds items ADVs and a pair of control buttons: Left and Right.

Figure 12. Amazon’s improved index

71

Modular and Systematic Interface Design for Rich Internet Applications

Figure 13. Carrousel pattern definition





Item ADV: It is the element which will be shown and will be forced to shift positions with one of its neighbour items. Left and Right buttons: They trigger, after a click over them, the shift behaviour to the left and to the right respectively.

Example Amazon store: It uses carrousel pattern in its scrollable indexes. (www.amazon.com) Yahoo: Uses carrousel for presenting video indexes. (www.yahoo.com)

fUtUre reseArCh direCtions We are currently working in several research subjects: First, we are improving tool support to be able to derive implementations in an easier way. We are also formalizing an approach for refactoring conventional Web applications into RIA by means of oblivious compositions as described in (Rossi, 2008). We are also building tools to animate ADV-charts in order to have a more “agile” development methodology for RIA.

ConClUsion In this chapter we have discussed the problem of specifying the interface of Rich Internet Applica-

72

tions, those applications which offer the user a better interaction experience than the hypertextlike style of Web 1.0 software. We have outlined our approach for taming the complexity of RIA interfaces. The OOHDM approach offers a set of modelling primitives to describe, in a modular way, the structure and dynamics of a RIA interface. Particularly, we have described how to use Abstract Data Views and their associated ADVCharts, a variant of Statecharts. We have also shown how to further decompose the interface space by decoupling interface objects belonging to different concerns (Urbieta, 2007), e.g. when evolution occurs. In this way we can support a seamless evolution of the application, as concern composition is performed obliviously. We illustrated our approach with some simple examples and presented some RIA interface patterns and their corresponding specifications.

referenCes Ajax Patterns. (2008). Retrieved in September 2008, from http://ajaxpatterns.org/ Aspect Oriented Programming and Javascript. (2007). Retrieved in September 2008, from http://www. dotvoid.com view.php?id=43 Bozzon, A., Comai, S., Fraternali, P., & Toffetti Carughi, G. (2006). Conceptual modeling and code generation for rich Internet applications. ICWE, 2006, 353–360. doi:10.1145/1145581.1145649

Modular and Systematic Interface Design for Rich Internet Applications

Ceri, S., Fraternali, P., & Bongio, A. (2006). Web modeling language (WebML).Amodeling language for designing Web sites. Computer Networks and ISDN Systems, 33(1-6), 137-157. Cowan, D., & Pereira de Lucena, C. (1995). Abstract data views: An interface specification concept to enhance design for reuse. IEEE Transactions on Software Engineering, 21(3), 229–243. Consortium, U. W. A. (2002, October 16-18). Ubiquitous Web applications. In Proceedings of the eBusiness and eWork Conference 2002, (e2002), Prague, Czech Republic. Van Duyne, D., Landay, J., & Hong, J. (2003). The design of sites. Addison-Wesley. Gamma, E., Helm, R., Johnson, R., & Vlissides, J. (1995). Design patterns. Elements of reusable object-oriented software. Addison Wesley. Garrett, J. (2005). Ajax: A new approach to web applications. Adaptive path. Retrieved from http:// www.adaptivepath.com/publications/essays/archives/000385.php Ginzburg, J., Rossi, G., Urbieta, M., & Distante, D. (2007, July 16-20). Transparent interface composition in Web applications. In Proceedings of the 7th International Conference on Web Engineering (ICWE2007) (pp. 152-166), Como, Italy. Springer. Gordillo, S., & Rossi, G. Moreira, A., Araujo, A.,Vairetti, C., & Urbieta, M. (2006). Modeling and composing navigational concerns in Web applications. Requirements and design issues. Proc. of 4th Latin American Web Conference (pp. 25-31). Jacobson, I. (1996). Object-oriented software engineering. ACM Press. Koch, N., Kraus, A., & Hennicker, R. (2001). The authoring process of UML-based Web engineering approach. In Proceedings of the 1st International Workshop on Web-Oriented Software Construction (IWWOST 02) (pp. 105–119), Valencia, Spain. Kolbitsch, J., & Maurer, H. (2006, June). Transclusions in an HTML-based environment. Journal of Computing and Information Technology, 14(2), 161-174.

Linaje, M., Preciado, J., & Sanchez-Figueroa, F. (2007, July 16-20). A method for model based design of rich Internet application interactive user interfaces. In Proceedings of the 7th International Conference on Web Engineering (ICWE2007) (pp. 226-241), Como, Italy. Springer. Meliá, S., Gómez, J., Zhang, G., Kroiß, C., & Koch, N. (2008, July 14-18). A model-driven development for GWT-based rich Internet applications with OOH4RIA. In Proceedings of the 8th International Conference on Web Engineering (ICWE2008), New York. IEEE Press. Nelson, T. H. (1981). Literary machines. Mindful Press Nelson, T. H. (1995). The heart of connection: Hypermedia unified by transclusion. Communications of the ACM, (8): 31–33. doi:10.1145/208344.208353 OpenLaszlo. (2008). Retrieved in September 2008, from http://www.openlaszlo.org/ Pastor, O., Abrahão, S., & Fons, J. (2001). An object-oriented approach to automate Web applications development. In Proceedings of EC-Web (pp. 16–28). Preciado, J. C., Linaje, M., Sanchez, F., & Comai, S. (2005). Necessityof methodologies to model rich Internet applications. IEEE Internet Symposium on Web Site Evolution, 7-13. Preciado, J., Linaje, M., Morales-Chaparro, R., SanchezFigueroa, F., Zhang, G., Kroiß, C., & Koch, N. (2008, July 14-18). Designing rich Internet applications combining UWE and RUX-method. In Proceedings of the 8th International Conference on Web Engineering (ICWE2008), New York. IEEE Press. Rossi, G., Nieto, A., Mengoni, L., Lofeudo, N., Nuño Silva, L., & Distante, D. (2006). Modelbased design of volatile functionality in Web applications. Proc. of 4th Latin American Web Conference.

73

Modular and Systematic Interface Design for Rich Internet Applications

Rossi, G., Urbieta, M., Ginzburg, J., Distante, D., & Garrido, A. (2008, July 14-18). Refactoring to rich Internet applications. A model-driven approach. In Proceedings of the 8th International Conference on Web Engineering (ICWE2008), New York. IEEE Press.

Rossi, G., Urbieta, M., Ginzburg, J., Distante, D., & Garrido, A. “Refactoring to Rich Internet Applications. A Model-Driven Approach” In Proceedings of the 8th International Conference on Web Engineering (ICWE2008: July 14-18, 2008; New York, USA), IEEE Press, 2008.

Schwabe, D., & Rossi, G. (1998, October). An object-oriented approach to Web-based application design. [TAPOS]. Theory and Practice of Object Systems, 4(4), 207–225. doi:10.1002/(SICI)10969942(1998)4:43.0.CO;2-2

Urbieta, M., Rossi, G., Ginzburg, J., & Schwabe, D. “Designing the Interface of Rich Internet Applications” Proc. of 5th Latin American Web Conference (LA-WEB 2007, Santiago, Chile), IEEE Press, 2007.

UML. the Unified Modeling Language. (2008). Retrieved in September 2008 from http://www. uml.org/ Urbieta, M., Rossi, G., Ginzburg, J., & Schwabe, D. (2007). Designing the interface of rich Internet applications. In Proc. of 5th Latin American Web Conference (LA-WEB 2007), Santiago, Chile. IEEE Press. Vilain, P., Schwabe, D., & de Souza, C. S. (2000). A diagrammatic tool for representing user interaction in UML (pp. 133-147). York, UK. Wright, J., & Dietrich, J. (2008). Requirements for rich Internet application design methodologies. XSL. The Extensible Stylesheet Language Family. (2008). Retrieved in September 2008 from http:// www.w3.org/Style/XSL/ Yahoo! Patterns. (2008). Retrieved in September 2008, from http://developer.yahoo.com/ypatterns/

AdditionAl reAding Linaje, M., Preciado, J., & Sanchez-Figueroa, F. “A Method for Model Based Design of Rich Internet Application Interactive User Interfaces” In Proceedings of the 7th International Conference on Web Engineering (ICWE2007: July 16-20, 2007; Como, Italy), pp. 226-241, Springer, 2007.

74

Key terMs And definitions ADV (Abstract Data View): A model which allows specifying the structure of interface objects and their relationships with other software components. The behavioural aspects of the interface are specified using ADV-charts, which are a variant of StateCharts Crosscutting Concern: A concern that affects other concerns. These kinds of concerns often cannot be cleanly decomposed from the rest of the system in both the design and implementation. OOHDM: Object Oriented Hypermedia Design Method is a method for the development of Web applications which consists of five activities, requirements gathering, conceptual design, navigational design, abstract interface design and implementation Separation of Concerns: The ability to identify, encapsulate and manipulate those software artifacts which are relevant to a specific concept, task or purpose User Interface Pattern: A general and reusable solution for recurrent user interface design problems. Volatile Functionality: A kind of functionality that is presented in an application during a short period of time.

75

Chapter 5

Towards Web 2.0 Applications: A Conceptual Model for Rich Internet Applications Alessandro Bozzon Politecnico di Milano, Italy Sara Comai Politecnico di Milano, Italy Piero Fraternali Politecnico di Milano, Italy Giovanni Toffetti Carughi Università della Svizzera Italiana, Switzerland

AbstrACt This chapter introduces a conceptual model for the design of Web 2.0 applications relying on rich Internet application (RIA) technologies. RIAs extend Web application features by allowing computation to be partitioned between the client and the server and support core Web 2.0 requirements, like real-time collaboration among users, sophisticated presentation and manipulation of multimedia content, and lexible human-machine interaction (synchronous and asynchronous, connected and disconnected). The proposed approach for the design of Web 2.0 applications extends a conceptual platform-independent model conceived for Web 1.0 applications with novel primitives capturing RIA features; the conceptual model can be automatically converted into implementations in all the most popular RIA technologies and frameworks like AJAX, OpenLaszlo, FLEX, AIR, Google Gears, Google Web toolkit, and Silverlight.

introdUCtion Rich Internet Applications (RIAs) extend traditional Web architectures by allowing computation to be reliably partitioned between the client and the server, by supporting data storage on the client,

rich interactive interfaces, and rich communication modalities between the client and the server, without sacrificing the openness and universality of browser-based user interfaces. They are an essential ingredient of the Web 2.0, because they blend the best of Web-enabled and desktop architectures and address core Web 2.0 requirements, like real-time

DOI: 10.4018/978-1-60566-384-5.ch005

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Towards Web 2.0 Applications

collaboration among users, sophisticated presentation and manipulation of multimedia content, and flexible human-machine interaction (synchronous and asynchronous, connected and disconnected) (Bughin, 2007). As RIA adoption is growing, a multitude of programming frameworks have been proposed to ease their development. These increase productivity, but are bound to a specific technology and therefore not easily portable across different platforms. This chapter proposes a different approach based on the conceptual, platformindependent design of rich Internet applications, in the tradition of Model Driven Development. The essential innovation is the presence of a high-level, platform-independent, and technology-neutral model of the RIA application, which can be used to describe all its relevant features and can be automatically converted into implementations in all the most popular RIA frameworks. The model is visual and intuitive, so to be usable also by non-programmers, and at the same time rigorous and formal, so to enable automatic code generation. The proposed conceptual model comprises simple and intuitive extensions of the concepts and notations used for modelling traditional Web applications. The Chapter shows the conceptual model at work, by illustrating several real-life RIA design patterns relevant to the design of effective Rich Internet Applications, and describes a prototype for the visual specification of the conceptual model and the automatic code generation of the final application. The Chapter is organized as follows: Section 2 offers an overview of Web 2.0 and RIA, identifying the novel issues that affect the design and development of RIAs, while Section 3 describes the state of the art of Web engineering methods w.r.t. RIA design and development. Then, in Section 4 we introduce the conceptual model for RIA design, which extends the Web Modelling Language (WebML) (Ceri, 2002), a visual notation for Web 1.0 applications, with novel primitives capturing RIA features. In order to keep the exposition concrete, we introduce a running example,

76

which conveys the methodological concepts and the notation. In Section 5 we discuss how the proposed conceptual model can be exploited to derive the implementation code. Finally, in Section 6 conclusions are drawn.

bACKgroUnd The Web and desktop applications are rapidly converging: Web applications keep adding new features overcoming the traditional Web sites capabilities, and desktop applications are quickly becoming Internet-enabled to offer functionalities typical of distributed, collaborative, on-line systems (Brent, 2007). The term Web 2.0 has been used to describe a new class of Web applications strongly centred around a prominent role of the end-users. Web 2.0 applications demand also a novel development paradigm (Farrell, 2007), to overcome the limitations of the traditional HTML interfaces w.r.t. desktop applications both in terms of content presentation and manipulation (HTML was designed for documents, not GUIs, and multimedia support is limited) as well as in terms of interaction (server-side computation implies full page refresh at each user-generated event). Rich Internet Applications provide the technological core for Web 2.0 development enabling powerful, in terms of content presentation and manipulation, and reactive interfaces for collaborative applications that work seamlessly in a connected and disconnected fashion. In the next sections we illustrate how RIAs achieve improved interactivity, thanks to their flexible architecture, we report a short overview of the main RIA technologies, and describe the state of the art of RIA development methodologies.

riAs Architecture RIAs extend the traditional Web architecture by moving part of the application data and computation logic from the server to the client. The aim

Towards Web 2.0 Applications

is to provide more reactive user interfaces, by bringing the application controller closer to the end user, allowing fast partial interface updates and minimizing server round-trips. The general architecture of a RIA is shown in Figure 1: the system is composed of a (possibly replicated) Web application server and a set of user applications running on client machines. These applications are generally either implemented as 1) JavaScript, Flash animations, plug-in-interpreted code, or applets running inside a Web browser, or as 2) downloadable binaries (e.g., Java Web Start applications, Adobe AIR) interpreted and executed in a specific runtime environment. In both cases, client-side applications are downloaded from the server and executed following the code on demand

paradigm of code mobility (Carzaniga, 1997). From the development perspective, one of the most relevant aspects of RIAs client-side architectures is the neat separation of concerns stemming from the Web development legacy: most approaches use declarative mark-up languages for interface specification, a scripting language for event handling, interface update, client-side business logic, and (a)synchronous communication with the server to exchange data and event notifications. Persistent and temporary data on the client-side are generally stored in XML, JSON (JavaScript Object Notation), or relational format. The server-side of the overall system remains consistent with traditional Web applications, and is generally composed of a three-tier architecture.

Figure 1. RIA architecture

77

Towards Web 2.0 Applications

riA technologies RIAs can be implemented with a number of different technologies. Focusing on their functionalities, we can broadly classify them in four categories; similar classifications of RIA technologies can be found in the papers of Brent (2007) and Farrell (2007): 1.

Scripting-based: the client side logic is implemented via scripting languages (JavaScript) and interfaces are based on a combination of HTML and CSS.

The main advantage of this class of solutions is that they do not need plug-in installation as they build on browser JavaScript support and W3C standards such as HTML and CSS. In addition, JavaScript supports XML fairly well. The drawbacks are insufficient rich media support (video, audio, graphics, animations), poor debugging and development tools, browser constraints forbidding, for instance, file system access or persistent storage, and inconsistent browser behaviour. Because of the latter aspect, the developer community has seen the flourishing of a vast number of frameworks promising to abstract from browser idiosyncrasies (e.g., Backbase, Rico, DWR, Dojo, Scriptacolous, Prototype, GWT, etc. – for further details on all the technologies cited in this chapter the reader may refer to the additional reading section). 2.

78

Plug-in-based: advanced rendering and event processing are granted by browser’s plug-ins interpreting specific scripting languages, XML or media files (e.g., Flash, Flex, OpenLaszlo, Google Gears, Silverlight, AIR). Plug-in players like Flash are available in 96% of Web-enabled user terminals, including hand-held devices, and behave consistently on any browser. An advantage common to all these plug-ins is that they support media interaction natively, generally allow

client-side persistence, and provide better performances than interpreted Javascript. However, many browser-based functions, like bookmarking and HTML+CSS rendition, are not supported natively. 3.

4.

Browser-based: rich interaction is natively supported by some browsers that interpret declarative interface definition languages. The most relevant browser-based solution is Mozilla XUL. Runtime environments: applications are downloaded from the Web but can be executed outside the browser, (e.g., Java Web Start, JavaFX, XULRunner, AIR, Silverlight). These solutions offer the most in terms of client-side capabilities and off-line use with (compiled) programming languages and full access to the file system and the underlying operative system. However, they require a dedicated runtime environment, which force users to install additional software on their machines.

In this chapter we propose a conceptual model that supports RIA application design, by abstracting from specific implementation technologies; this model captures the essential features offered by RIAs such as: distribution of computation and logic across client and server, temporal and persistent data storage at the client-side, and asynchronous client-server communication. Although abstract and platform-independent, the proposed model is amenable to implementation on top of stateof-the art RIA technologies: Section 5 describes a prototype implementation of a RIA runtime environment supporting our solution.

state of the Art of riA development Methodologies Several tools supporting the development of RIAs have been proposed, and the increasing number of available development platforms confirms the

Towards Web 2.0 Applications

growth of their acceptance among developers. Tool vendors have adapted to RIA requirements their existing solutions for Web and desktop development, typically offering WYSIWYG composition of interfaces using widgets and integrated code editing (like, for example, in Visual Studio, Expression Blend, FlexBuilder, DreamWeaver, Eclipse, NetBeans). In most cases, IDEs are used to model the interface and the behaviour at the client-side, leaving the specification and development of the service / server tier to other tools and methodologies. The focus is on implementation for a specific framework / platform, rather than modelling, providing little chances of reusing specifications. On the contrary, the approach described in this chapter provides the designer with a perspective spanning the complete runtime behaviour of a Rich Internet Application, addressing both clientand server-side computation and data, as well as client-server communication. Web Engineering approaches build on Web architectural assumptions to provide simple but expressive notations enabling complete specifications for automatic code generation of (data-intensive) Web applications. Several methodologies have been proposed in literature like, for example, Hera (Vdovjak, 2003), OOHDM (Schwabe, 1996), WAE (Conallen, 2002), UWE (Koch, 2004), WebSA (Melià, 2006), OO-H (Gomez, 2001), and W2000 (Baresi, 2001), but, to the best of our knowledge, none of them has yet completely addressed the lack of modelling concepts that traditional approaches show w.r.t. the novel architecture, functionalities, and behaviour introduced by Rich Internet applications. Current Web modelling approaches as well as hypermedia methodologies have been investigated in (Preciado, 2005): the survey shows the limits of current methods and tools with respect to RIA development support. The extensions we propose in this chapter represent a consistent effort to overcome such limits, even if some concepts have partially been defined in previous methodologies. In particular, the WAE (Conallen,

2002) methodology uses different stereotypes to denote components running on the client and on the server. While the original idea referred to either simple DHTML or thick clients, it can be used to represent generalized client-side computation and data, as required by RIAs. However, the limits of the WAE approach reside in being too close to implementation and in not providing a clear separation between data and business objects on client and server tier. Furthermore, WAE does not easily enable automatic code generation, due to the lack of precise semantics of the methodological concepts. UWE (Koch, 2004) extends Conallen’s concept of “client page” in UML deployment diagrams, to specify the distribution of application components. Also in this proposal, model semantics and code generation for RIA clients is not addressed. As of today, none of the Web engineering methodologies studies the implication of distinguishing between client and server components in the overall design and code generation of a Web application. Some recent contributions concentrate on the important (yet partial) aspect of providing formal notations to specify the user interface behaviour of RIAs (Linaje, 2007 – Urbieta, 2007 - MartinezRuiz, 2006 – Dolog, 2007). The limits of these proposals lay in focusing mainly on user interfaces issues, overlooking the full novelty of the RIA paradigm: real-time collaboration among users, sophisticated presentation and manipulation of local and remote content, and flexible humanmachine interaction (synchronous and asynchronous, connected and disconnected). With respect to the cited proposals, our approach provides a general underlying modelling framework upon which richer interfaces can be specified. The modelling concepts that will be presented in this chapter provide the primitives to address the novel concerns that are specific to RIA development focusing mainly on the business logic of the application and on its functional aspects. The aim is to empower designers to produce specifications defining the complete functioning of RIA com-

79

Towards Web 2.0 Applications

ponents and contexts: client and server, on and off-line, carefully considering the specific tradeoffs of distributed data and computation.

Modelling riA APPliCAtions A RIA application can be described by its structure and behavior. The former comprises a data model, which specifies the content objects underlying the applications, and an interface model, which describes the front-end exposed to the user. The latter is represented by a dynamic model that describes what happens when the user or other sources of events interact with the application. After establishing a concrete running case in Section 4.1, Section 4.2 introduces the essential aspects of the data model, Section 4.3 presents the structure of the interface model, and, finally, Section 4.4 describes the dynamic model.

Case study Application To ease the exposition, throughout the paper we will use, as case study, a simplified version of a travel agency application, offering users search and reservation functionalities for flights, hotels, and car rentals by means of a RIA. In particular, we will show the features that can be added to a traditional Web 1.0 application considering the new architecture of RIA technologies. Such features are not bound to the chosen example, but may be applied to any other traditional application. The main interaction object of the application is represented by the planned trip, specifying the information about the start and end date for the trip, as well as its total cost. Registered users are allowed to create new trip plans and to manipulate them by searching and selecting flights to their selected destinations, hotels for their stay and, possibly, car rentals for their on-site transportation. Trip plans are stored at the client-side for later reuse, possibly in a disconnected manner: given a planned trip the user may choose differ-

80

ent alternatives for hotels, flights, and car rentals, also when she is offline. The application must also offer a reconciliation function that aligns the information stored on the client (like the total cost of a trip) with their updated status on the server. Finally, the user can confirm a trip plan stored on the client by issuing a purchase order. Besides supporting off-line usage, the application must incorporate also typical RIA features, such as the minimization of server round-trips and partial page refreshes.

data Model The data model specifies the data used by the application. In traditional Web 1.0 applications content resides solely at the server-side, either in the form of database tuples or of main memory objects contained in server-side user sessions. With RIA technologies, content can also reside in the client, as main memory objects with the same visibility and duration of the client application, or, in some technologies, as persistent client-side objects. Data are therefore characterized by two different dimensions: (1) The tier of existence, which can be the server or the client, and (2) The level of persistence, which can be persistent or temporary. The tier of existence and the level of persistence of data are added as a refinement to the initial data model that identifies the core objects of the application and their relationships. When interaction requirements become clear and specifications start to consolidate into a design document, these properties of the data model become relevant. The enrichment of the data model with the tier of existence and the persistence level can be done by following a few refinement guidelines. Data shared among multiple users and accessed in multiple application runs, must be persistent at the server-side. Conversely, content owned by the

Towards Web 2.0 Applications

individual user or created and manipulated by the user can be stored on the client (either temporarily or persistently). Persistent client-side data can be specified for offline usage, so that the application can be used in a disconnected way like any desktop application. The distribution between the two tiers may require data replication between the client and the server. For example, client-originated content may be initially stored at the client-side and then, at the end of a usage session, may be saved persistently at the server-side; on the contrary, server-side content that needs to be manipulated in complex ways (e.g., as in record-set editing) could be replicated on the client. Therefore, the refined data model may contain replicated entities, which express the need of having content objects both at the server- and client-side. From the notation standpoint, the data model can be any formal model capable of expressing finite collections of typed objects stored extensively. For example, UML class diagrams, EntityRelationship diagrams, or a relational database can be adopted as data models (at different levels of abstraction). In the sequel, we will represent the data model by means of the essential Entity-Relationship concepts: entities, i.e., named collections of objects characterized by a set of attributes, and binary relationships, i.e., relationships over the population of two entities. For capturing the two dimensions characterizing RIA data, the conventional notation is extended with the graphical specification of the tier of existence and of the level of persistence, on both entities and relationships: we denote server and client entities/relationships with a “S” or a “C” icon respectively, and persistent and temporary entities/relationships with a filled or an empty icon, respectively. In order to guarantee the correctness of the data specification when different tiers and persistence levels are combined, the following constraints hold: it is not possible to connect temporary objects with persistent relationships or to define

a server-side persistent relationship between client-side persistent entities. These constraints guarantee that at the end of the application session all the persistent relationships connect existing persistent objects, while temporary relationship instances are automatically eliminated when the application terminates. Figure 2 shows a possible data model for the case study application. The entities on the lefthand side are marked with a filled “S” icon and represent persistent server data that will be stored in a server-side database: the User entity represents a registered user, while the Flight, Hotel, and Car entities represent the resources offered to such users by the Web application, i.e., the available reservation items for their trip planning. The User entity is also associated with the entity Reserved Trip, which contains all the data needed to represent a purchased travel plan; the relationships between Flight, Hotel, and Car with the Reserved Trip entity represent the associations between the purchased plan and the resources selected by the user. All such relationships are persistent and server-side (when not explicitly specified, relationships inherit their type from the entities they connect; if they have different persistence levels, they must be stored with temporary persistence; otherwise, if they are on different tiers, they are stored on the client). To support navigation and updates of the data of a planned trip also when the user is disconnected some data are duplicated and saved persistently on the client: in particular, the selected Flight, Hotel and Car entities and the data of the Planned Trip are replicated on the client (they are represented with a filled “C” icon). To avoid data replication inconsistencies, client entities should contain only data that, for their nature, can be consistently processed also in a disconnected way: in the running example, since the prices of a hotel or flight might vary according to their reservation status, their latest updated values is not saved on the client but are always retrieved dynamically from the server. Nonetheless, in order to provide

81

Towards Web 2.0 Applications

Figure 2. Data model for the Travel Agency sample application

some significant information also during off-line usage, the total price for a trip plan is stored on the client to offer a snapshot of the trip’s cost at a given time: reconciliation with updated server data is achieved by means of a synchronization function (discussed in the next section).

interface Model The interface model specifies the organization of the front-end of a Web application, by addressing the definition of the presentation and business logics, i.e., the content of the pages and the mechanisms to support user’s navigation and interaction. RIA technologies allow the distribution across client and server of the presentation and business logic of the application. In particular, the designer can specify how the computation of the page and of its content is distributed between the client and the server, how distributed data are managed (to minimize data transmissions), how and when replicated data are synchronized, etc. In the sequel, we will model the front-end of the application using the WebML notation (Ceri, 2002), a visual and intuitive notation that allows to express in a precise and natural way the con-

82

cepts and mechanisms of Web applications. The proposed approach is very general, can be mapped also to other standard notations (e.g., UML - for WebML a UML 2.0 profile has been defined in (Moreno, 2006)) or can be applied to other Web engineering notations and methodologies that allow the specification of the interface composition and navigation. Structure of the application. From the technological standpoint RIAs have a different physical structure than traditional Web 1.0 applications: the former typically consist of a single application ”client-container” (e.g., a Java applet or a FLASH movie), which loads different data and components based on the user’s interaction. The latter consist of multiple independent templates, processed by the server and simply rendered by the client. In terms of Web pages, conventional Web applications typically consist of a collection of ”flat” independent pages, atomically computed by the server and rendered by the client; in RIAs, instead, the structure of the interface consists of a topmost page (eventually contained into a traditional, server-computed HTML page) partitioned into peer-level sub-pages, independently calculated and rendered by the client, possibly in collaboration with the server. As a consequence,

Towards Web 2.0 Applications

it becomes important to distinguish between the two types of pages, by classifying them into: 1.

2.

Server pages: they represent traditional Web 1.0 pages; content and presentation are calculated by the server, whereas rendering and event detection are handled by the client. For script and plug-in based RIA technologies (Section 2.2), server pages might be the outmost container of a RIA application. Client pages: they represent pages incorporating content or logics managed (at least in part) by the client. Their content can be computed at the server or client side, whereas processing, rendering and event handling occur at the client side. To reflect the complex, single-application shell structure of RIA applications, client pages can contain other client sub-pages.

In WebML pages are depicted as rectangles and are associated with a name. To distinguish the two kinds of pages we mark them with a circled

“S” or “C” icon to denote that they are server or client pages, respectively. As an example consider the client page in Figure 3, depicting a fragment of the interface model of the travel agency application: it represents a client page (named RIA Travel Application and highlighted with (1)) marked with a circular “C” icon; this page includes different sub-pages: MyFlights (2), MyAccommodations (3), and MyTrips (4). Each sub-page addresses a particular task and corresponds to a distinct part of the user interface. Content of the application. For each page (or sub-page) the interface model specifies the data to be shown, the available interaction mechanisms, and the operations that may be triggered by the user using the provided interaction mechanisms. According to the WebML notation for Web 1.0 applications, pages comprise content units, representing components for content publishing: the content displayed in a unit typically comes from an entity of the data model, and can be determined by means of a selector, which is

Figure 3. Extract of the interface model for the Travel Agency sample application showing its general structure

83

Towards Web 2.0 Applications

a logical condition filtering the entity instances to be published. The set of the entity instances retrieved by the unit will be referred to also as its population. Instances selected for display can be sorted according to ordering clauses. Units are connected to each other through links, which allow the user to navigate the hypertext and to carry link parameters, i.e., values that are transported from the source unit to the destination unit. The destination unit can use these parameters (also called its input parameters) in its selectors. WebML also allows specifying operation units implementing arbitrary business logic; in particular, a set of data update operations is predefined, whereby one can create/delete/modify the instances of an entity, and create or delete the instances of a relationship. To support RIA design all the WebML concepts are refined with the explicit specification of distribution between the server and the client: content and operation units, selectors, and ordering clauses can be defined either as server or as client, with some constraints on the possible combinations. Units contained in a server page are computed by the server and are defined as server units, while units contained in a client page are computed by the client (possibly invoking the server) and are defined as client units. For a client unit it is possible to: 1)

2)

3)

84

publish or manipulate content locally stored at the client-side or (by invoking the server) at the server-side (i.e., the reference entity/ relationship of the unit can be either a server or a client one, persistent or temporary, as seen in the data model); have client-side selector conditions and/or server-side selector conditions; the former are computed locally at the client, whereas the latter are executed at the server-side. have client-side or server-side ordering clauses; the former are computed locally at the client, whereas the latter are executed at the server-side.

Instead, server units are entirely computed by the server and therefore cannot use client-side entity/relationships and cannot comprise client-side selectors and ordering clauses. Indeed, in Web 1.0 applications all the computations performed by the server must rely only on data and operations computable at the server side to cope with the asymmetric nature of the Web, where the client calls the server and not vice versa. In this chapter we will focus on client pages and units and on the typical features of RIAs. Figure 4 shows a fragment of interface model of the case study; the most interesting cases include those distributing and mixing client and server concepts and those exploiting the client storage capacities and client computation logic. They will be explained in more detail in the following paragraphs. As a first example consider the fragment of interface model in Figure 5, supporting the search for a flight. The search function for a new flight is provided in the main page of the application (RIA Travel Application): here the user can enter the source and destination location, and the start and end date of the desired flight, through the Specify Trip Location entry unit. When the user submits his request, the link exiting such unit is followed, carrying a set of parameters (SLocation, ELocation, StartDate, EndDate) that will be used in the selector conditions of the destination unit. An initial list of flights can then be retrieved from the server, represented by the AvailableFlights index unit, defined over the Flight server entity (denoted with a filled “S” icon like in the data model), filtered by means of the [ELocation == Destination], [SLocation == Source], [StartDate = ArrivalDate] server-side selector conditions (denoted with a “S” icon), and sorted according to the (SortBy: Price) server-side ordering clause. Once an initial set of flights has been retrieved from the server, client-side selectors and ordering clauses allow to refine the filtering conditions and to sort the data according to different criteria, directly on the client.

Towards Web 2.0 Applications

Figure 4. Interface model for the Travel Agency sample application

Figure 5. Fragment of the interface model of Figure 4: flight search functionality

In particular, the retrieved instances can be locally filtered by means of the Refine Trip Search entry unit, which provides inputs for the

two client-side conditions on price and operator of the given flight ([Price . Now suppose that all the units in Figure 8 have already been computed according to the scenario just described and that the user wants to change location (possibly keeping the same choices on the price limits and the operator): the set of available flights will change and as a consequence the selected flight details should be invalidated. The sequence associated with link SearchF is: . Notice that, similar extensions apply also to the other links of the page. In this example, we can see that some units may need to be completely invalidated, while for others the invalidation concerns only the current population but not the input parameters (and therefore are only emptied). In traditional Web applications, all the sequences associated with the navigational links of the page require the invalidation (possibly keeping the parameters associated with the interaction) of all the units of the page and the computation of all the units that have to be displayed. For example, the sequence associated with SearchF would be: . The parameters of the previous choices on the price limits and the operator can be preserved by including them in the HTTP request associated with the SearchF link: in the dynamic model, the preservation of the input parameters is represented by the empty operator applied to the AvailableFlights index unit. Differently from traditional Web applications, in RIAs it is also possible to refresh a unit. To understand the need for the refresh operator, let us consider the interface model in Figure 8 and suppose that all the units contained in it have already been computed and that the user wants to update the price of his planned trip, by navigating link RefreshTripPrice: in this case, after the computation of all the operation units in the chain triggered by the link, only the content of the TripDetails unit needs to be recomputed, to show the updated price, i.e., the computation sequence for the link will be: .

iMPleMentAtion on toP of eXisting frAMeworKs A prototype of the RIA modelling primitives discussed in this chapter has been implemented in WebRatio (WebRatio, 2008), a CASE tool for the visual specification and the automatic code generation of Web applications. Modelling and generating RIA applications required the extension of all three major component of the WebRatio suite: 1) the IDE to edit data, navigation, and presentation models, 2) the code generator, and 3) the runtime environment. In order to refine the modelling primitives discussed in this chapter, we modified the WebRatio IDE by adding custom properties to all model elements for which data or computation can be distributed across the client or the server (e.g., entities, relationships, pages, etc.). We also enriched the model-checking rules of the tool to take into account the set of constraints introduced in Section 4 concerning client and server computation. The runtime of WebRatio had to be integrated with a brand new client runtime environment designed and implemented from scratch. The complete architecture of our prototype is shown in Figure 10. This solution is general enough to be implemented with any of the technologies presented in Section 2.2. For our experiments, we adopted Laszlo LZX (an object oriented, tagbased language that uses XML and JavaScript for declarative specification of the presentation layer of RIAs, available at www.openlaszlo.org) as the client implementation technology. The choice is motivated by the ability of LZX to transparently deploy both Flash and AJAX-based interfaces, allowing us to generate code which situates on both sides of the scripting-based vs. plugin-based border of the classification provided in Section 2.2. The internal architecture of a client application is coded in LZX and it is organized according to an MVC pattern (see Figure 10). We have:

91

Towards Web 2.0 Applications

Figure 10. Extended WebRatio run-time architecture for RIAs







A controller, built on the existing controller mechanism of Laszlo, which is responsible of handling events triggering, the computation of client-side pages, content units and operations. It is conigured by means of XML descriptors produced by the code generator and compiled with the LZX application. Model elements, which are (a) LZX components for client-side WebML page/unit/ operation runtime services, and (b) clientside state objects (DataSets) containing the data content of client-side units encoded in XML. Such elements are implemented in common libraries, as model instances, conigured through LZX descriptors. View components, consisting in presentation templates produced by WebRatio and compiled into the Flash application. There are view components only for pages and content units; operation units have no presentation.

The presence of a client-side runtime also affects components of the previous architecture for what concerns client application download and instantiation, client-to-server and server-to-

92

client communication, as well as communication granularity and format: we preserved the original usage of one of the most common MVC model 2 implementations, Apache Struts, sided by the OpenLaszlo’s LPS (server classes) for client application compiling and distribution. Additionally, the granularity level for the server components has been increased w.r.t. traditional Web applications, in order to provide Struts actions addressed to single unit computation, in contrast with the old paradigm where every action corresponded to a whole page computation. The code generator of WebRatio, which initially produced only server-side code, has also been extended to generate the client side applications. We designed all the libraries and runtime descriptors of the client-side application, implementing: configurable LZX libraries for WebML primitives, XSLT stylesheets to create runtime configuration descriptors, auxiliary libraries used for client-server communication and authentication, and XSLT stylesheets to create the LZX view components stemming from the presentation models. On the server-side, instead, the implementation steps covered: the XSLT stylesheets for the configuration of the Struts controller to support content unit invocation through HTTP,

Towards Web 2.0 Applications

the runtime class libraries for the new units, the stylesheets for the XML responses to invocations from client applications. With the current prototype implementation, automatic code generation covers only a subset of the WebML standard units (entry, data, index, multidata, data-management operations), additional units can be added by plugins. Concerning presentation generation, automatic generation is achievable as in the original WebRatio suite by deriving a stylesheet from a mockup interface.

fUtUre reseArCh direCtions The approach presented in this chapter can be extended in several directions. First of all, an extended presentation model of the application will be studied, since also the layout and the look&feel of user interfaces are affected by some features of the new RIA technologies. In particular, temporal behaviours, advanced user’s interactions, and users’ events implying only changes to the layout or to the look&feel of the application should be included at the presentation level. Other aspects, like the single-page paradigm or the computation of the business logic of the page as a consequence of users’ events have instead already been considered in this chapter. Moreover, we plan to study advanced features for displaying and navigating multimedia contents through sophisticated Rich Internet Application interfaces. Indeed, traditional Web application do not provide multimedia native support and they need plug-ins to show video and audio at the client side. Multimedia contents and animations are instead natively supported in several RIA technologies and can be exploited in several applications, like for example, in audiovisual search-enabled applications.

ConClUsion In this chapter we have presented a conceptual model for the specification of RIAs, as an extension of a notation conceived for Web 1.0 applications. Novel primitives have been introduced, focusing mainly on the distribution of data and computation between the client and the server. With the help of a case study we have discussed the trade-offs of such distributions, described some typical patterns induced by such distributions, and exemplified specific RIA functionalities such as off-line usage. We have also described the extensions needed for the specification of the computation of RIAs upon user (or, possibly, other events) interactions. Finally, we have seen how the proposed model can be automatically converted into implementations based on RIA technologies. The proposed approach has also been applied to some industrial cases (Bozzon, 2006), characterized by interfaces requiring sophisticated interactions and complex layouts. Our experience demonstrated the value of having a unique framework for modeling and implementing complex Web 2.0 applications, leveraging on Rich Internet Applications technologies to support typical requirements of this class of applications.

referenCes Baresi, L., Garzotto, F., & Paolini, P. (2001). Extending UML for modeling Web applications. In Proceedings of the 34th Annual Hawaii International Conference on System Sciences, Maui, HI. Bozzon, A., Comai, S., Fraternali, P., & Toffetti Carughi, G. (2006). Conceptual modeling and code generation for rich Internet applications. In Proceedings of the International Conference on Web Engineering (pp. 353-360), California.

93

Towards Web 2.0 Applications

Brent, S. (2007). XULRunner: A new approach for developing rich Internet applications. IEEE Internet Computing Magazine, 11(3), 67–73. Bughin, J., & Manyika, J. (2007). How businesses are using Web 2.0: A McKinsey global survey. The McKinsey Quarterly. Carzaniga, A., Picco, G. P., & Vigna, G. (1997). Designing distributed applications with a mobile code paradigm. In Proceedings of the 19th International Conference on Software Engineering (pp. 22-32). Boston, MA. Ceri, S., Fraternali, P., Bongio, A., Brambilla, M., Comai, S., & Matera, M. (Eds.). (2002). Designing data-intensive Web applications. San Francisco: Morgan Kauffmann. Conallen, J. (Ed.). (2002). Building Web applications with UML, 2nd edition. Addison Wesley. Dolog, P., & Stage, J. (2007). Designing interaction spaces for rich Internet applications with UML. In Proceedings of the 7th International Conference on Web Engineering (pp. 32-47), Como, Italy. Farrell, J., & Nezlek, G. S. (2007). Rich Internet applications: The next stage of application development. In Proceedings of the 9th International Conference on Information Technology Interfaces (pp. 413-418), Cavtat/Dubrovnik, Croatia. Gomez, J., Cachero, C., & Pastor, O. (2001). Conceptual modeling of device-independent Web applications. IEEE MultiMedia, 8(2), 26–39. doi:10.1109/93.917969 Koch, N., Kraus, A., Cachero, C., & Meliá, S. (2004). Integration of business processes in Web application models. [NJ: Rinton Press.]. Journal of Web Engineering, 3(1), 22–49.

94

Linaje, M., Preciado, J. C., & Sánchez-Figueroa, F. (2007). Engineering rich Internet application user interfaces over legacy Web models. IEEE Internet Computing, 11(6), 53–59. doi:10.1109/ MIC.2007.123 Martinez-Ruiz, F., Muñoz Arteaga, J., Vanderdonckt, J., Gonzalez-Calleros, J., & Mendoza, R. (2006). A first draft of a model-driven method for designing graphical user interfaces of rich Internet applications. In Proceedings of Fourth Latin American Web Congress (pp. 32-38), Cholula, Puebla, Mexico. Melià, S., & Gòmez, J. (2006). The websa approach: Applying model driven engineering to Web applications. Journal of Web Engineering, 5(2), 121–149. Moreno, N., Fraternali, P., & Vallecillo, A. (2006). A UML 2.0 profile for WebML modeling. Workshop Proceedings of the 6th International Conference on Web Engineering, ICWE 2006, Palo Alto, CA. Preciado, J. C., Linaje, M., Sánchez, F., & Comai, S. (2005). Necessity of methodologies to model rich Internet applications. In Proceedings of Seventh IEEE International Workshop on Web Site Evolution (pp. 7–13), Budapest, Hungary. Schwabe, D., Rossi, G., & Barbosa, S. D. J. (1996). Systematic hypermedia application design with OOHDM. In Proceedings of the Seventh ACM Conference on Hypertext (pp. 116-128). Washington, D.C. Toffetti Carughi, G., Comai, S., Bozzon, A., & Fraternali, P. (2007). Modeling distributed events in data-intensive rich Internet applications. In Proceedings of the 8th International Conference on Web Information Systems Engineering (pp. 593-602), Nancy, France.

Towards Web 2.0 Applications

Urbieta, M., Rossi, G., Ginzburg, J., & Schwabe, D. (2007). Designing the interface of rich Internet applications. In Proceedings of the Fifth Latin American Web Congress (pp.144-153), Santiago de Chile, Chile. Vdovjak, R., Frasincar, F., Houben, G., & Barna, P. (2003). Engineering semantic Web information systems in Hera. Journal of Web Engineering, 2(1-2), 3–26. WebRatio. (2008). Retrieved from www.webratio. com

AdditionAl reAding AIR. http://www.adobe.com/products/air/ Backbase, http://www.backbase.com/ Dojo, http://dojotoolkit.org/ DWR. http://getahead.org/dwr/ Flash, http://www.adobe.com/products/flash/ flashpro/ Flex, http://www.adobe.com/products/flex/ Google Gears. http://gears.google.com/ GWT. http://code.google.com/webtoolkit/ In order to better understand the underlying technology the reader may refer also to the Web sites of the main RIA technologies. Here we provide the list of the technologies cited in the chapter, together with their Web sites (last visited in August 2008). JavaF. X.http://sun.com/javafx Java Web Start. http://java.sun.com/products/ javawebstart/ MozillaX. U. L.http://www.mozilla.org/projects/ xul/ OpenLaszlo. http://www.openlaszlo.org/

Prototype, http://prototypejs.org/ Rico, http://openrico.org/ Scriptacolous, http://script.aculo.us/ Silverlight, http://silverlight.net/ XULRunner. http://developer.mozilla.org/en/ docs/XULRunner

Key terMs And definitions Client-Server: Computing architecture separating a client from a server, typically implemented over a computer network. A client is a software or process that may initiate a communication session, while a server can not initiate sessions, but is waiting for a request from a client Data Design: Design process aiming at the definition of the application’s data. Dynamic Modelling: modelling process aiming at the definition of the behaviour of the application. Model Driven Development: Software development approach based on the systematic use of models as the key artifacts throughout the engineering lifecycle, from system specification and analysis, to design and testing Rich Internet Applications: Web applications that have the features and functionality of traditional desktop applications, offering online and offline capabilities, sophisticated user interfaces, the possibility to store and process data directly on the client-side, and high levels of user interaction Web Architecture: Organization of a Web system defined in terms of structure, behaviour, communication, and composition of its components Web Engineering: Discipline studying the approaches, methodologies, tools, techniques, and guidelines for the design, development, evolution, and evaluation of Web applications

95

96

Chapter 6

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX Marco Brambilla Politecnico di Milano, Italy Piero Fraternali Politecnico di Milano, Italy Emanuele Molteni Web Models S.r.l., Italy

AbstrACt This chapter describes how the design tool WebRatio (and its companion conceptual model WebML) have been extended to support the new requirements imposed by rich Internet applications (RIAs), that are recognized to be one of the main innovations that lead to the Web 2.0 revolution. Complex interactions such as drag and drop, dynamic resizing of visual components, graphical editing of objects, and partial page refresh are addressed by the RIA extensions of WebRatio. The chapter discusses what kinds of modelling primitives are required for specifying such patterns and how these primitives can be integrated in a CASE tool. Finally, a real industrial case is presented in which the novel RIA features are successfully applied.

introdUCtion The advent of Rich Internet Applications (RIA, for short) has allowed a much broader set of user interaction possibilities within Web applications. Complex interactions such as drag and drop, dynamic resizing of visual components and graphical editing of objects were once a prerogative of desktop applications, DOI: 10.4018/978-1-60566-384-5.ch006

while now are available as standard patterns in many Web applications too. These patterns enable more flexible and usable interfaces, but at the same time require a more complicate application logics, both at client side and server side. Correspondingly, if model-driven design is adopted, new primitives and design patterns must be devised. This chapter aims at discussing what kinds of modelling primitives are required for specifying Rich Internet Applications and discusses how these

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

primitives can be integrated in a CASE tool. In addition, a real industrial case is presented in which the novel RIA features are successfully applied. The viewpoint presented here is somehow opposite to the typical academic research paper, where an abstract solution to the investigated problem is first designed and verified formally, and then applied top-down to a prototype implementation. In this chapter we report on a bottom-up approach, which has extended a real world modelling notation and tool progressively, following the penetration of RIA features in the market and the raise of interest in the customers. The chapter deals with four main aspects related to the coverage of RIA requirements in Web application design: • • • •

extensions to the conceptual model; extensions to the CASE tool elements and properties; architectural issues and code generation aspects; implementation examples in real industrial scenarios.

The conceptual modeling primitives cover the following aspects of RIAs: management of new hypertextual link behaviour, including partial page refresh, in-page popup windows, splash screens, dynamic tooltips, and animations; interaction among page objects through drag and drop and dynamic dependencies; and advanced form specifications, including text autocompletion, on-event actions, and field dependencies. Besides the modeling aspects, the chapter will describe how they are implemented within the WebRatio tool and how they are exploited through automatic code generation. The architectural description of the adopted design framework is provided, together with the analysis of the best mix of technologies that can be leveraged for implementing this kind of features. The designed architectural framework extensively exploits the XMLhttpRequest method and consists of imple-

menting each conceptual hypertext page with two dynamic pages that interact for providing the rich interface features: the first is a back-end dynamic XML page that stores the data of interest for a specific navigation context; the second is the front-end JSP page (including the required JavaScript needed for event management) that is shown to the user. The latter invokes extraction of data fragments from the back-end XML page according to the user behaviour. The original contribution of the chapter stands in the mix of conceptual aspects and industrialbased solutions that lead to a comprehensive conceptual view of the development of RIAs. To our knowledge, this is the first attempt to bring together academic research and industrial implementation in conceptual modeling of RIAs. To validate the approach, we exemplify the usage of the devised components in a real business scenario in which WebRatio has been adopted for designing and implementing RIA applications. The chapter is organized as follows: we start describing the role of RIAs in the context of Web 2.0; then we summarize some background information about RIAs and about WebML. Subsequently, we move to the core part of the chapter, describing the new conceptual modelling primitives for supporting RIAs; hence we describe the WebRatio architecture and the extensions needed for RIAs, and an industrial case study where the approach has been applied; finally, we draw some conclusions on the work.

the role of riCh internet APPliCAtions in web 2.0 RIAs represent one of the mainstream evolutions of Web applications that are recently taking place. Together with other evolution aspects, they contribute to the innovation of the Web in a subtle but radical way. Among these aspects, we can cite the success of the Web 2.0 applications, whose main characteristic is the deep involve-

97

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

ment of end-users in the success of applications, based on community behaviour, on continuous and pervasive user interaction, and on contents mainly provided by end users. The novelty of these trends does not stand only in technical innovations introduced in the Web applications, but on the new ways of using the existing technologies for achieving different objectives. Among the other phenomena that we are witnessing, rich interfaces are not usually considered as a mainstream innovation of Web 2.0. However, Web 2.0 sites very often feature a rich, user-friendly interface based on AJAX or similar rich media. In some sense, RIAs can be seen as one enabling technique for strictly speaking Web 2.0 applications. Indeed, a lot of community and interaction features rely on user friendly interfaces. Without them, many user activities involved in Web 2.0 applications (although technically feasible) would be so complex and boring that many users would probably simply give up interacting.

bACKgroUnd Our proposal of extension towards AJAX and RIAs of a well known and established CASE tool for Web application design is positioned in a quickly changing scenario of technologies and tools. We now examine the current state of the art in the field, considering contributions from the different classes of: AJAX toolkits and libraries, AJAX comprehensive IDE tools and frameworks, and conceptual model proposals for AJAX applications. We also briefly describe the WebML models that have been extended in the context of this work.

AJAX libraries and toolkits Following the exceptional growth of the RIA interfaces, several application frameworks have been proposed in the context of AJAX development. Among them, several opensource and commercial

98

projects provide libraries for rich internet application development. Among them, we can cite the most established ones: Dojo (2008), Ext (2008), Google Web Toolkit (2008), jQuery (2008), MooTools (2008), Prototype (2008) and Scriptaculous (2008), and Yahoo (2008) User Interface Library, but there are hundreds more that are flourishing. Among the advantages of AJAX libraries, we mention the fact that often developers get full access to the source code, even if not released as opensource, thanks to the fact that libraries are almost entirely developed in JavaScript, which is normally visible to developers. The most successful toolkits are probably Dojo, currently supported by IBM and Sun; Ext, a fast-growing toolkit offering both opensource and commercial licenses; and the Google’s GWT library, built on Java. To further enrich the scenario, there exist also hybrid versions now that mix and match several of the major projects, like GWT-Ext (Google, 2008, 2) and MyGWT (2008), that mix GWT and Ext, and Tatami (2008), that mixes GWT and Dojo. Other hybrid approaches are positioned between the scripting libraries and the browser desktop-like function libraries, such as XUL (Mozilla, 2008), that are often provided with appropriate development approaches (Brent, 2007). In the WebRatio runtime framework Prototype and Scriptaculous have been adopted, and thus the code generator of WebRatio produces Javscript code that exploits these libraries. Prototype provides a simple approach to manipulating a Web page, with a relatively light layer that offers both shorthand versions for popular functions and a good amount of cross-browser abstraction. Scriptaculous is a set of special effects and simple widgets built on top of Prototype. Among the various options, the choice of Prototype is motivated by the simplicity of the library, on the widespread usage, and on the good quality and reliability of the results.

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

AJAX development tools Besides the various AJAX libraries, a smaller set of development tools for AJAX exist. The four leading tools are Backbase (2008), Bindows, JackBe (2008) NQ Suite, and Tibco (2008) General Interface. All of them offer broad widget collections, rich tools, sophisticated debuggers, and development platforms that rival any of the IDEs for traditional languages. Differently from the toolkits described in the previous sections, these are full frameworks that function best when one builds the entire application on top of their structure. All of these systems are built around collections of widgets that are joined with a central backbone of events and server calls that link the widgets in a cohesive set of panes. Events flow across a central bus to all parts of the system, with an approach that is closer to desktop application than to Web page design. The major differences among these packages lie not in the capabilities, but in the server support and in the finer details of their approach. Although it may be easy to find one widget or design structure in each of the packages that outshines the others, the cores of the four packages are similar, and they’re all built around a core set of UI widgets. All these tools manipulate the DOM tree. The main drawback of these tools is that the resulting presentation tends to be rather boilerplate, not fully showing the interaction quality that might be expected from JavaScript applications. For the interaction with the server, some of the tools expect the data to be packaged in Web services, while others include extensive server frameworks that integrate the client application with backend databases. With respect to our proposal, these tools are positioned in the IDE field, supporting the developers for writing the code. They don’t provide any high-level, model-based design facility.

Model-driven Approaches to riA design Several researches have tried to highlight the capabilities and features of RIAs. The paper by Preciado et al. (2007) discusses how RIAs extend the behaviour of Web application at different levels: data, business logic, presentation, and communication. • •





In the data layer, the client can be exploited for storing non persistent data; At the business logic level, some operations (data iltering, numeric operations, and so on) can be delegated to the client; At the presentation level, new user events can be managed and new interaction paradigms are allowed (drag and drop, modal windows, and so on) At the communication level, new synchronous and asynchronous communication patterns can be exploited, allowing pushing of information from the server and partial page refresh.

The approach presented in this chapter is a pragmatic extension of WebML and WebRatio for supporting those features. These extensions are defined in terms of new properties of WebML components. Other more comprehensive works, like Bozzon (2006), aim at proposing a new structured framework for event management and interaction design. Specific works address the client-server communication issues (Toffetti, 2007). Other approaches (Brambilla, 2008) exploit workflow modeling for achieving a larger separation of concerns regarding (i) the data and business logic distribution, (ii) the interface behaviour, (iii) the hypertext navigation, and (iv) the presentation aspects. The proposal by Kadri (2007) exploits the UML notation and hierarchical design of components for specifying complex (possibly distributed) Web applications including rich in-

99

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

terface behaviours. The proposal is provided with a design and code generation tool. However, the focus is more on the application structure than on the GUI behaviour. The RUX-Method (RUXProject, 2008) proposes a more clear separation of the presentation at design-time through a stack of models specifically designed for RIAs. Other recent proposals in the Web Engineering field represent the RIA foundations (e.g., Urbieta, 2007) by extending existing Web engineering approaches. Urbieta et al. (2007) suggests a design approach that extends OOHDM with a good separation of concerns and a UML-like notation. Some works offer insights and experiences on the migration of traditional desktop or client-server applications to Web based application that exploit rich interfaces. Among them, Samir et al. (2007) proposes an approach to translate Java Swing applications to XUL applications.

The WebML methodology and language are fully supported by the CASE tool WebRatio 5.0 (2008), an Eclipse plugin representing a new generation of model driven development (MDD) and engineering (MDE) tools for Web applications. Besides taking advantage of the Eclipse features, WebRatio provides advanced capabilities in terms of support of model extensions, model checking, code generation, project documentation, and collaborative work. •

webMl and webratio background The WebML language and methodology is a high-level notation for data-, service-, and process- centric Web applications. It allows specifying the data model of a Web application and one or more hypertext models that can be based on business process specifications and can exploit Web service invocation, custom backend logic, and rich Web interfaces. The WebML approach to the development of Web applications consists of different phases. Inspired by Boehm’s spiral model, the WebML process is applied in an iterative and incremental manner, in which the various phases are repeated and refined until results meet the application requirements. The WebML language is a Domain Specific Language (DSL) for designing Web applications. This section summarizes the basic WebML concepts, with particular attention to data model and hypertext model.

100



WebML Data Model. For the speciication of the underlying data of the Web application, WebML exploits the existing EntityRelationship data model, or the equivalent subset of UML class diagram primitives. The data model can also include the speciication of calculated data. Calculated attributes, entities, and relationships are called derived and their computation rule can be speciied as a logical expression written using declarative languages like OQL or OCL. WebML Hypertext Model. The hypertext model enables the deinition of the frontend interface of the Web application. It enables the deinition of pages and their internal organization in terms of components (called content units) for displaying content. It also supports the deinition of links between pages and content units that support information location and browsing. Components can also specify operations, such as content management or user’s login/logout procedures (called operation units).

A site view is a particular hypertext, designed to address a specific set of requirements. It consists of areas, which are the main sections of the hypertext and comprises recursively other sub-areas or pages. Pages are the actual containers of informa-

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

tion delivered to the user. Several site views can be defined on top of the same data schema, for serving different user roles or devices. Pages inside an area or site view can be of three types: the home page (“H”) is the default address of the site view; the default page (“D”) is the one presented by default when its enclosing area is accessed; a landmark page (“L”) is reachable from all the other pages or areas within its enclosing module. Pages are composed of content units, which are the elementary components that publish pieces of information within pages. In particular, data units represent some of the attributes of a given entity instance; multidata units represent some of the attributes of a set of entity instances; index units present a list of descriptive keys of a set of entity instances and enable the selection of one of them; scroller units enable the browsing of an ordered set of objects; entry units allow to publish forms for collecting input values from the user. Units are characterized by a source (the entity from which the unit’s content is retrieved) and a selector (a restriction predicate on the result set of the contents). Units and pages are interconnected by links, thus forming a hypertext. Links between units are called contextual, because they carry some information from the source unit to the destination unit. In contrast, links between pages are called non-contextual. Different link behaviours can be specified: an automatic link (marked as “A”), is automatically “navigated” in the absence of a user’s interaction when the page is accessed. A transport link (dashed arrow), is used only for passing context information from one unit to another and thus is not rendered as an anchor. Parameters can be set as globally available to all the pages of the site view. This is possible through global parameters, which abstract the implementation-level notion of session-persistent data. Parameters can be set through the Set unit and consumed within a page through a Get unit.

WebML also supports the specification of content management, custom business logic, and service invocation. WebML offers additional primitives for expressing built-in update operations, such as creating, deleting or modifying an instance of an entity (represented through the create, delete and modify units, respectively), or adding or dropping a relationship between two instances (represented through the connect and disconnect unit, respectively). Other utility operations extend the previous set. Operation units do not publish the content to be displayed to the user, but execute some business logics as a side effect of the navigation of a link. Each operation can have two types of output links: the OK link is followed when the operation succeeds; the KO link when the operation fails. Like content units, operations may have a source object (either an entity or a relationship) and selectors, and may have multiple incoming contextual links, which provide the parameters necessary for executing the operation. Two or more operations can be linked to form a chain, which is activated by firing the first operation. Figure 1 shows the WebML model representing a Web site area (Product area) marked as Landmark (L), comprising two pages: Products page (default page of the area) contains the index of All products available in the database and a form. By clicking on the index, the used follows the Details link to the Product Details page, where the information about the selected product is shown. By submitting information through the New product form, the user triggers the execution of the Create product unit, which instantiates a new Product tuple in the database. In case of success, the OK link is followed and the Product details of the newly created element are shown; in case of failure, the KO link is followed toward the Product page.

101

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

Figure 1. The WebML model for uploading a file with asynchronous behaviour using AJAX

Modeling PriMitives for riAs As previously mentioned, RIAs enable a wide set of user interaction patterns that mimic the behaviour of desktop application interfaces. In this section we examine the most common interactions and describe how they can be specified with the WebML conceptual modeling language. A comprehensive specification of the possible interaction patterns can be found in Preciado (2007). In this chapter, the discussion will highlight how the conceptual modeling of the new RIA behaviours differs from traditional web application design. The main areas where AJAX mechanisms can be applied are the refinement of navigation of links, content publishing in the pages, and user input management. If any of these aspects is expected to be managed with AJAX, at the WebML modeling level the involved objects (pages, units, links) must be marked with the property “AJAX enabled”.

Partial Page refresh Thanks to the new features introduced by AJAX, links can represent new kinds of interactions in the pages. In particular, it is possible to define

102

partial refresh of pages and management of separate popup windows with respect to the main application interface. The partial page refresh consists in the possibility of reloading only some portions of the page, thus making the user interaction with the application quicker and more effective. The (partial) refresh is usually caused by a user click on the interface, although it could be in principle activated by any kind of user event. The behaviour of the partial page refresh is fundamental in the modeling of RIAs. Therefore, its representation at the modeling level must be straightforward. Webratio allows one to describe this feature simply by marking the link that triggers the page refresh as “AJAX”. In this case, by clicking on the AJAX link the user will activate the refresh only of the target of the link (a page, subpage, or group of content units) that are affected by the navigation of the link. The calculation of page contents, of dependencies among units, and of partial page refresh criteria is a complex task. The visual WebML model relies on a solid page computation algorithm that manages all these issues. The main aspect to be considered is the definition of which parts must be recalculated after a user interaction. WebML speci-

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

fies calculation semantics (which is implemented in the WebRatio tool too) for its models that relies on the topology of the links: basically, any unit with an incoming link can be seen as dependent with respect to the source unit of the link. When a user submits or selects a value in a unit, all the (recursively) dependent unit must be recomputed. In a traditional Web 1.0 approach, the full page is refreshed anyway, while in a RIA application only the dependent units must be refreshed. Figure 2 shows a page with an AJAX link that allows the user to select a customer from a list. The effect of clicking on the link is that the page is partially refreshed: the Customer details and Order list units are refreshed, because by clicking on the AJAX link the user triggers partial page refresh of all the dependent components. In the example, the content of Customer details and Order list units depend on the selection of the user. Indeed, they are both connected by (possibly

indirect) links to the Customer List unit, where the user interaction took place. Notice that Text and Customer list units are not refreshed, since they are not affected by the user selection.

on-Page Popup and window Management Another important feature related to the page management is the possibility of defining AJAX Windows within the Web application. Windows can be opened as popups or message boxes in the interface upon user events. AJAX Windows can be defined as modal or not; modal windows do not allow one to interact with the other parts of the application while they are opened. From an implementation point of view, AJAX windows are placed on a new layer above the page, thus automatically disabling the user interaction on the main page. When the popup window is closed,

Figure 2. The WebML model for a partial refreshing page

103

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

the content of the main page becomes active again and the user can browse the application. The WebML model of Figure 3.a represents the navigation from a list of customers (AJAX Window page) to a popup Modify Customer popup window that displays a form containing the data of the selected customer and allowing to modify the details. The popup window is opened by the AJAX link from the Customers List unit. Once opened, the Modify Customer popup page retrieves the current Customer value and shows a form for updating such values. The popup is closed once the user clicks on the outgoing AJAX link towards the ModifCustomer operation, that perform the database updates on the customer instance retrieved through the dashed transport link coming from the Customer unit. The rendering of the page is shown in Figure 3.b: the main page contains the Customers List, while the AJAX popup contains the form Customer Data for updating the information of the selected customer.

dynamic tooltips on Page data The tooltip feature allows one to render some information in a dedicated area that is loaded when the user selects an element of the page, without the need to reload the whole page. Tooltips can be shown upon a user action over a page object. At the modeling level, this can be specified with a simple notation: the tooltip behaviour is associated by a set of properties with the unit where the tooltip is expected to appear. Those properties include: the anchor of the tooltip, i.e., the position where the tooltip is activated (for instance, the link anchors or the attribute values shown in the unit); the triggering event (mouse click, mouse double click, mouseOver, …); the tooltip size (possibly dynamic on the size of the content); the options for drag and drop of the tooltip; and so on. Once this has been specified, a link exiting the unit can be marked as a tooltip link, representing that the activation of the tooltip concretely consist of traversing that link. The actual tooltip content

Figure 3. The WebML model of a popup window for editing the customer information

104

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

consists of a full WebML page and therefore can be dynamically extracted from a datasource by any set of WebML units. To make a page behave as a tooltip, one of its units must be the destination of the tooltip link. Figure 4 a shows a simple example of tooltip usage: a page contains an index of Customers, which is enabled for tooltips. The outgoing link activates the tooltips upon event. The link leads to a page containing an index that fetches from the database the list of orders previously submitted by the current customer. If we specify that the triggering action is the OnMouseOver event on the attribute values of the index, the content of the Order List page will be dynamically shown once the user rolls over the values of the Customers List index unit. Figure 4 b shows the rendered page with the tooltip window opened for the current customer.

drag and drop among Content Components The Drag and Drop feature allows one to perform some operations in a web page by dragging some elements of the page on others. After data objects are selected and dragged, the drop action causes the execution of any associated side effect. This is a powerful interaction paradigm that can replace the traditional selection of objects in the page, making the interaction more intuitive in several contexts, such as adding a product to the cart, moving an element to a specific folder, and so on. This behaviour can be modelled simply by specifying that a WebML unit is enabled for the dragging event. Then, the outgoing links marked as AJAX links will behave as drag&drop paths. A drag&drop path is defined as a link from a source unit to a side effect operation (or set of operations) that is performed when the drop event occurs.

Figure 4. The WebML model for a dynamic tooltip on an index

105

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

Figure 5. The WebML model for a drag and drop behaviour for an ecommerce cart

Figure 5 shows an example of drag and drop in a simple ecommerce application. The user may reach the Add to cart page by some other page in the site through the incoming arrow in the topleft angle of the figure. There, the user can see the list of available products (Product List unit) and can drag and drop them into his own cart, represented by the Order summary data unit. The drag link is represented by a symbol of moving mouse pointer, and the effect of the dragging is represented by the Connect product unit. The other link that reaches the Connect product unit is not meant to be navigated, but instead associates the recipient of the dragging (i.e., the Order Summary unit) to the dragging action. Therefore, notice that the drop event can happen only on the Order summary unit. In general, the allowed component for dropping objects can be identified because it is connected to the component that is the destination of the drag link. The drag link in the example leads to the Connect product operation, that connects the dropped product to the Order. Once Order and Product are connected, the user is redirected to the page through the OK link exiting the Connect product operation.

106

event Management and dynamic dependencies in the Page While traditional Web applications only rely on the onClick event for the behaviour of the user interface, AJAX allows to handle a much wider set of events, thus making the user interface more usable. Thanks to AJAX events, the designer can specify the actions to be performed by the system after the occurrence of a specific event, usually associated to form fields. The elementary events that can be managed are: •





onChange event: allowing one to deine update policies among preloaded ields; the typical application is the refresh of the contents of a drop down list depending on the changed value of another ield; onFocus event: allowing one to calculate and show some contents when the user moves the cursor to a speciic ield; this may be useful for showing some instructions or hints when the user enters a ield; onBlur event: allowing one to execute some action when the cursor leaves a ield;

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

typically, this is used for applying validation checks on the input data. These options can be speciied as ield properties: each ield can be marked as sensitive to a speciic event and can be associated with an outgoing link that manages this event.

ateCustomer unit, which actually creates a new Customer instance in the database (including the city information). Analogous behaviours can be defined by exploiting the other kinds of events mentioned above.

Figure 6 shows a simple WebML page aiming at adding a new customer in a database. In this page, the set of existing countries is immediately extracted (by the Countries query unit) from the database and used for populating a dropdown field in the Customer Data entry unit. The field is enabled for catching the OnChange event, associated to the highlighted link. When the user makes his choice for the country of the customer, the OnChange link is triggered and the list of available cities is recalculated based on the selection (by the CitiesOfCountry query unit). Therefore, the choice of the customer city is limited to those belonging to the country previously selected. Once the user finally selects a city, he triggers the Cre-

Autocompletion consists of a set of suggestions that is shown to the user while he types textual contents in some field of a form. In this way, the user can type only the initial letters of the word or of the value that has to be inserted in the field and the application shows different options among which the user can choose. The AJAX autocomplete feature allows one to specify an information source that is used as suggestion. This can be specified by setting the AJAX Autocomplete property on the interesting fields of the form; then for each autocomplete field one outgoing link marked as AJAX Autocomplete identifies a page that is rendered within the autocomplete drop down menu.



field Content Autocompletion

Figure 6. The WebML model for managing the OnChange event on the customer data submission

107

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

Figure 7. The WebML model enabling autocompletion on the customer surname field of a form

Figure 7 a shows a page containing a search form for customers. The form has an outgoing link toward the Search Result index that displays all the customers whose surname matches the input of the user. Besides the traditional search behaviour, the form is equipped with the autocompletion facility represented by the Ajax link leading to the Customer suggestion page. This link associates a query that extracts the list of customers beginning with the letters typed by the user. This list is shown as a set of autocompletion options to the user. This behaviour can be generalized by exploiting complex, possibly multiple, dependencies between fields. A typical scenario is autocompletion of geographical information. Once a country is specified in a field, the autocompletion of the city field will consider both the value of the country field and the partial text that the user is entering as city name. Figure 7 b shows the online application that displays the autocompletion suggestions when the user has typed the “a” character in the field.

108

Management of background file Uploads The AJAX file upload feature allows the user to upload files into a Web application with asynchronous behaviour, allowing the interaction with the application during the file upload. In standard Web applications this is not possible: the user has to wait until the upload operation is finished before continuing the navigation. To implement this particular feature there is a technological limitation in Javascript that must be solved: JavaScript does not allow access to the local file system resources for security issues. This means that no selection and upload of local files can be performed. The AJAX file upload can only be implemented by defining a separate page containing the upload field, which is then included as an iframe in the main page, so that the upload request is processed independently from the rest of the application. To capture this behaviour, two different pages must be defined. The first page represents the iframe containing the upload field and the upload operation. The second page represents the main application page from which the user will start the file upload.

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

Figure 8. The WebML model for uploading a file with asynchronous behaviour using AJAX

Figure 8 shows the WebML model representing a background upload, supposing to have one entity called File in the data model, with a BLOB attribute. The AJAX Home page is the main Web site page, which embeds the upload iframe, while the Upload page models the iframe containing the actual uploading. The Home page contains an AJAX form with a single field. This form is associated through a property to the Upload page iframe, that will be properly rendered in the page. The Upload page contains the Upload form with the actual upload field, which triggers the Create file operation. The same page contains also the list of files uploaded up to now (File Index unit). Once the user clicks on the Submit button in the Upload form, the link is followed and the Create File operation is performed.

webrAtio design tool And rUntiMe ArChiteCtUre WebRatio 5 (2008) is an Eclipse plug-ins that fully supports design and development of Web applications based on the WebML language and methodology. The tool provides a set of visual editors for the WebML models, some model checking and design facilities (wizards, property panels, and so on) for the developer, and a set of code

generators for producing the running application. The full description of the tool is available in Acerbis (2008).

design facilities for AJAX Properties At design-time, the WebML editors allow one to model the application and to save it as an XML project. The WebRatio Eclipse perspective comprises several panels: visual model editors, advanced text editors, form-based editors components and properties, wizards, and documentation editors. The editing of the AJAX features can be performed mainly through the property panels of the WebML visual components. Figure 9 shows three examples of property panels for setting AJAX features. Figure 9 a sets the tooltip behaviour of a unit: in this example, the event is OnMouseOver; the active position is set on the attributes; and the link to be followed is the See orders link. This means that when the user moves the mouse over any content attribute shown by the unit in the page, a tooltip appears displaying the contents of the page reached by the See orders link. Figure 9.b, referring to the example in Figure 7, enables the autocompletion of a field and defines the link to the page containing the autocomplete hints (Surname Autocomplete). Finally, Figure 9.c, referring to the example in Figure 6, activates the OnChange

109

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

Figure 9. AJAX property panels for describing various RIA behaviours within WebML components

event on the Country field and specifies the Link to be followed when the event is triggered. The WebML models are enriched by these specifications and can hence be used as starting point for automatic AJAX code generation.

runtime Architecture The run-time WebRatio framework exploits a set of off-the-shelf object-oriented components for organizing the business tier, as shown in Figure 10.

For every page, there is one main JSP template in charge of displaying the full page interface with the needed contents, and one auxiliary JSP template, which contains the data elements to be retrieved by AJAX. Every kind of WebML unit is associated to one service class. At runtime a single service class is deployed for each type of unit and one runtime XML descriptor is generated for each unit instance used in the application design. For instance, every Data unit in the project will be executed by the same Data unit component, which will be configured by several descriptors (one for

Figure 10. Runtime view of Java components and XML descriptors for WebRatio units

110

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

each actual Data unit instance), to provide the respective functionalities.

AJAX templates organization Code generation and runtime support of AJAX features have been developed by means of a mix of opensource technologies, extensively exploiting the XMLhttpRequest method. Namely, the AJAX toolkits of choice are Prototype (2008) and Scriptaculous (2008). Their characteristics and the comparison with other solutions are presented in the Background Section. The adopted runtime architectural solution consists of implementing each conceptual hypertext page with two dynamic pages that interact through XMLhttpRequest for providing the rich interface features: the first page is a back-end JSP page that stores the chunks of page that could be requested by an AJAX request; the second page is the front-end JSP page (including the required JavaScript needed for event management) that is shown to the user. The latter invokes extraction of pieces from the back-end JSP page according to the user behaviour, and shows the results and the interface options to the user. Figure 11 pictorially represents this behaviour.

As described in the modeling section, special attention must be devoted to asynchronous file upload. Indeed, it requires two distinct dynamic templates. The main template contains a form with the reference to the subpage for the actual upload, which is hence included. The two pages involved in the AJAX upload are generated separately by the code generator. The subpage is generated as a page without any menu and header, and then is included as an iframe in the main page. This is possible simply by specifying that the page graphical style is the Empty page layout, which prints only the page content.

Automatic Code generation The WebML models are transformed into running code by the WebRatio automatic code generation modules. WebRatio code generators produce J2EE Web applications starting from the WebML models. They are developed using the ANT, XSLT, and Groovy technologies. Groovy is a light weight language using a Java-like syntax and fully integrated in the Java Platform, since it is actually translated to Java programs before being executed. It provides many features and facilities that are inspired by scripting languages,

Figure 11. Sequence diagram describing the manipulation of page contents on AJAX events

111

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

but also allows one to exploit all the existing Java libraries. The generated code includes all the XML descriptors of the units, the front-end JSP pages including the Javascript for the AJAX features, and the back-end pages used as a source of data to be shown upon AJAX events. Besides existing WebML primitives, it is possible to specify new components (i.e., new WebML units), that can in turn be AJAX-based, and include them in the application design and code generation framework. A custom WebML unit consists of a component made of: • •

a Java class that implements the service of the component; and a set of XML descriptors, deining the component interface in terms of inputs and outputs.



The application is therefore a facility for (i) identifying the emerging trends in the reference market and in other related markets; (ii) capturing and structuring the data resulting from these trends; (iii) interpreting and sharing the findings; and (iv) feed the new results to the creative people for the definition of new ideas. Eight main use cases are identified: •



riA indUstriAl CAse: eKrP • The presented approach has been applied in the development of several industrial applications that required flexible and user friendly interface. We present now one of these applications, called eKRP (electronic Knowledge Repository Process), to demonstrate the feasibility and effectiveness of the proposal. The eKRP application has been developed for a primary Italian textile enterprise, well known at European level for its home textile production, which includes curtains, linens, towels, and bathrobes.



requirements The eKRP Web application aims at providing the research and development division of the company a tool for:







112

the technological and social analysis of the state of the art and of the user requirements, based on market analysis, on

competitor and internal products, and on user’s opinions; the management of the creative process of invention and of concept deinition.

User proile development: the administrator can dynamically manage and organize user proiles and categories, and apply different access rights to different proiles; Stimuli proiling: users can dynamically deine the proiles for the various stimuli, i.e., inputs coming from the market, that can feed the creativity process; Data entry: users must be able to access content management interfaces for the deined stimuli; Cluster development: users can cluster the stimuli, thus aggregating inputs based on the expertise and sensitivity of the speciic persons, by means of visual diagrams and interactive graph editing. This leads to a irst mapping of the interesting areas and trends. The management of the clusters must be provided at two levels: a view “in the large” allows one to visualize and edit the position of the clusters, while a view “in the small” allows one to manage the internal structure of a cluster, in terms of trends and stimuli that pertain to the cluster; Polarization development: users can create and manage the polarization of the clusters that are considered strategic for the company; Orientation and design direction: users can manage and organize the processes of creative development of new ideas;

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX



Backofice management: users can access the usage history of the eKRP system and can check the system performances in terms of measurable advantages on the resulting products.

design The whole eKRP application has been developed using WebRatio. All the interface requirements have been fulfilled, thanks to a deep use of AJAX features throughout the project. The design and development followed the steps specified by the best practices of Web engineering, according to the WebML methodology (Ceri, 2002). The resulting design consists of two siteviews, one devoted to the company user and one for the administrator: the overall application is organized

in 79 pages and 1265 WebML units. Around 60 pages incorporate some AJAX feature. Some new custom components (units) have been developed too: the ErrorsCheckUnit (for checking the correctness of inputs coming from multiple and possibly alternative forms), the ThumbnailCreateUnit (for automatically generating thumbnails from uploaded images), and the WGetUnit (for downloading and storing a Web page, including all its resources, such as css, javascript, images, and so on). The peculiar aspect of the developed application is the sophisticated interface of most pages. This reflects a typical trend of RIAs: pages tend to be more and more complex, while their number decreases (because several functions are gathered into one page). For space reasons, the design cannot be reported entirely in this chapter. To give a flavour

Figure 12. WebML model of the Manual Clustering page in the eKRP application

113

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

of the features that can be obtained with the AJAX components, we show in Figure 12 the WebML model of the Manual Clustering page, one of the most complex of the eKRP application. The page allows to manually create and modify clusters of stimuli. The main function of the page is provided through the search for available stimuli within the system, according to some criteria (Chapter, Argument, and Keyword). The search is implemented by the Manual Clustering entry unit, the component (1) in Figure 12, where the user submits the search criteria. The form is equipped with full-fledged AJAX validation and comprises many dynamic dependencies among fields. The submitted keywords are split into separated strings by the Split keywords unit (2) and are fed to the parametric query, together with the other criteria. The results of the search are displayed in the Search results unit (3), that shows all the stimuli match-

ing the criteria. The user can define (or redefine) a cluster by dragging one or more results to the Working on cluster data unit (5). This is obtained through the link (4), which is associated with the drag event. When the user drops an element in the data unit (5), the side effects defined for the link (4) are performed.

implementation The implementation of the application completely relies on the code generation features of WebRatio: a visual style has been defined for satisfying the requests of the customers, and has been applied by the code generators. The only handwritten code is the one needed for the ad hoc business logic of new custom units. Figure 13 shows the Manual Clustering page interface, corresponding to the model shown in

Figure 13. Snapshot of the Manual Clustering page interface in the eKRP application

114

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

Figure 12. The top central area (1) of the page represents the form that collects the search criteria (in terms of Chapters, Main arguments, and Other arguments). When the user clicks on the search button, the query is performed and the resulting stimuli are shown in the subpage (3). Stimuli can be dragged from the subpage (3) to the Working on cluster panel on the left (5), thus redefining the members of the cluster. Other pages are then devoted to the visualization of clusters. For instance, Figure 14 shows a page with a graph of clusters in the top central part of the screen. This graph is implemented by a custom AJAX-enabled unit that allows one to show, edit, and save the set of clusters.

ConClUsion This chapter presented the features available in the WebML language for supporting RIA behaviour

of Web applications, based on AJAX extensions. The primitives have been implemented in the WebRatio design tool, which now features visual modeling of AJAX characteristics and automatic generation of the code. The proposed approach is very pragmatic and relies on relatively simple design abstractions, if compared to more comprehensive design proposals. The reason of this choice stands on the need of a practical and efficient way for describing the typical RIA interactions. In addition, the approach based on simply extending the WebML features with a set of properties of the standard objects allows full backward compatibility with respect to traditional WebML models and do not require significant effort for the WebML designers to become accustomed to the new features. Another need covered by the approach is the relatively easy and quick implementation of the new features, both at the design level (in terms of new panels, properties, and primitives in the hyper-

Figure 14. Snapshot of the Cluster Visualization page interface in the eKRP application

115

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

text model editor) and at the code generation level (in terms of updates to the code generator). Thanks to the model-driven design approach adopted in our proposal, good separation of concerns between conceptual components and implementation details is achieved. Indeed, the conceptual components specified in WebML and the corresponding interaction paradigms presented in Section 4 are not bound to the AJAX technology. They represent general RIA behaviours that could be implemented in any technology. Therefore, components (and existing application models) could be reused without any change in case of technology switching. It would be enough to reimplement the runtime classes of the components in the new platform of choice (i.e., other AJAX libraries or different paradigms, like Laszlo, Flex, or XUL) and to regenerate automatically the existing applications. The result would be an application implemented in the new technology. Currently, the approach is already adopted for the implementation of real enterprise Web applications, as shown in the discussed case study. The percentage of developed applications that include AJAX features is continuously increasing since the first day of availability within the tool. This is a clear symptom of the success of the RIA interfaces among the users, which definitely calls in the Web Engineering field to develop comprehensive design facilities for this kind of applications.

Bindows. (2008). Retrieved from http://www. bindows.net/ Bozzon, A., Comai, S., Fraternali, P., & Toffetti Carughi, G. (2006). Conceptual modeling and code generation for rich Internet applications. In Proceedings of ICWE 2006, International Conference on Web Engineering (pp. 353-360). ACM Press. Brambilla, M., Preciado, J. C., Linaje, M., & Sanchez-Figueroa, F. (2008). Business processbased conceptual design of rich Internet applications. In Proceedings of ICWE 2008. Yorktown Heights, USA: IEEE Press. Brent, S. (2007). XULRunner: A new approach for developing rich Internet applications. IEEE Internet Computing, 11(3), 67–73. doi:10.1109/ MIC.2007.75 Ceri, S., Fraternali, P., Bongio, A., Brambilla, M., Comai, S., & Matera, M. (2002). Designing data-intensive Web applications. San Francisco, CA: Morgan Kauffmann. Daniel, F., Yu, J., Benatallah, B., Casati, F., Matera, M., & Saint-Paul, R. (2007). Understanding UI integration: A survey of problems, technologies, and opportunities. IEEE Internet Computing, 11(3), 59–66. doi:10.1109/MIC.2007.74 Dojo. (2008). Retrieved from http://dojotoolkit. org/ Ext. (2008). Retrieved from http://extjs.com/

referenCes Acerbis, R., Bongio, A., Brambilla, M., Butti, S., Ceri, S., & Fraternali, P. (2008). Web applications design and development with WebML and WebRatio 5.0. In Proceedings of TOOLS Europe 2008. (LNBIP 11, pp. 392–411. Springer. Backbase. (2008). Retrieved from http://www. backbase.com/

116

Google. (2008). Google Web toolkit. Retrieved from http://code.google.com/webtoolkit/ Google. (2008). GWT-Ext. Retrieved from http:// code.google.com/p/gwt-ext/ JackBe. (2008). Retrieved from http://www. jackbe.com/ jQuery. (2008). Retrieved from http://jquery. com/

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

Kadri, R., Tibermacine, C., & Le Gloahec, V. (2007). Building the presentation-tier of rich Web applications with hierarchical components. In Proceedings of WISE 2007, Web Information Systems Engineering. (LNCS 4831/2007, pp. 123-134). Springer. ISSN 0302-9743, ISBN 9783-540-76992-7. Linaje, M., Preciado, J. C., & Sánchez-Figueroa, F. (2007). Engineering rich Internet application user interfaces over legacy Web models. IEEE Internet Computing, 11(6), 53–59. doi:10.1109/ MIC.2007.123 MooTools. (2008). Retrieved from http://mootools.net/ Mozilla. (2008). XUL. Retrieved from http://www. mozilla.org/projects/xul/ MyGWT. (2008). Retrieved from http://mygwt. net/ Preciado, J. C., Linaje, M., Comai, S., & SanchezFigueroa, F. (2007). Designing rich Internet applications with Web engineering methodologies. In Proceedings of International Symposium on Web Site Evolution (pp. 23-30). IEEE Press. Preciado, J. C., Linaje, M., & Sánchez-Figueroa, F. (2007). An approach to support the Web user interfaces evolution. In ICWE Workshop on Adaptation and Evolution in Web Systems Engineering (pp. 94-100). Springer. Preciado, J. C., Linaje, M., Sánchez-Figueroa, F., & Comai, S. (2005). Necessity of methodologies to model rich Internet applications. In Proceedings of International Symposium on Web Site Evolution (pp. 7-13). IEEE Press. Prototype. (2008). Retrieved from http://www. prototypejs.org/ RUXProject. (2008). Retrieved from http://www. ruxproject.org/

Samir, H., Stroulia, E., & Kamel, A. (2007). Swing2Script: Migration of java-swing applications to Ajax Web applications. In Working Conference on Reverse Engineering 2007 (WCRE 2007) (pp. 179-188). ISSN: 1095-1350, ISBN: 978-0-76953034-5. Schwabe, D., Rossi, G., & Barbosa, S. (1996). Systematic hypermedia design with OOHDM. In 7th ACM International Conference on Hypertext (pp. 116-128). Washington, D.C.: ACM Press. Scriptaculous. (2008). Retrieved from http:// script.aculo.us/ Tatami. (2008). Retrieved from http://code.google. com/p/tatami/ Tibco. (2008). Tibco general interface. Retrieved from http://gi.tibco.com/ Toffetti Carughi, G., Comai, S., Bozzon, A., & Fraternali, P. (2007). Modeling distributed events in data-intensive rich Internet applications. In Proceedings of International Conference on Web Information Systems Engineering (pp. 593-602). Urbieta, M., Rossi, G., Ginzburg, J., & Schwabe, D. (2007). Designing the interface of rich Internet applications. In Proceedings of Latin-American Conference on the WWW (pp. 144-153). IEEE Press. WebRatio. (2008). Retrieved from http://www. webratio.com/ Yahoo. (2008). Yahoo user interface library. Retrieved from http://developer.yahoo.com/yui/

AdditionAl reAding W3C (2007). XMLHttpRequest, http:// www.w3.org/TR/2007/WD-XMLHttpRequest-20070227/ W3C (2008). Rich Web Clients Activity, http:// www.w3.org/2006/rwc/Activity.html

117

A Tool for Model-Driven Design of Rich Internet Applications Based on AJAX

W3Schools (2008), AJAX Tutorial, http://www. w3schools.com/Ajax/Default.Asp Adobe, Flex Developer Center, http://www.adobe. com/devnet/flex/ Curl (2008), RIA Knowledge Center, http://www. curl.com/knowledge-center/ Macromedia (2002). Requirements for Rich Internet Applications, http://download.macromedia. com/pub/flash/whitepapers/richclient.pdf

Key terMs And definitions Asynchronous JavaScript and XML (AJAX): Set of Web technologies and development techniques used for developing RIAs. Asynchronous client-server interactions are achieved thanks to the XMLHttpRequest object. Automatic Code Generation: Software engineering technique that allows to automatically generate application code starting from (platform independent) conceptual models.

118

Model Driven Development: Software development approach based on the systematic use of models and their transformations throughout the engineering lifecycle of a software system. Partial Page Refresh: Possibility of refreshing single pieces of a Web page upon the occurrence of an event. Rich Internet Application (RIA): Web application that implement sophisticated user interfaces, and advanced user interaction patterns with respect to traditional Web applications, including partial page refresh, client-side calculation and data storage, drag&drop, and other features Web Engineering: Scientific discipline studying models, methodologies, tools, techniques, and guidelines for the design, development, evolution, and evaluation of Web applications Web Modeling Language (WebML): Conceptual model and methodology for the visual design of data-intensive, process-intensive, and service-intensive Web applications WebRatio: CASE (Computer Aided Software Engineering) tool for the specification of Web applications according to the WebML modeling language and for the automatic code generation.

119

Chapter 7

Web 2.0: Self-Managing System Based on SOA Model and Grid Computing Overlay Wail M. Omar Sohar University, Sultanate of Oman

AbstrACt Web 2.0 is expected to be the next technology in the interaction between the enterprise applications and end users. Such interaction will be utilized in producing self-governance applications that are able to readjacent and reconigure the operation framework based on users’ feedback. To achieve this, huge numbers of underneath resources (infrastructures and services) are required. Therefore, this work proposes the merge of Web 2.0 technology and grid computing overlay to support Web 2.0 framework. Such merge between technologies is expected to offer mutual beneits for both communities. Through this work, a model for managing the interaction between the two technologies is developed based on the adapting of service oriented architecture (SOA) model, this model is known as SOAW2G. This model manages the interaction between the users at the top level and resources at the bottom layer. As a case study, managing health information based on users’ (doctors, medicine companies, and others) experiences is explored through this chapter.

1. introdUCtion Web2.0 is considered as the next era of the interaction between the web applications and users, where Web2.0 offers the framework for community-based collaborative. The community can be varying based on the community patterns, behaviours and functions, i.e. health community, engineering comDOI: 10.4018/978-1-60566-384-5.ch007

munity, software community, mobile community, application community, intelligent services community and others. Therefore, the Web2.0 framework is the merge of users’ experiences, feedback and services with the system resources (services and infrastructures), like wikis, blogs, RSS...etc. Therefore the interaction between the communities (users) at the top level with resources at the bottom level should be managed in a way that enhances performance, reliability, fidelity, and security of the

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Web 2.0

framework. Such system requires dynamic model that has the ability to manage and re-manage or re-adjacent the underneath resources based on the experiences and feedback from the users in order to provide better services. Web 2.0 successes in supporting different types of complex web applications (Web2.0 sites) (O’Reilly, 2005) that are available nowadays, such as Google Maps, YouTube.com, flicker, and many other irrefutable websites and web applications that are built up using the users’ participations themselves. Despite the success that Web2.0 idea has achieved so far, most of the experts see that it can be pushed and developed further forward in the direction of creating large scale enterprise applications. This is achieved through the use of other smaller components and user contributions (SematicGrid). Others are dreaming of the idea of creating customized application on the fly by utilizing the same concept. These ideas and ambitions are so alive now a days and are breaking their paths through to existence, but one of the main obstacles in creating such a thing is the need for large computational resources that are not easily available to everyone(O’Reilly, 2005; SematicGrid). To overcome the problem of resources lacking, the use of grid computing has been proposed in this work. In this chapter, grid computing offers the fabric for deploying different types of resources including applications and general services, infrastructures, monitoring and controlling systems and others. On the other hand, Web2.0 can be used to provide the Grid community with high quality and survivable services from the users’ contributions which generate self-manageable and self-governor framework that is able to reshape the application and resources based on the interaction with environment. In this work, we will try to answer the questions of “can the two concepts be combined to achieve a mutual assistance to each other in a crucial step toward the futuristic information technology

120

world?”. Which leads to another question: “can Web2.0 and grid computing boost each other?”. To answer these questions, a model for managing the interaction between the users at the top level and grid resources at the fabric layer based on the use of Services Oriented Architecture (SOA) model has been adopted and developed. The SOA has been adopted here to offer a platform for controlling the interaction between different layers of the system. Principally, our approach for the merging process is to invoke and use grid computing resources through developed Web2.0 framework that have been already developed based on SOA philosophy (Wail M. Omar, May/Jun, 2007) in a manner that would make this invocation seamless to users and other agents, so the semantic feature would be for all resources, infrastructures and services adequately. But does that means both technologies are strongly mixed together in this SOA based framework in a way that they will be one solid structure that cannot be dissolved into its basic two components? The answer is definitely no, because simply both technologies are still wobbling and yet in forming phases, to a different levels for both, and of course in different development paces for each one, so one of our main goals that we have focused on in this formation is the scalability of the framework and its two main components (Grid & Web2.0) and our vision about achieving that is to provide upper layers for both technologies that can hide the underneath complexities and structures, then we begin shaping our merging structure by working these high level layers. This chapter is structures as follow: introduction is described in section 1. Then, the background is discussed in section 2, followed by SOAW2G model in section 3. How to use the model in action is explained in section 4. Then, describing for the Resources Markup Language (RML), classifying resources at the bottom layer with a case study of classifying drug information is coming in section 6 with case study on classifying the health

Web 2.0

information resources, ending with conclusion and future work in section 6.

2. bACKgroUnd There are many web applications appear based on the use of Web2.0 technologies, such as face book and YouTube. In 2006, Hogg et al. (C. Schoth, 2007; R. Hogg, 2006) Conducted an in-depth investigation of 40 successful Web2.0 applications. They condensed their respective characteristics to describe the phenomenon of Web2.0 communities and to provide a systematic overview of current and emerging business models. Such applications show the important of the use of Web2.0. To achieve the future vision of how Web2.0 will support the enterprise applications, grid computing is proposed to be an attractive solution for offering different types of resources. However, till now there is no clear vision of how Web2.0 can merge with grid computing in order to get benefit from the powerful resources that available in the grid community. This chapter is trying to focus on bringing the two technologies together in a way that each one boosts the other. No much work has been conducted regarding the linkage between grid and Web2.0. However there are some efforts to imitate Web2.0 concepts and try to adapt them into Grid technology. The leader group in that matter is called semantic grid group (SematicGrid). The main notion in Web2.0 that can be extended is the semantic web, so their vision about the grid is called the semantic grid and it relies on revolutionizing the resources to a process able common knowledge that can be understood and dealt with by all parties taking part to create a specific grid, which will yield in a drastic improvement in the dynamics of grid technology. As has been described before, the merge between grid computing and Web2.0 will return benefit for both by producing a robust framework for enterprise application. Such framework merges

between the high quality and availability of resources through the use of grid computing and the knowledge that can gain from the users to enhance the operational framework. The framework is also required an architecture that control the follow of processes and information within the framework. Therefore SOA is used her to offer the required architecture. The following sections will describe the grid computing and SOA.

2.1. grid Computing Over the coming years, the utilities and services will become an integral part of future socioeconomical fabric. The realisation of such a vision will be very much affected by many factors including; cost of access, reliability, dependability and securit. Hoschek (Hoschek, 2002) defined grid computing as; “… collaborative distributed Internet systems characterized by large scale, heterogeneity, lack of central control, multiple autonomous administrative domains, unreliable components and frequent dynamic change …”. Whereas, Berman et al. (F. Berman, 2003) defined grid computing as; “…The Grid is the computing and data management infrastructure that will provide the electronic underpinning for a global society in business, government, research, science and entertainment…” From the above definitions, the benefits of grid computing to support enterprise business application are accrued through collaborative distributed resources and information sharing including; software, hardware and associated content, to build one large system serving all subsystems and consumers. The important capabilities of grid computing that assist in clarifying the expected usability

121

Web 2.0

of such technology are (G. Menkhaus, 2002; I. Foster, C. Kesselman, J. Nick, S. Tuecke, 2002; I. Foster, C. Kesselman,S. Tuecke, 2001; IBM, 2003b; L. Ferreira, 2003): •







122

Exploiting Resources: As Ferreira et al. (L. Ferreira, 2003) pointed out that the ideal use of grid computing is to run distributed application on different machines. The machine on which the application is normally run might be busy due to the peaks in activity; therefore, it could be run on an idle machine elsewhere on the grid. Resources Allocation: availability, reliability, interpretability and Service Level of Agreement (SLA) are improved by the merge of the grid computing competence in performing the process of exploring resources with other technologies, like SOA. Parallel Processing: Such computing power is driving a new evolution in industries like inancial modelling, motion picture animation, and many others that can be partitioned into independently running parts to reduce the time of processing by splitting the application among many CPUs. Offer a Fabric for Running Large-Scale Enterprise Applications: One fact that must be understood that not all applications can be transformed to run in parallel on a grid and achieve scalability according to the fact that not all the programs can be partitioned (L. Ferreira, 2003). Therefore, the programming direction of using Web services in grid computing is increasing rapidly (G. Menkhaus, 2002) because web services are considered as classes that can be distributed over the grid environment to get more processing power (IBM, 2003b). Therefore SOA can get beneit from grid computing to offer an operation framework for enterprise applications.







Virtual Resources Collaboration: Another important capability of the grid computing contribution is the facilitating and simplifying the collaboration among a wider audience to give better services. In the past, distributed computing promised this collaboration and achieved it to some extent (I. Foster, C. Kesselman, J. Nick, S. Tuecke, 2002). However, grid computing is often presented as the next step towards resources virtualisation and sharing for the wider community. This assumption characterised large virtual computing systems offering a variety of virtual resources. This feature is so important for the Web2.0 community, in order to create distributed virtual community. Resource Balancing: Grid computing consists of a huge numbers of resources from services, infrastructures and networking. These resources collaborate together in order to offer reliable services to the consumers with high performance (L. Ferreira, 2003). Thus, resources load balancing can be get beneit from grid computing in order to enhance the utilisation of grid computing in terms of resources availability, reliability and QoS (I. Foster, C. Kesselman, J. Nick, S. Tuecke, 2002). Reliability: Ferreira et al. (L. Ferreira, 2003) described the reliability of the grid computing resources as “…High-end conventional computing systems use expensive hardware to increase reliability…”. The reliability of the resources is discussed from two perspectives; hardware, networks, and software services failure viewpoints. The next gain in building reliable systems is now focused on software and software services reliability and resilience. Therefore grid computing emerged to address (I. Foster, C. Kesselman,S. Tuecke, 2001) the development of low cost high-performance, highreliable and high-availability computing.

Web 2.0





Management: The goal of the resources availability on the grid is to disperse information technology’s infrastructure and handle heterogeneous systems (L. Ferreira, 2003). In such heterogeneous, decentralised and distributed information, administrators (consumers and/or control system) provide the system with policies, rules and strategies that handle and manage how the different organisations might share or compete for the resources after getting real-time information from the web or distributed computing. In addition, grid computing environment is expected to manage the resources in a way that improve the critical parameters such as; reliably, idelity, QoS and others. Open Standards: The idea is to convince the community of software engineers currently developing the grid, including those from major IT companies, to set common standards for the grid up-front. The open standards community assists the applications to communicate with infrastructures, services and tools in formalization and semantic ways in addition to offer a way for describing the use of resources in standard format. The feature is so important in merging the SOA with grid computing, because the key feature of SOA is depending on open standard services.

to bring the Object Oriented (OO) mentality to the distributed large scale enterprise applications, where the new distributed applications are proposed to be structured from numbers of small object models (B. Borges, 2004), such objects can be a web services. SOA is the architecture for the next generation of enterprise applications which depend on the use of software and hardware as services for the applications. The use of open standard format for developing services is vital in order to form a standard framework for the applications. Web services technology is used for developing services for SOA applications, where web services provided the required open standard format. The open standard format is achieved based on the use of Simple Object Access Protocol (SOAP), Web Services Description Language (WSDL) and eXtensible Markup Language (XML). SOA and Web2.0 is each one completes the other. Web2.0 offers a framework for dealing with users. This includes recording user requests, user behavior, user categories, user experiences, framework parameters, requested services, user interaction with other users, and user community. On the other hand, SOA offers all the underneath architecture for receiving the user requests and providing services to the user. Moreover, SOA is in charge of connecting, managing, and controlling the Web2.0 framework with grid computing resources.

2.2. service oriented Architecture 3. soAw2g Model Services Oriented Architecture (SOA) is anticipated to offer a generic model for implementing large scale enterprise applications (B. Borges, 2004; M. Endrei, 2004), such as e-health, ecommerce and e-government. Hence, SOA is a model for hiding the complexity of the usability of distributed services from the consumer in one hand, and provide a framework for services provider in the second hand (B. Borges, 2004; Fellenstein, 2005). Moreover, SOA is adopted

The proposed model SOAW2G is the combination of SOA, Web2.0 framework and Grid resources to form a generic framework. This framework is able to use the user experiences based on Web2.0, the ultimate resources based on grid computing overlay, and SOA to manage the interaction between the units of the model. The model consists of six layers to manage the interaction between the user and the resources.

123

Web 2.0

Figure 1. SOAW2G model

There are other three layers responsible for controlling the security, managing the system and offering ontology for assisting in exchanging information between the layers of the system in open standard format.

3.1. resources layer The resources layer in SOAW2G includes three main categories, which are Services, Computational and Data Process. These categories consist of varieties of components which include range of services types, infrastructures, communication systems, monitoring resources, storage system and controlling facilities. The services resources consist of all types of application and management services that are offered to the customers, such as health services(W. Omar, A.Taleb-Bendiab, 2006), financial services

124

(B. Ahmad, 2006), dictionary services, monitoring services and others. The computational category covers the resources that are required for processing tasks requested from the services part in this layer or other specific tasks that are demanded by the users. The data process resources consist of all the storage system and the access ways to the storage data, such as data grid. This category also includes data processing services like data mining. The resources layer at SOAW2G is proposed to serve huge variety types of applications from financial, research, education, government, health and others. Therefore, there is a need to hide the complexity of integrating resources from the top layer (users). Web Services technology would be an attractive solution for most of the developers due to the broad compatibility that this technology has. This will offer an open standard at resources

Web 2.0

layer which also requires a standard language for describing resources.

3.2. resources Management layer This layer is responsible for classifying the resources that are deployed by the provider according to the nature, functionality and behavior of resources. Such classification process has been proposed to improve the functionality of the below layer by enhancing the manageability and fidelity of selecting resources. Moreover, this layer is responsible for looking after requests coming from upper layers in order to find the best services that match the requests. Managing users’ experiences to manage the resources layer is the other job of this layer. This layer consist of three main components to accomplish this task, which are classification for classifying resources, predication for predicating the category of resources that may be requested by upper layers and reasoning for understanding the requests and interact according to it.

3.3. Control layer Control layer is in charge of managing the resources found in the resources layer in a way that offers high reliability, quality of services, availability (fault tolerance) and maintainability. This layer consists of a number of tools and services, which assist in carrying out the task, such as replication, fault tolerance, load balance, mirroring and others. For example, this layer is in charge of detecting the overload requests on specific services, and hence finding the way for keep the service alive and manages the load. This can be achieved through load balance techniques, such as replicate the service in another server, manage the access to service based on priority or first shortest job, and schedule the service usage in advanced based on advanced and on demand access .

3.4. support functions layer This layer is required for managing the processes of deploying, discovering and invoking resources. The deploy function coordinates the deployment of the resources from providers to resources containers in the resources layer, the discovery function manages the process of discovering resources by the users, while invoke function controls and advices the way of accessing the resources in the lower layers by the users. The provider should provide the system with prosperous information in order to assist the layer in discovering the most suitable resources to the user. Such information will help in improving the fidelity of the system.

3.5. User interface layer This layer is responsible for dealing with the users through presenting the data in different forms like forums, RSS, wikis, blogs which show the interests of the users. Also this layer should have the ability to monitor the users’ activities and record them in a history log in order to provide them to the management layer for re-adjustment of the framework to be suited for the running application. This layer should be flexible to include different rules and policies that describe the operations of the user based on the nature of the applications (i.e. health, science, game and other frameworks).

3.6. User layer The user layer represents the consumers as well as applications. The user in this framework is an active user and not a passive one. In another words, the user interacts with the system to improve the operation of the framework through providing the system with experiences, arguments (rules), application specification and other information assists in reconfiguring the framework to give better services. Plus this active user would feed

125

Web 2.0

the whole system with a very vital and important ingredient, the Data that would be the main engine to accelerate the process of budding and raising the Web2.0 sites and applications.

task. Different services and infrastructures from resources layer will be used in this layer, such as monitoring system, intelligent services (for planning), effectors, and others.

3.7. security layer

3.9. Knowledge layer

To protect the system, the SOAW2G model consists of security layer working with the other layers of the model. The security layer commences by checking the authority of the users, then checking the Service Level of Agreement (SLA) in support function layer. The SLA is required to control and protect the users who have the right and privilege for deploying and using resources from those who do not have. A file system is proposed to be used to record the authority and SLA’s for different users, and an encryption mechanism can be very useful to encrypt these files and protected them against sniffing and vandalizing activities. The security at the control layer is to manage the administrator access to control services. At the end, security at the resources layer is to protect the layer from the different types of attacks from outside or inside which can be implemented in sourced or outsourced security systems (virus protection, worms protection applications, etc…).

The knowledge layer in this model is proposed to offer wealthy information to all layers of the system that would assists in efficient usage of the framework. This layer should be in attached to all layers of the system for gathering and providing information to each one of them. For example, this layer should collect and provide information regarding the available control services to control layer, the security policies and SLA to security layer, user information, experiences to user interface layer, classification and prediction service to resources management layer, and the available sensors, actuators, loggers and other monitoring resources to the monitoring system,. Furthermore, this layer assists the user in selecting the services from resources layer based on the information provided by resources management. Because this layer is involved with all layers, it should use an open standard format that can be readable and understandable by all components. Therefore, in this work, we designed a Resources Markup-Language (RML) for describing the resources in open standard format as explained in the following sections. Other types of description languages are used here to describe the processes and components of the system such as Sensors and Actuators Description Language (SADL) and Monitor Session Description Language (MSDL) (W. Omar 2005). So the existence of such layer would offer a storage and retrieval mechanism for all layers in the framework, and this would facilitate the process of information exchange between layers (not necessarily contiguous layers) in a great way, also it can be a backup for the information found in each layer which will add a precious amount of robustness to the whole system. Obviously

3.8. Management layer The management layer works in corporation with all layers of the model for managing the Web2.0 Grid framework. This layer consists of numbers of capabilities working together for managing the framework. Such capabilities are framework configuration, optimization, adaption, healing, protection, organizing, and others which assist in improving the operational framework and moving it to on demand framework (Fellenstein, 2005). All these capabilities should be selected, executed, blocked and destroyed in an automated way. Therefore, autonomic computing (IBM, 2003a; Murch, 2004) is proposed to be used in this case for implementing the self management

126

Web 2.0

a huge amount of data will have to be stored in some way, and this way can be achieved through utilizing the data grid.

4. soAw2g in ACtion In this section, the process of deploying, discovering, invoking and managing resources in SOAW2G framework would be explained. This would be done by providing a walkthrough for a basic expected scenario of service providing by providers and service consuming by users. As shown in Figure 2, the resources providers would start using the Deploy function from Support Function Layer to upload services (or other type of resources) to resources containers. The resources management layer is in charge of understanding the functionality, behavior and nature of the deployed resources in order to categorize it to one of the sub domain in one of the three resources category. This process improves the functionality of the system in the usability of the resources, which will be reflected in improving the reliability, performance, availability, maintainability and fidelity. For this layer to serve other layers in an efficient way, it requires to categorize and syndicate the resources containers in an intelligent and automated way to achieve the self-governor system. The resources layer needs a control system to manage it and insure the high availability and maintainability of the resources. For control and management systems to work in an efficient way, they require robust monitoring system that is able to record all activities of the resources container. Therefore, in this scenario the monitoring system uses the monitoring resources from resources containers to monitor the resources layer. The control system, which is found in the control layer, would manage the replications, QoS, Fault Tolerance, load balance and other control services attributes and requirements for controlling the resources layer. The framework operation security is one of the important aspects that keep the framework

running in safe and protected mode. Therefore, monitoring system provides the security system with prosperous information regarding the security situation in the resources container. The security system specifies the security attributes and authorizations needed for the usability of resources in the resources layer as well as the administrator and user rights. To this end, we have resources deployed by the providers and categorized through resources management layer inside resources container. Such resources are monitored by the monitoring system, which in its turn provides the control and security systems with the required information. Now, after deploying resources, it is time for the user to interact with the system and play his/ her/it’s role in requesting services and providing experiences, rules, arguments and knowledge to the framework. The user interacts with the application within the SOAW2G framework through User Interface Layer, which would be taking information from knowledge layer regarding the available resources, security levels, and other users’ information. This information would be delivered to the users, so the user would have a clear idea about the service and its shape after sending request and discovering the required resources before the invocation process. This assists in improving user time consuming, QoS and fidelity in selecting resources by requesting the most suitable services to the user instead of invoking each resource and search for the required one. In addition, the user interface layer has another important job, which is recording the users’ experiences, preferences, environment, characteristics and applications features in order to adjust the framework to be more effective for the user. This information is used also, to manage and re-classify the underneath layer (resources layer) according to the user preference. After that, the user can call and use services from the resources container using the Invoke function found in the Support Functions layer.

127

Web 2.0

Figure 2. SOAW2G in action

At the end, all the components of SOAW2G interact with each other through the Knowledge Layer. The knowledge layer consists of all the information about the framework in an open standard format that makes the components of the framework interact between each other in a smooth and efficient way, i.e. the deployed resources by the providers should be easily understood and classified by the resources management layer and then requested and used by the consumers. As has been mentioned before the next section would talk about the RML.

5. resoUrCes MArKUPlAngUAge Resources Markup Language (RML) is simply a description language for describing the services and resources available in SOAW2G framework sites and applications in an open standard format

128

to offer a resources metadata, in which can be considered as a guidance for the resource discovery, and hence assist the user at the top level of the SOAW2G model to achieve high fidelity in selecting resources. In order to achieve the crossplatform consuming goal, RML has been written based on XML data format describing the different properties for the resources. The RML is flexible in describing the resources at the underneath layer. In this chapter, we only illustrate the RML for describing health services, but it can be used to describe any type of resources. Figure 3 illustrates RML for health applications (Health Resources Markup Language (HRML)). The HRML is used to describe five main categories that are required for describing drug information. These categories are: general information, Used For, Ingredients, Side Effects, and Similar Drugs. The development of RML has been motivated by the following:

Web 2.0

• • • •

Resources metadata to describe the resources. Resources management through resources categorization and classiication. Automated processing of resources information by software agents. Interworking among resources by combining data from several resources to structure new resources.

Designing RML is the first step in categorizing the resources layer, where it offers the fabric of information for the classification process. Different algorithm would be used for clustering the resources according to resources’ functionality, behavior, SLA and nature. Such algorithms are Multiple Regression Algorithm, Data mining, and machine learning. Next section describes the classifying of drugs information resources based on users (doctors) experiences.

6. resoUrCes ClAssifiCAtion: CAse stUdy Classifying resources is proposed in this work to achieve the goals (availability, performance, maintainability, and reliability) of using grid computing resources with Web2.0 framework. The system clusters the resources based on the users’ experiences. The users’ experiences are different from framework to another framework, for example the experiences will be medical experiences if the framework is concerning with health. The users in this case would be doctors, nurses, and even patients. The information in this case would be medical information. For the process engineering framework, the experiences would be based on the feedback from engineers and the information describes the processes. For education framework, the experiences would be based on the feedback and comments from the academic staff and students, and so on. On the

same line, the resources at resources layer are different based on the operational framework. For example: medicine information, medical sensors, medical test, health services and related health resources in case of health framework. Therefore, as a case study we will depend on the health framework to show how the user experiences that are collected from the top layer will manage the underneath layer (resources) through classifying the health resources (in this case medicine or drug information) based on the user feedback. For sure, there is a need for classifying the experiences according to the characteristic of the framework, type of users, type of the experiences, and the target resources at the resources layer. Classifying the experiences will assist in classifying the resources. Autonomic computing (Miller, 2005; Murch, 2004; W. Omar, A. Taleb-Bendiab, Y. Karam, 2006) is adopted in this work to classify the experiences according to the framework and parameters of the system. The classifying would be helpful in classifying resources based on these experiences. For example, in health environment, the feedback from the doctor can be used to classify medicine information (as resources) to be as treatment for disease. In this case a collaboration environment will be generated between different doctors (each one does not know the others) for sharing information on using medicines. Also the feedback from doctors can be used to indicate if there is a need for specific test to diagnose a suspension cases. The sharing experiences will not be only useful in classifying resources, but also in giving advices for other doctors in case there a suspicion cases. SOAW2G manages the process of sharing resources in the user interface layer, which is responsible for recording and collecting the experiences from the users. In this case study, the classification of medicine information based on the feeding experiences from the users (doctors) is demonstrated. Each medicine has the features of treating number of diseases. But many of the medicines can share

129

Web 2.0

Figure 3. HRML

the same characteristics of treating same disease. For example, Loratadine (Generic Name) is used to temporarily relieve the symptoms of hay fever (allergy to pollen, dust, or other substances in the air) and other allergies. These symptoms include sneezing, runny nose, and itchy eyes, nose, or throat. Loratadine is also used to treat itching and redness caused by hives (MedlinePlus; Weka). There are number of medicines from different brand names use Loratadine to treat the allergies disease. For example CLARINASE® REPETABS

130

– CMI can be used for nasal and sinus congestion, sneezing, runny nose, watery, and itchy eyes (APPGuide). The same inductions are for drug Claritin (RxList), which is used for relief of nasal and nonnasal symptoms of seasonal allergic rhinitis and for the treatment of chronic idiopathic urticaria in patients 2 years of age or older. From this simple example, the similarity between the two medicines in treating syndrome can be recognized. But which one is better to be used? We can’t say which one is better, because

Web 2.0

such decision is depending on the doctor opinion at the first place, the side effects of the drug and how this can effect on the patient if he/she has allergies against one of the medicine ingredient, availability of the drug in the local market, and the local price of the drug. All the information that is used as features and rules for classification processes is provided through the users starting from the materials that are used in drugs manufacturing, doctors’ opinions, market study, allergy effect, side effects and other information. All the mentioned information will be provided to the system through the user interfaces layer. The knowledge layer from SOAW2G is in charge of receiving the information and be sure that the resources management system receives the collected information in open standard format. The resources management layer uses the information to classify the medicine information (as resources) in a way that be dynamic and different from health framework to another health framework depending on the features and characteristics that are received from the users. Grid computing resources are requested in this work to offer fabric for collecting information and experiences from users, saving and managing the information in large scale distributed storage system, offering classifying services, and providing tools for converting the information into semantic format. Autonomic computing is adopted, as we mentioned before, in this work to be utilized in the resources management layer for offering self-management facility to resources layer. The autonomic computing requires prosperous information for carrying out the task. Such information is flooded by the users as explained previously. The autonomic computing service uses the information for classifying resources as the first step in selfmanagement life cycle. Intelligent services are required for conducting the classification process within autonomic computing services. Supervised machine learning algorithm is proposed in this work, like Neural Net, VSM, Decision Tree and others to implement the intel-

ligent classification stuff. The work now is under the process of selecting the best algorithm(s) for classifying resources. One of the suggestions is to use more than one algorithm at the same time. Weka software (Weka) is used for implementing the intelligent stuff based on the use of machine learning. Weka is a collection of machine learning algorithms for data mining tasks. Weka contains tools for data pre-processing, classification, regression, clustering, association rules, and visualization. Weka is aiming to be developed as web service in order to be included in the grid computing resources.

7. ConClUsion This chapter presents a model for managing the interaction between Web2.0 framework and grid computing resources based on the use of SOA model. The proposed model consists of number of layers. This work is at the beginning, we start with describing the underneath layer which represents the resources container. Resources Markup – Language (RML) has been developed for offering the required information for managing the resources layer. Clustering resources is proposed to be used in this work. Case study of classifying resources is illustrated through this work to show how SOAW2G can use to support the large scale enterprise applications. The work is still under process, but the initial results are promising and show the possibility of using the model for improving the use of resources for supporting the applications. The future work will focus on the use of the SOAW2G model for predicting the required resources for each framework. In this case, the system can be migrated from one framework to new framework through selecting the basic required resources based on the new framework characteristics.

131

Web 2.0

8. referenCes Ahmad, B., Omar, W., & Taleb-Bendiab, A. (2006). Intelligent monitoring model for sensing financial application behaviour based on grid computing overlay. Paper presented at the 2006 IEEE International Conference on Services Computing (SCC 2006). APPGuide. http://www.appgonline.com. au/drug.asp?drug_id=00097529&t=cmi. Ahmad, B., Omar, W., & Taleb-Bendiab, A. (2006). Intelligent Monitoring Model For Sensing Financial Application Behaviour Based On Grid Computing Overlay. Paper presented at the Submitted to 2006 IEEE International Conference on Services Computing (SCC 2006), USA. Berman, F. G. fox, A. Hey. (2003). Grid Computing: Making the Global Infrastructures a Reality. Chichester, West Sussex, England: John Wiley and Sons Ltd. Borges, B. K. H., A. Arsanjani. (2004). ServiceOriented Architecture. Endrei, M. J. A., A. Arsanjani, S. Chua, P. Comte, P. Krogdahl, M. Luo, T. Newling. (2004). Patterns: Service-oriented Architecture and Web Services: IBM Redbook. MedlinePlus. http://www.nlm.nih. gov/medlineplus/druginfo/medmaster/a697038. html.

Hogg, R., et al. (2006). Overview of Business Models for Web 2.0 Communities. Paper presented at the Gemeinschaften in Neuen Medien, Technische Universitat Dresden. RxList. http://www. rxlist.com/cgi/generic/lorat.htm. Hoschek, W. (2002). Peer-to-Peer Grid Databases for Web Service Discovery. IBM. (2003a). Autonomic Computing. May 2004, from http://www.research.ibm.com/autonomic IBM. (2003b). On Demand Glossary. from http:// www-3.ibm.com/e-business/doc/content/toolkit/ glossary_o.html Menkhaus, G., Pree, W., Baumeister, P., & Deichsel, U. (2002). Interaction of Device-Independent User Interfaces with Web services. from http:// www.softwareresearch.net/site/publications/ C048.pdf Miller, B. (2005). The Autonomic computing edge: The “Standard” way of autonomic computing. July 2005 Murch, R. (2004). Autonomic Computing: Prentice Hall. O’Reilly, T. (2005). What Is Web 2.0. O’Reilly Network, from www.oreillynet.com/pub/a/oreilly/ tim/news/2005/09/30/what-is-web-20.html

Ferreira, L., Berstis, V., Armstrong, J., Kendzierski, M., Neukoetter, A., Takagi, M., et al. (2003). Introduction to Grid Computing with Globus: IBM.

Omar, W., Ahmad, B., Taleb-Bendiab, A., & Karam, Y. (2005, 24-28, May). A Software Framework for Open Standard Self-Managing Sensor Overlay For Web Services. Paper presented at the 7th International Conference on Enterprise Information Systems (ICEIS2005), MIAMI BEACH- FLORIDA-USA.

Foster, I., Kesselman, C., Nick, J., & Tuecke, S. (2002). The Physiology of the Grid: An Open Grid Services Architecture for Distributed Systems Integration., from http://www.globus.org/ogsa/

Omar, W., & Taleb-Bendiab, A. (2006). E-Health Support Services Based On Service Oriented Architecture. IEEE IT Professional, 8(2), 35–41. doi:10.1109/MITP.2006.32

Fellenstein, G. (2005). On Demand Computing: Technologies and Strategies: IBM press.

Foster, I., Kesselman, C., & Tuecke, S. (2001). The Anatomy of the Grid. 2003, from www.globus. org/research/papers/anatomy.pdf

132

Web 2.0

W. Omar, A. Taleb-Bendiab, Y. Karam. (2006). Autonomic Middleware Services for Just-In-Time Grid Services Provisioning. Journal of Computer Sciences. Schoth, C., & Janner, T. (2007). Web 2.0 and SOA: Coverging Concepts Enabling the Internet of Services. IEEE IT Professional, 9(3), 36–42. doi:10.1109/MITP.2007.60

SematicGrid. www.semanticgrid.org Wail, M. Omar, Ali Dhia K. Abbas, Taleb-Bendiab. (2007, May/Jun). SOAW2 for Managing the Web 2.0 Framework. IT Professional, 9(3), 30–35. doi:10.1109/MITP.2007.56 Weka. http://www.cs.waikato.ac.nz/ml/weka/.

133

Section 3

Web Architecture

135

Chapter 8

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures Marcel Linnenfelser Synlag Web Engineering, Germany Sebastian Weber Fraunhofer Institute for Experimental Software Engineering (IESE), Germany Jörg Rech Fraunhofer Institute for Experimental Software Engineering (IESE), Germany

AbstrACt An important aspect of Web 2.0, mentioned by Tim O’Reilly, is the rich user experience. Web 2.0 applications offer the user a desktop-like interface to bring back eficiency and productivity. The click-wait-andrefresh-cycle of normal Web applications leads to a less responsive, and thus less eficient, user interface. To serve the needs of these so-called rich Internet applications (RIA), many different approaches have emerged, based either on Web standards or on proprietary approaches. This chapter aims at deining a qualiied criterion system for comparing RIA platforms. Thereafter, those RIA platforms are selected and analyzed in terms of the criterion system that is most likely to become widely accepted.

introdUCtion In his essay “What Is Web 2.0”, Tim O’Reilly (2005) collected attributes that qualify a Web platform as Web 2.0. The key features of Web 2.0 platforms from the technological point of view are: User Generated Content, Tagging / Folksonomy, Content Syndication, and Rich User Experience. While the other features mentioned affect only some minor DOI: 10.4018/978-1-60566-384-5.ch008

parts of the technological side of a Web platform, the Rich User Experience requires a fundamental architectural decision. This chapter will focus on Web X.0 technologies that enable a Rich User Experience and state several criteria for the differentiation and evaluation of these technologies for Web 2.0 services. Web applications themselves have many advantages in software distribution and deployment. But as Kevin Hakman (2006) from TIBCO Software Inc. showed by changing over from a fat client to a Web

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

client in Siebel Systems software in 2002, Web clients may have a serious impact on productivity. At one call center, there was a 30% productivity loss caused by the Click-Wait-and-Refresh-Cycle, as he called it. RIA technologies enable Web clients to measure up to fat clients regarding GUI usability, thus combining the advantages of Web clients and fat clients. This chapter aims at providing a criterion system for evaluating RIA platforms and frameworks, which is defined in section “Definition of a criterion system”. It is designed as a tool for decision makers to compare such platforms and to help them select an appropriate one for a specific project. The section “Platform Outlines” applies the criterion system to evaluate and compare those currently available platforms that are most likely to become widely accepted.

standards. The best of communication means incorporating two-way interactive audio and video” (Duhl, 2003). Asynchronous JavaScript + XML (AJAX) “Ajax isn’t a technology. It’s really several technologies, each flourishing in its own right, coming together in powerful new ways. Ajax incorporates:

definitions

Offline Web applicationOffline Web Applications utilize Web technologies to build desktop applications. To enable the applications to run online and offline, they have to be granted file access in order to be able to save states (“TR10: Offline Web Applications”). RIA runtime environment An RIA runtime environment provides an environment that allows running platform independent RIAs. Usually, the runtime environment is available for different operating systems. RIA framework The term RIA framework is used in this text to describe an application framework that supports the development of RIAs for one or more RIA runtime environments.

Web application A Web application is an application “accessed over the World Wide Web by using a Web browser” (“WHATWG FAQ”). Rich Internet Application (RIA) The term was coined by Macromedia in 2002. “Macromedia defines RIAs as combining the best user interface functionality of desktop software applications with the broad reach and low-cost deployment of Web applications and the best of interactive, multimedia communication. The end result: an application providing a more intuitive, responsive, and effective user experience. Specifically, the best of the desktop includes providing an interactive user interface for validation and formatting, fast interface response times with no page refresh, common user interface behaviors such as dragand-drop and the ability to work online and offline. The best of the Web includes capabilities such as instant deployment, cross-platform availability, the use of progressive download for retrieving content and data, the magazine-like layout of Web pages and leveraging widely adopted Internet

136

• • • • •

standards-based presentation using XHTML and CSS; dynamic display and interaction using the Document Object Model; data interchange and manipulation using XML and XSLT; asynchronous data retrieval using XMLHttpRequest; and JavaScript binding everything together” (Garret, 2005).

bACKgroUnd Today RIA is a hyped topic, causing many companies to enter the market of RIA platforms and frameworks. There exist numerous JavaScriptbased frameworks enabling the development of RIAs that run directly in the browser, without

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

the need for additional plug-ins. Adobe offers a framework called Flex, which enables the development of applications with desktop-like interfaces targeting the Flash player. Even Microsoft and Sun have come up with their own solutions for RIAs, but so far, only Microsoft has released a final version. And there are a lot more frameworks and platforms for building RIAs, e.g., Lobo1, Curl2, Omnis3, Mozilla Prisma4, to name but a few. The large number of competitors in the RIA platform and framework market makes it difficult to gain a general overview. Douglas Engelbart presented the concept of an “oNLine System” (NLS) in December 1968. The system allowed two people to communicate via audio and video and to collaborate on a shared screen to create and edit text documents (Norman, 2005, p. 41). In 1969, a team led by Leonard Kleinrock established a connection between two host computers over a network switch (Norman, 2005, p. 41). The first network was established. Terminals such as DECwriter II and later DEC VT-100 made the first distributed applications possible, at least from the users’ point of view. In the 1970s, the Personal Computer (PC) appeared, and after 1985, computer networks fanned out (Ceruzzi, 1998, p. 6). Personal Computers allowed the use of fat client network applications. While being normal desktop applications, those fat clients could be as powerful as any other desktop application. The disadvantage of such applications is the need to install them on all client computers. The Web was invented by Tim Berners-Lee in 1989-1990. He created the HTML and implemented the first HTTP server and the first browser (Norman, 2005, p. 100-102). Web applications are lightweight and do not need to be installed on the client computer. But as mentioned in the introduction, traditional Web applications may have a serious impact on productivity. With the release of Internet Explorer 5 in 1999, Microsoft introduced XMLHTTP ActiveX Control (“About Native XMLHTTP”), which allowed a type of

applications later known as AJAX-based RIAs (Garret, 2005). The term RIA was coined by Macromedia in 2002 to describe desktop-like Flash-based applications. In 2002, Macromedia released Flash MX featuring Flash UI Components (Mook, 2003). In the following, a criterion system will be defined that allows characterization of RIA platforms, and thus comparison of the different RIA platforms. The outline of characteristics given in this chapter will help to evaluate the technologies regarding specific needs.

Client-server Cross-seCtion RIAs are basically client-server applications. They differ in the amount of processing done on the server and the client. For this examination, a layer model with three layers is assumed. The first layer hosts the data access, the second the business logic, and the last one the presentation logic. Figure 1 shows five approaches to allocating the layers on client and server. Approach one shows a Fat Client (Mahemmoff, 2006, page 317), where all the processing takes place on the client side. Communication with the server is only necessary for manipulating or retrieving data. Data may be cached on the client. Approach two hosts only a part of the business logic on the client and the other part on the server, while approach three only performs retrieval and presentation on the client. To update the presentation in approach three, data has to be sent to the server. After the data has been processed by the business logic, the results are forwarded to the client, where the presentation is then calculated. While approach four splits the presentation logic and calculates a part of the calculation on the server side, the last approach calculates the presentation on the server and sends a description (e.g., HTML) of it to the client, where the presentation is rendered following the description.

137

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

Figure 1. Cross-section through client-server architecture

The first three designs may be implemented with any of the discussed RIA technologies. Approach four has to have some kind of rendering engine on the client side as well as a processing engine. In an example implementation of approach four, implemented using AJAX, the server-side presentation logic would process data obtained from the business logic and would generate HTML and JavaScript code to be sent to the client. Client-side events are handled by the generated JavaScript code. The server is requested via XMLHttpRequest (see section “AJAX”). The server-side processing logic generates HTML snippets to replace the HTML code of the updated sections of the page. Client-side JavaScript is used to replace the obsolete HTML. Approach five is a traditional Web application where all the HTML is generated on the server. The HTML is sent to the client and rendered by the rendering engine. On every client-side event, a new request is sent and the whole page is refreshed.

138

definition of A Criterion systeM In the following, a criterion system is defined that allows evaluating and comparing RIA platforms. The criteria cover development aspects as well as runtime and availability issues. •





Tool support: Available development tools like debugger and proiler. Tool support is important for eficiency in development and testing. Availability: Availability on different operating systems. Depending on the audience, the RIA platform used has to support the target operating systems. Available APIs and functionalities: Available APIs such as collections, threading, and special language extensions. The availability of a functionality or type of API such as a 3D API may enable or

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures













disable a platform or framework for a certain project. Language characteristics: Characteristics like object-orientation, inheritance, and such. Language characteristics can determine the applicability of a platform or framework regarding the size of a software project. For example, use of an un-typed language may be inappropriate for large systems. Runtime environment: Characteristics of the runtime environment. The runtime environment determines the features, the platform independence, and the performance of a platform. Extension challenges: Challenges in creating and modifying UI elements. The RIA platforms differ in architecture. This leads to various extendibility challenges and different lexibility of the styling capabilities. For all-audience applications, styling capabilities may be essential. For some frameworks, extended tool support exists to assist designers. Market penetration: Market penetration regarding all Internet-connected client computers. Market penetration is important especially for all-audience applications. Beyond the browser: Many platforms allow developing applications to be deployed not only in the browser. Ofline Web applications close the gap between desktop applications and Web applications. Applications with direct access to the ile system can offer improvements in terms of usability. Interoperability: Interoperability with JavaScript and thereby interoperability with other plug-ins. For some applications, a mix of the browser’s native RIA capabilities and plug-ins is the best choice. Therefore, interoperability may be important.









Separation of design and logic: Special language features and approaches to separate design (UI) and logic. The separation of design and logic can help to increase maintainability, but also makes it possible to split responsibilities between programmers and designers. Supported media types: Supported video and audio formats, as well as bitmap and vector graphic formats. The supported media formats may determine the decision for a speciic RIA platform. Installation: Download size, install experience, and such. The download size and simplicity of installation of the runtime environment of a chosen framework may be irrelevant for business applications if automatic update mechanisms are in use at the customer’s site. But at least for all-audience applications, these parameters are critical in terms of acceptance. Supported devices: Supported input and output devices. Access to webcams and microphones enables different kinds of applications, such as collaborative applications and applications that require taking cam shots (e.g., barcode reader via webcam).

The criterion system defined in this section will be used in the following section “Platform Outlines” to evaluate and compare the four platforms and frameworks AJAX, Microsoft Silverlight, Adobe Flex, and JavaFX.

PlAtforM oUtlines AJAX The AJAX platform is made up of Web browser applications. It is based on the W3C and ECMA standards HTML, CSS, XML, JavaScript / ECMA Script (including JSON), and XMLHttpRequest-

139

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

Object and similar mechanisms. Due to the multitude of existing browsers, the platform is highly heterogeneous. The browsers differ in their support of the standards. Thus, cross-browser compatiblity is an important subject for the AJAX platform and is sometimes hard to deal with.

tool support Countless text editors are available with syntax coloring capabilities. Editors like Adobe GoLive and Adobe Dreamweaver allow WYSIWYG editing of HTML. These products have been used by designers for years, so there is much expertise available in the designer community. The programs support developers in creating HTML, CSS, and JavaScript, and feature tools for dealing with browser incompatibilities. Several IDEs provide support for JavaScript. One of the most feature-rich IDEs for JavaScript is the Eclipse-based Aptana Studio5.. Aptana comes with syntax coloring, code assist for HTML, CSS, and JavaScript, as well as JavaScript debugging for Firefox and Internet Explorer (IE). Debugging supports breakpoints and watched variables. Firebug is a debugger extension for Firefox. It allows setting break points, analyzing the network traffic of AJAX applications, inspecting the page structure, and profiling JavaScript code. The number of JavaScript frameworks available makes it a very time-consuming task to get an overview. Also, the number of frameworks that provide JavaScript widgets6 is anything but small and includes Dojo Dijit, Backbase, TIBCO General Interface, Ext JS, and Adobe Spry, to name but a few. Because there is not one single standard framework with one standard set of widgets, the only way of providing WYSIWYG editing of GUIs is for each framework to provide its own tools for this purpose. The GUI builders of TIBCO and Ext JS are written using their respective frameworks and run in the browser. An interesting approach is the Google Web Toolkit (GWT). GWT uses a Java-to-JavaScript

140

translator to add an abstraction layer. This allows using Java tools and the GWT development tools, including a full-featured debugger.

Availability HTML, CSS, and JavaScript are available in nearly every browser in use, but many of the installed browsers are not fully standards-compatible. Bugs also induce non-standard-conformant or even unpredictable behavior7. The heterogeneity of the AJAX platform leads to increased development and testing effort. More than 98% of installed browsers are AJAX-capable (“Browserwatch”) (“The Counter. com”) (at least in Germany), which means they support XMLHttpRequest (“The XMLHttpRequest Object”) objects or XMLHTTP ActiveX control (“About Native XMLHTTP”).

Available APis, libraries, and functionalities Designed to manipulate an HTML page after it is loaded by the browser, the key API of JavaScript is the Document Object Model (DOM). The DOM grants access to the elements of the currently loaded HTML page and to the attributes of the page elements, by providing methods and properties to retrieve, modify, update, and delete parts of the document (“The DOM and JavaScript”). The DOM has been standardized by the W3C (“Document Object Model (DOM)”) to provide a language-neutral and compatible interface for accessing content structure and style of documents. Manipulation of the HTML DOM using JavaScript is often referred to as Dynamic HTML (DHTML). JavaScript provides support for working with arrays, doing calculations, and working with regular expressions. Several libraries including Dojo Toolkit and Prototype8 address problems with JavaScript version and browser incompatibilities by build-

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

ing abstraction layers and extending the browser DOM. With the help of the Dojo Toolkit, JavaScript supports a technique called comet (“Comet: Low Latency Data for the Browser”). Comet uses long-lived HTTP connections to allow the server to push data to the client.

Language Characteristics A Web developer who develops for the AJAX platform should at least know three languages for doing client-side web development: HTML, JavaScript, and Cascading Style Sheets (CSS). JavaScript is an object-based language based on the ECM 262 specification (“ECMAScript Language Specification 3rd Edition”). JavaScript’s functional language features allow functional programming, which may lead to more compact code. For more information on functional programming, see (“Functional programming in the real world”). Some language features like E4X (“ECMAScript for XML (E4X) Specification”) and XSLT are not available on all browsers. CSS is used to style HTML elements of a Web page by applying properties.

Runtime Environment The runtime environment of HTML- and JavaScript-based applications consists of the rendering engine and the JavaScript interpreter. The rendering engine renders the visual representation of the HTML and CSS code and provides access to the DOM. The JavaScript interpreter parses and interprets the embedded JavaScript code and accesses the DOM provided by the rendering engine to manipulate the page display. The runtime is heterogeneous. There are serious differences between the available browsers, even between versions of the same browser (“Microsoft’s Interoperability Principles and IE8”).

Extension Challenges Describing how to extend the many different frameworks would go beyond the scope of this chapter. The one challenge common to all the extensions of the various frameworks is the heterogeneous runtime environment mentioned before. Creating widgets requires writing HTML, CSS, and JavaScript code for and testing of all the browsers to be supported.

Market Penetration Most, if not all Internet-connected client computers have support for HTML, CSS, and JavaScript, because all common client operating systems9 include a Web browser. Most installed browsers are AJAX enabled; thus, they allow development of RIAs.

Beyond the Browser Since Internet Explorer 4.0, Microsoft has supported HTML Applications (HTA). HTAs are run like every other executable on Windows; thus, they have access to the file system and other privileges. The available platform features depend on which version of Internet Explorer is available (“Introduction to HTML Applications (HTAs)”). Adobe has developed a runtime to build RIAs that deploy to the desktop. It is called AIR and supports development with HTML and JavaScript as well as with Flash and Flex. Unlike a browser or the Flash plug-in, AIR grants executed applications access to the file system.

Interoperability Most plug-ins that can be embedded in Web pages can be accessed using JavaScript, and vice versa. For example, Java enables accessing JavaScript functions and the DOM (“Java-to-Javascript Communication”). A developer can also define methods that can be called from JavaScript

141

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

(“JavaScript-to-Java Communication (Scripting)”). The External Interface class of the Flex framework allows Flex applications to access any JavaScript function and JavaScript to access defined Actionscript functions. Silverlight has a similar functionality.

Supported Devices

Separation of Design and Logic

Silverlight offers a browser plug-in based on the Windows Presentation Foundation (WPF). This text will mainly focus on aspects only found in Silverlight 2.

Usually JavaScript and HTML are mixed together in one file and JavaScript code is written directly into the event attributes of the tags. While the described code layout can cause readability problems, it is possible to assign all event handler functions without any JavaScript code in the HTML. Libraries like jQuery10 and Dojo behavior11 make it easier to access DOM nodes and assign event handler functions. Frameworks often use some kind of templating mechanism to separate JavaScript code from HTML code. Dojo, for instance, uses templates to separate the implementation of widgets from the HTML code.

Supported Media Types Normally, the only media types directly supported by browsers are images in the formats GIF, JPEG, and PNG12. Safari and Firefox also support a subset of Scalable Vector Graphics (SVG), an XML vector format. Internet Explorer supports VML, which was mentioned above. Other media types are supported through plug-ins, if available for the particular operating system and browser.

Installation Browsers are installed with all relevant desktop systems. Microsoft Windows comes with Microsoft Internet Explorer, Mac OS X comes with Safari, and the available Linux distributions mostly come with Firefox and / or other browsers with the Mozilla Gecko rendering engine.

142

Generally, no special devices are supported without special plug-ins being installed.

silverlight

Tool Support XAML (eXtensible Application Markup Language) is unlikely to be written by hand, but will be created by specialized software. Microsoft Visual Studio helps to create basic user interfaces with XAML from a developer’s point of view. Applying a sophisticated design should be done by designers using Microsoft Expression Blend. Thus, Silverlight and XAML allow separating design and logic, as well as giving developers or designers an environment they are used to (MacDonald, 2007, page 22).

Availability Microsoft provides a Silverlight player for Windows and Mac OS in Version 1.0 and 2.0 beta (as of March 19, 2008). The Mono project13, which is supported by Novell, is developing a compatible open source alternative to the Silverlight player, called Moonlight14. Microsoft presented Silverlight 1.0 for Mobile on MIX08, which is restricted to JavaScript-based Silverlight 1.0 content. Nokia has announced Silverlight support on S60 on Symbian OS and Series 40 (“Nokia to bring Microsoft Silverlight powered experiences to millions of mobile users”). Neither the free SDK for one of the two Silverlight versions, nor the tools Visual Studio and Expression Blend are available for any platform

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

other than Windows. This means that developers and designers are tied to the Windows platform, at least at the moment. The Mono project is planning to integrate an XAML designer into MonoDevelop, the Integrated Development Environment of the Mono project, which is based on Alan McGovern’s Lunar Eclipse (“MonoTorrent”).

Available APIs, Libraries, and Functionalities Silverlight comes with a base class library, which is a compatible subset of the full .NET framework that includes collections, IO, generics, threading, globalization, XML, local storage, cryptographic services, libraries for the definition of global methods and types, generation of assemblies at runtime, events and delegates, and more (Guthrie, 2008) (“Common Language Runtime and Base Class Library in Silverlight”). Many of the functionalities mentioned above are known from other languages like Java. Silverlight libraries allow accessing a so-called isolated storage in which a partial trust application can store files. Thus, an application can store caches and user settings on the user’s machine. Another major feature is the .NET Language-Integrated Query (LINQ) (“LINQ: .NET Language-Integrated Query”), which turns out to be a general-purpose query language extension of the C# language. LINQ looks much like SQL. It is used to query XML as well as other data. The 3D namespace of the .NET framework 3.5 is missing in Silverlight’s base class library. Silverlight features rich network support, including support for calling REST, SOAP, POX, RSS, and standard HTTP services. Cross-domain network access and networking sockets are also included (Guthrie, 2008).

Language Characteristics As of Silverlight version 2.0, the Common Language Runtime (CLR)15 is included. This allows using every language supported by the .NET framework with Silverlight, including C#, Python, and Ruby. Since C# is Microsoft’s preferred language for the CLR, this text will focus on it. C# was developed by Microsoft and has been standardized by the ECMA (“C# Language Specification (ECMA-334 4th Edition)”) and the ISO (“ISO/IEC 23270:2003”). It is a high-level language similar to Java. Other useful language features are LINQ, delegates, enums, structs, and generics. User interfaces of Silverlight applications are defined using a language called eXtensible Application Markup Language (XAML16). Because of the hierarchical nature of XML, an XML-based language is a good choice for defining GUI component trees.

Runtime Environment The runtime environment of Silverlight is Microsoft’s CLR, an implementation of the Common Language Infrastructure (CLI) (“Common Language Infrastructure (CLI) (ECMA-335 4th Edition)”). The CLI defines an infrastructure that is able to execute multiple high-level languages. The languages are compiled into the Common Intermediate Language (CIL), the instruction set understood by the Virtual Execution System (VES). The infrastructure allows assemblies to run without modification on every platform the infrastructure is available on. In the managed environment, a garbage collector does automatic memory management. To increase execution speed, the infrastructure includes a Just-in-Time (JIT) compiler.

143

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

Extension Challenges Silverlight and the technologies used are well structured; thus, it is relatively easy to create custom UI controls. A ControlTemplate written in XAML defines the user interface of the control. The event handling and controller logic can be embedded into the XAML file, but placing it into a partial class is far more readable and maintainable (“Creating Custom Controls for Silverlight”). Unlike HTML and JavaScript, Silverlight applications are executed on a homogeneous platform; thus, testing a large number of alternative runtime environments is not necessary. The only compatibility problems that might occur affect Moonlight, the upcoming open-source implementation of Microsoft Silverlight.

access other plug-ins from Silverlight embedded in the same page (“How to: Call Managed Code from JavaScript”) (“Accessing the HTML DOM from Managed Code”).

Separation of Design and Logic Partial classes allow splitting a class and prorating it over several files. Since an XAML file is usually translated into a class, a partial class can be used to add methods to the generated class. This way, no programming code has to be in XAML files. Event handlers for certain events can be specified with the corresponding attribute, e.g., the Click attribute is given the name of the method for handling the click event, which is defined in the partial class.

Market Penetration

Supported Media Types

There are no official numbers from Microsoft at the time of this writing, but the demand for Silverlight developers is low (Lai, 2008), meaning that only few companies are creating Silverlight content at all.

The Microsoft Silverlight plug-in has built-in support for various media formats. The Windows Media Audio (WMA) format is supported, as is MP3 audio. The Silverlight plug-in also supports the WMV7-9 video codecs. WMV9 is Microsoft’s implementation of the standard VC-1 codec. VC-1 codec enables 720p HD Movies. Progressive downloading18 and streaming19 are also supported.

Beyond the Browser There is no desktop runtime environment available for Silverlight, except for the complete .NET framework, which is only available on the Windows platform.

Interoperability Silverlight allows accessing so-called managed code17 from JavaScript and vice versa. A developer is able to access properties and methods of managed code from JavaScript and to connect managed methods with JavaScript events. On the other hand, managed code can access the DOM and the access properties and methods of the DOM. JavaScript functions can be connected to managed events. JavaScript may also be used to

144

Installation The Silverlight version 2.0 (currently beta as of March 18, 2008) runtime for Windows has a file size of 4.4 MB. A developer embedding a Silverlight movie into a Web page is able to provide an attribute called ‘pluginspage’ to point to a download location for Microsoft Silverlight. Microsoft does some player detection using JavaScript, which embeds the movie if the player is detected20.

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

Supported Devices Supported devices are unknown at the time of writing. See (“Silverlight FAQ”) for up-to-date information.

flex Flex is a framework originated by Adobe (formerly Macromedia) to enable a more developer-like approach to creating Flash-based applications than Flash CS Professional, which aims at designers wanting to create animated Web sites and other animated content.

Tool Support Since Flex is an Adobe product, Flex Builder provides the most extensive support for Flex development, such as syntax highlighting, code-assist, life-error highlighting, refactoring, debugging, and profiling. Flex Builder also includes a GUI designer for visually creating Flex-based GUIs. The GUI designer creates MXML code, an XMLbased language used to define UI component trees. Source code editors are available for MXML and Actionscript 3.0, the language Flex is based on. Another IDE with Flex support is IntelliJ IDEA, but it can only be used as an editor, including code-assist and syntax highlighting (“IntelliJ IDEA, JavaScript Editor”). In addition to Flex Builder, Adobe offers a large lineup of tools for changing the appearance of and designing new Flex components. Since Flex 3, Adobe has offered the Flex Skin Design Extensions for Fireworks CS3, Flash CS3 Professional, Illustrator CS3, and Photoshop CS3, which allow creating skins for Flex components that offer more options for changing the visual appearance of components than styles21. For Flash CS3 Professional, Adobe offers an extension called Flex Component Kit. The kit allows creating Flex components with Flash CS3 Professional.

Currently, Adobe is working on a new software similar to Microsoft Expression Blend, codenamed Thermo (“Thermo”). The Flex SDK is available as open source under the Mozilla Public License (MPL). It features a compiler and a debugger for MXML and Actionscript 3.0 and the Flex framework, as well as the core Actionscript libraries.

Availability The Flex framework is based on the Adobe Flash player. For Flex versions 2 and 3, Flash player version 9 is the minimum requirement. Directly supported by Adobe are the following platforms (“Adobe Flash Player: System Requirements”): Windows 98, ME, 2000, 2003 Server, XP, Vista, Mac OS X 10.1.x to 10.4.x, Red Hat and SUSE Linux, Solaris 10. Although Flash Lite 3 is available for many cell phones, Flex version 2 and higher is not supported due to Flash Lite’s restriction to Flash 8 content. RIAs for Flash Lite 3 can be developed using the Flash Professional authoring environment and Actionscript 2.0 instead of Actionscript 3.0. The Flex 3 SDK is available for Windows 2000, 2003 Server, XP, Vista, Mac OS X 10.4.7-10.4.10 and 10.522, Red Hat, SUSE Linux, and Solaris 9 and 1023. The Adobe website provides different kinds of information for the Windows platform (“Adobe - Flex 3: System requirements”). Flex Builder 2 and 3, Adobe Fireworks CS3, Illustrator CS3 and Flash Professional CS3 are available for Windows and Mac OS X only. Since Flex Builder is based on Eclipse and the SDK is available for Linux, it should not be a problem for Adobe to provide a Linux version in the future.

Available APIs and Functionalities Besides a rich pool of UI controls, Flex comes with the functionalities of Actionscript 3.0, the Flex API, and Flash Player API. Actionscript

145

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

3.0 features an XML language extension called ECMAScript4XML (E4X) (Moock, 2007, page 353), which offers easy access to XML data and allows selecting XML elements. The Flash 2D display API allows dealing with interactive visual objects, bitmaps, and vector content (Mook, 2007, page 457). Further features are: Animation API (Mook, 2007, page 610), effects and transitions (Kazoun and Lott, 2007, page 232), back-button handling24, data bindings (Kazoun and Lott, 2007, page 268), RPC APIs25 and sockets, validation and formatting (Kazoun and Lott, 2007, page 288), and loading external content (Mook, 2007, page 762). Flash player allows an SWF file to store data locally on the user’s computer. The default maximum data that can be stored is 100 kb, but the user can agree to store more (“Flash Player Help”, “Local Storage Settings”).

Language Characteristics Because Actionscript 3.0 (AS 3.0) follows the ECMAScript Edition 4 specification, which is currently under development, most of the statements on JavaScript in section “AJAX” are correct in this case, too. But AS 3.0 has learned the advanced features of modern object-oriented languages. AS 3.0 features single inheritance, interfaces, data types, namespaces, metadata, and exception handling. Although AS 3.0 is a typed language, it still has the dynamic abilities of ECMAScript Edition 3. A variable can be defined without a type26 or with a wildcard. AS 3.0 allows adding instance variables and instance methods at runtime (Mook, 2007, page 279). Something similar to Ruby’s mixins (“Programming Ruby”) is also possible (“Ruby-like Mixins in Actionscript 3.0”). Using interfaces makes the AS 3.0 “mixins” type-safe. E4X extends Actionscript 3.0 with XML. It allows using XML directly in the source code and provides convenient handling of XML data (Mook, 2007, page 353).

146

MXML, the markup language of Flex, is an alternate way of defining a class. It allows defining component trees declaratively (Kazoun and Lott, 2007, page 43). Actionscript features garbage collection, like most modern languages that target a virtual machine.

Runtime Environment The runtime environment of Flex is Flash player 9 and higher. Flash player includes the Display API (Mook, 2007, page 457), among other things. Flash player also includes two virtual machines, but only the second one – the AVM2 – can run Flex 2 and 3 applications. Actionscript 3.0 bytecode runs in the new AVM2 (“Adobe Flash Player: Features”) virtual machine included in Flash player 9 and above and Adobe AIR 1.0 and above. AVM2 is open-sourced under the name Tamarin (“Tamarin Project”). It features a Just-in-Time (JIT) compiler (“Adobe/ Mozilla Tamarin Project Frequently Asked Questions”) to increase execution speed by creating native code for a particular hardware platform.

Extension Challenges A custom Flex component is a subclass of UIComponent. Custom components are created using MXML or Actionscript. Subclasses of container components with children are called composite components. If a component is instantiated using an MXML tag, the attributes can specify events, styles, and values of properties. As with Silverlight, the applications are executed in a homogeneous environment. Some small problems seem to exist. For example, Flash player on Mac OS X is not able to handle the mouse wheel27.

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

Market Penetration

Separation of Design and Logic

Since Flex 2 and Flex 3 need, as a minimum, Flash player 9, only the market penetration of version 9 and above matters. At the time of this writing, Flash player 9 is the latest version available and has a market penetration of 95.7% in the mature markets28 and 93.3% in the emerging markets29 (“Adobe Flash Player Version Penetration”).

An MXML file is translated into a class and mx:Script allows embedding Actionscript code into an MXML file, such as methods and properties. The easiest way to separate MXML and Actionscript code is to use the MXML counterpart of the reserved word include, mx:Script with the attribute source to specify the Actionscript file to be included. This is an approach similar to C#’s partial classes used in conjunction with an XAML template. Another, but more complex, way to separate MXML and Actionscript is to use Adobe’s framework Cairngorm. Cairngorm is called a microarchitecture and provides an Actionscript-like implementation of J2EE blueprint patterns (“J2EE Patterns Catalog”), like the Front Controller and the Business Delegate.

Beyond the Browser Since the release of AIR 1.0 on February 25, 2008, a desktop runtime environment for Flex applications has been available. Flex on AIR has some additional functionalities. Besides the Flash player, Adobe also packaged the WebKit rendering engine from Apple’s Safari browser. This makes an HTML component available to Flex, which has the complete rendering capability of a modern rendering engine. Also, AJAX-based applications can run in AIR, and can also be combined with Flex applications. Applications running in AIR are allowed to access the local file system (“Adobe AIR Local File System Access”) and enable drag and drop from the desktop or other programs (“Flickr Floater”).

Interoperability Flex offers three ways of data communications on the client: local connections, shared objects, and external interface. Local connections allow .swf files to communicate as long as they are running on the same machine, no matter in what environment. Shared objects allow storing locally shared objects on the client, which can be loaded the next time the application is running. The data is stored in the meantime. Finally, the external interface allows accessing the .swf file from the host environment and vice versa. In the case of the Flash player running in a browser, the external interface allows interacting with JavaScript (Kazoun and Lott, 2007, page 355).

Supported Media Types Flash player supports various video codecs up to HD. It supports H.263 playback and encoding, H.264, and On2 VP6 playback. The supported audio formats are MP3 and HE-AAC (“Datasheet Adobe Flash Player 9”). Additionally, Flash player supports PNG, JPG, and GIF formats for displaying bitmap images. SWF files can be embedded and loaded at runtime, but SVG files can be embedded with Flex at compile time only30 (“Embedding Application Assets”).

Installation The Flash player’s installation package of the latest Windows version 9,0,115,0 (on March 18, 2008) is 1.5 MB. Installation is very easy. Adobe AIR is a separate download and the file size is 11.2 MB for the latest version 1.0 (on March 18, 2008). The attribute pluginspage mentioned in the installation paragraph of Silverlight is available for Flash, too. Adobe offers the Flash Player

147

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

Detection Kit, which includes Express Install. Express Install features a player-based installation process, which installs the Flash player and returns the user to the page that requested the plug-in (“Flash Player Detection Kit”). Adobe ships an interesting possibility for a seamless installation of AIR applications called badges. Badges allow installing an AIR application via an SWF embedded in a web page, regardless of whether AIR is installed or not. If AIR is missing on the system, it is automatically installed with the AIR application (“AIR Install Badges”).

Supported Devices The Flash player supports audio output, mic input (“Flash Player Help”, “Microphone Settings”), and video input through a camera (“Flash Player Help”, “Camera Settings”).

JavafX “The JavaFX family of products is based on Java technology, designed to simplify and speed the creation and deployment of high-impact content for a wide range of devices. JavaFX technology enables developers and designers to create and deploy consistent user experiences, from the web page to desktop to mobile device to set-top box to Blu-ray Disc” (“JavaFX Technology FAQs”). JavaFX comes with a new scripting language called JavaFX Script, which has a different syntax than Java. It can be executed in an interpreted mode, but may also be compiled directly to bytecode for the JVM.

Tool Support Tool support for the Java programming language is very extensive, since many IDEs have extensive support. Some of the best known tools are Eclipse, NetBeans, and IntelliJ IDEA. JavaFX plug-ins are available for all of them. The JavaFXPad editor31 and the NetBeans plug-in32 feature realtime render-

148

ing of JavaFX Script. Since JavaFX is young and no stable release is available at the time of writing, the features of the plug-ins are limited. Java support of the mentioned IDEs includes syntax highlighting, code-assist, life-error highlighting, refactoring, debugging, and GUI designing33. Currently, no design tools are available for JavaFX, except for the realtime rendering mentioned above. Sun is putting “a lot of effort into interoperability with Adobe tools”, because designers know them well and have worked with them for years (“Sun’s JavaFX tools to interop with Adobe”).

Availability Java itself is available for nearly every client and server operating system. Sun directly supports Windows, Linux, and Solaris (“System Requirements for JRE 6.0”). Apple delivers its own Java versions with Mac OS X; the latest version at the time of writing was Java Standard Edition (SE) 5.0. Java SE 6.0 for Mac OS X was available as a developer preview (“Java”). Since Java was released as open source in mid-2007, Java can be ported to any platform. Java is also available on many mobile phones. Java Micro Edition (ME) is a slim version of Java SE suitable for the limitations of mobile devices (“Java ME at a Glance”). Sun has announced a new mobile operating system, called JavaFX Mobile, built on top of a Linux kernel and providing a built-in Java Virtual Machine (JVM) and a JavaFX environment (“JavaFX Mobile - Overview”). Java SE 5 or 6 is needed for JavaFX Script development. JavaFX technology is planned to be made available for the Java ME profiles Connected Limited Device Configuration (CLIC) and Mobile Information Device Profile (MIDP).

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

Available APIs and Functionalities JavaFX can access the complete class library of the host Java Runtime Environment (JRE), thus it depends on the JRE which APIs are available. On a Java SE, a wide range of APIs is available including networking, IO, security, cryptography, formatting, regular expressions, threading, and more. Besides the bundled class library, many thirdparty libraries are available. For example, the Lobo Project offers a complete HTML rendering engine called Cobra34, the Apache project offers a lot of libraries for XML processing and many other common tasks35, and Java3D offers 3D capabilities36.

Language Characteristics Java is a modern programming language. It is object-oriented and statically typed. Java features classes, interfaces, and exception handling. Memory management is done by a garbage collector. JavaFX Script is a new scripting language targeting the JVM. Like Java, JavaFX Script is statically typed and object-oriented. Unlike Java, JavaFX offers multiple inheritance. Although complete programs may be written in JavaFX Script, the key concepts were designed with user interfaces, graphics, and animation in mind (“JavaFX != JavaFX Script”). Object literals are used to declaratively instantiate classes, which allows defining UI component trees in a readable manner. Object literals are used for the same purpose as XAML and MXML. The reserved word bind allows binding variables, attributes of objects, value expressions, or even return values of operations37 to a certain attribute. This means that the latter is updated every time the bound value changes. Bindings allow connecting UI components declaratively. Instead of class constructors and getters and setters, JavaFX offers SQL-like triggers. Triggers are declared to fire on certain events like insertion, deletion, and replacement of

data. For further information, see (“The JavaFX Script Programming Language”).

Runtime Environment The runtime environment of Java includes the Java Virtual Machine (JVM), a stack-based virtual machine, and the class library. Currently, there are several editions available. Java Micro Edition (ME) features many different profiles and is dedicated to embedded and mobile devices. The Java Standard Edition is appropriate for desktop computers. Sun plans to drop Java ME in a few years in favor for Java SE, because mobile devices are getting enough power (“Sun starts bidding adieu to mobile-specific Java”). The virtual machine runs so-called Java bytecode. Unlike Flash player’s AVM2 and Microsoft’s CLR, the JVM is a HotSpot VM (“Java SE Hotspot at a Glace”). A HotSpot VM identifies code “worthy” of being optimized and compiled into machine code, instead of compiling the whole application. JavaFX applications are embedded into the browser using the Java plug-in. The special flavor of a Java application to run in the browser is called Java Applet. Compiled JavaFX Script runs directly in the JVM, but JavaFX Script may also be executed by an interpreter, which is written in Java and can be embedded into Java programs.

Extension Challenges JavaFX Script user interface components inherit from Widget. Since UI components are normal JavaFX classes, modification of existing UI components is simply done by extending the components class using inheritance. Composite UI components are usually created by extending CompositeWidget, composite canvas components by extending CompositeNode. To build up the composite component, the object literal syntax may be used in the corresponding compose method.

149

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

Market Penetration Available numbers differ. Adobe sees Java on 84.6% (“Flash Player Penetration”) of all Internetenabled desktops in mature markets38. Danny Coward states that 91% of all PCs were running Java in June 2007. Six months after the release of Java SE 6, 13% of all PCs were running Java SE 1.6, according to him (Coward, 2007, slide 11).

Beyond the Browser Java SE features Java Web Start, which enables deployment of standalone applications over networks with a single click. A Java Network Launch Protocol (JNLP) file specifies the files to be downloaded and the main class. After downloading is finished, the application is started immediately.

Interoperability

with Flash and Silverlight, which both support high-quality movies up to HD with their modern codecs (“Java Media Framework API”). For version 1.0, extended support for high-quality audio and video is planned39.

Installation Sun offers a JavaScript library to facilitate the installation process if a user has no JRE or if the installed version is too old (“Deployment Toolkit”). The JRE is the heaviest download of all discussed technologies, with 15.18 MB for the multi-language Windows version. The developers from Sun refer to the installation process as slow and complicated and to the startup time as poor (Coward, 2007, slide 26). Sun plans to make the next version modular, and faster to start, and wants to improve the installation experience.

The Java browser plug-in provides an easy way to access the DOM of the embedding Web page and to call JavaScript functions (“Java-to-JavaScript Communication”). It is also possible to access properties and methods of applets (“JavaScriptto-Java Communication”).

Supported Devices

Separation of Design and Logic

Alexey Gavrilov created a benchmark called Bubblemark40, which offers implementations of the same animation for different platforms. The following were chosen for a comparison: DHTML, Silverlight with interpreted JavaScript (SL JS), Silverlight with Common Language Runtime (SL CLR), Flex running on Flash player 9, JavaFX Script interpreted (JFX), JavaFX Script optimized (JFX opti.), and Java Swing. The results of the Mac OS X benchmarks are shown in Table 1, the results of the Windows benchmarks in Table 2. The Mac test machine was an Apple MacBook Core Duo 2GHz running Mac OS X 10.4. The Windows test machine was a custom PC driven by an AMD 2500+ Barton with an ATI 9600 XT graphics card running Windows XP SP2. The

JavaFX Script does not differ that much from Java when it comes to action listeners. In JavaFX, a function is used as an action listener, instead of the usually used inner classes of Java. The strategies for separating the model from the view are the same in both languages. An example strategy could be to reduce the code in the action operations to a single call of a method of the model.

Supported Media Types With JMF, Java applications and Applets can playback video and audio. While JMF supports a wide range of codecs, it still cannot compete

150

No out-of-the-box support for webcams and microphones is available. Audio output is possible.

Performance tests

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

test browser on Mac OS X was Safari 3.1 and IE 6 on Windows. Silverlight version 1.1 alpha was used, because the benchmark for Silverlight CLR was incompatible with Silverlight 2.0 beta. Java on Mac OS X was the bundled Java SE 5.0; Windows had Java SE 6.0 installed. Unfortunately, there is no compiled JavaFX Script benchmark on the site, but the optimized version and the Java Swing version should provide an indication of how fast a compiled JavaFX version would be. The first thing to note is that the optimized JavaFX version reaches significantly higher frame rates than the one that is not optimized. The class that does calculations for collision detection is implemented as a Java class and therefore is compiled and not interpreted. The results of a completely compiled version should be somewhere between the optimized JavaFX Script benchmark and the Java Swing benchmark, since JavaFX Script’s UI components are based on Swing and Java2D. It is obvious that Apple highly optimized the JVM on Mac OS X. Compiled JavaFX should be a high-performance solution on Windows and Mac OS X. Safari’s rendering engine WebKit also offers high performance. Adobe uses the same rendering engine in its AIR.

As expected, Silverlight with .NET CLR and JIT compiled assemblies is faster than the interpreted JavaScript Silverlight. Hence, it is all the more disappointing that Silverlight CLR performs so poorly on Mac OS X compared to Silverlight JavaScript. Flex seems to be slow when the frame rates with 16 balls are compared, but if all results are taken into account, Flex becomes more competitive the more balls are displayed. But still, the results of Flex are disappointing compared with DHTML in Safari and Internet Explorer, if taking into account that Flash player 9 features a JIT compiler for Actionscript 3.0 assemblies.

ConClUsion And oUtlooK This chapter provided an overview of a sample of RIA technologies. A number of aspects have to be considered when the right technology for a certain project has to be chosen. First of all, the audience and the purpose of an application have to be determined. If the audience is the totality of all Internet users, then only those technologies can be taken into consideration that are available on enough client computers. If search engine

Table 1. Bubblemark Mac OS X results DHTML

SL JS

SL CLR

Flex

JFX

JFX (opti.)

Java Swing

16

94 fps

86 fps

75 fps

55 fps

15 fps

64 fps

185 fps

32

88 fps

52 fps

44 fps

46 fps

6 fps

42 fps

184 fps

64

65 fps

26 fps

24 fps

29 fps

3 fps

22 fps

134 fps

128

25 fps

11 fps

12 fps

15 fps

< 1 fps

12 fps

82 fps

Table 2. Bubblemark Windows results DHTML

SL JS

SL CLR

Flex

JFX

JFX (opti.)

Java Swing

16

31 fps

54 fps

65 fps

38 fps

7 fps

27 fps

52 fps

32

15 fps

25 fps

35 fps

21 fps

3 fps

17 fps

30 fps

64

6 fps

10 fps

25 fps

9 fps

1 fps

10 fps

17 fps

128

3 fps

4 fps

18 fps

4 fps

< 1 fps

5 fps

9 fps

151

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

marketing is important, naturally, plain HTML is the best solution, because search engines have been created with the specific characteristics of hypertext in mind. The use of AJAX techniques may cause a serious impact on the ability of search engines to index a website, because search engines do not evaluate JavaScript (“CSS, AJAX, Web 2.0 & Search Engines”). Although Google and other search engines are able to index Flash movies, usually the complete content is in one file. This prevents reasonable search engine optimization (“Get Flash Sites Ranked in Search Engines”). The dynamic contents of a Flash movie, Silverlight application, or Java Applet cannot be indexed by any search engine; thus, other techniques have to be applied (“Search enabling Silverlight and AJAX web applications”). One aspect to consider are the skills of the available developers. Windows .NET developers are better off with Silverlight, for instance. Another constraint of a project is the budget. Thus, it is important to minimize costs by choosing a technology that helps to achieve that goal. AJAX requires extensive testing on many different browsers on different operating systems, because of its heterogeneous platform. Especially on older browsers, it is likely to detect rendering problems. One advantage of AJAX is that it is available on nearly every Internet-connected desktop computer without the need to install a plug-in. Furthermore, plug-ins can be used to add additional functionalities and support for media types. The plug-in-based technologies are handicapped because they depend on a plug-in on the client computer. But if the appropriate plug-in is installed, a homogeneous platform is provided with only one environment to be tested. Silverlight is a new, but stable platform with good tool support. Currently, Silverlight is not widespread. JavaFX is also a relatively new platform. At the time of this writing, no stable release was available. A final release of JavaFX Desktop 1.0 is planned for the fall of 2008; JavaFX Mobile and TV 1.0 will ship in spring 2009 (“Sun offers JavaFX

152

road map”). The existing development tools lack many features; design tools are not available, but a tool for converting SVG into JavaFX does exist (“JavaFX SVG Translator Preview”). At the time of this writing, the only plug-inbased platform with sufficient market penetration and good tool support is Flash / Flex. This makes it the platform of choice for plug-in-based applications aimed at the general public. At the moment, Silverlight is only suitable for applications where it is possible to make sure that the plug-in will be installed on all client computers. JavaFX is currently not ready for productive systems and should not be used until a stable release is available. In the near future, offline Web applications will be widespread at least in companies. They offer features similar to desktop applications, but can be installed easily by accessing an URL. No administrator privileges are required for installation. Today Adobe Flash and AJAX are the dominating RIA platforms. But the other two competitors JavaFX and Silverlight may be at the heels of Flash and AJAX in the near future, if the installation is hassle-free. While Java already has a widely installed basis, for JavaFX, an update to Java 6.0 will be necessary. Silverlight as a new platform has nearly no installed basis. However, if the installation process is easy and seamless and Sun and Microsoft can convince some big players in the Internet business to use their platform (such as Youtube, which is using Flash right now), the basis can grow very fast. On mobile devices, Java is in the front. JavaME is installed on over 70% of today’s mobile handsets41. The Flash Lite player is also widespread, but it is not sufficient for running Flex applications. Microsoft has a mobile runtime, too. As mentioned before, Nokia agreed to include this runtime in future handsets. Current smart phone operating systems such as iPhone OS, Windows mobile, Google Android, and Symbian, one realizes that every one of them has its own incompatible programming model. RIAs

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

on smart and other phones would help to reduce the costs of making an application available on all relevant mobile platforms. Only the runtime has to be ported. Since today’s phones offer more and more services, to enable a broader range of application types, the runtime should offer access to special phone features, such as GPS. In our opinion, the mobile RIA market will become one of the most exciting markets for future RIAs. New packet-based data transfer technologies as well as current and future combined phone and data contracts will allow everybody to be “always on”. The data rates have increased continuously during the last few years. Today’s rates are higher than the home Internet access most people had a few years ago. The user experience of those applications is one of the major challenges for user acceptance. Apple’s iPhone currently leads the way in mobile user interface design and interaction. Future RIAs will be able to utilize advanced types of input devices such as multi-touch displays and to interpret user gestures. To keep the cell phones handy, they have to stay rather small. Thus, the space on the screen is limited. RIAs have to deal with this constraint by introducing new ways of organizing user interfaces. Zooming the presentation, like Apple’s iPhone Safari does, is no viable way for easily and effectively usable RIAs. One solution towards an optimal usage of the available screen space may be the context-sensitive rearrangement of the user interface. To help the user follow the transitions between the states of the user interface, these have to be animated. The concept of states and animateable transitions is already implemented in Adobe Flex. Another issue where no standard solution exists is searchability. While Google and Yahoo have extended their indexers in order to be able to process Flash movies, this is insufficient for RIAs. Indexers can only work with content included in the indexed documents. This is a common problem of all RIA platforms. RIAs load the bulk of the content at runtime. Because search engine

crawlers cannot navigate those applications, they are unable to index the content. In a project of Fraunhofer IESE called SOP (Software Organization Platform) 2.0 (Weber et al., 2008), we offer a hybrid user interface. A hybrid interface in terms of our implementation offers both an HTML and an Adobe Flex interface at the same time. Thus, search engines can crawl the HTML interface to explore the complete content. To enable the changeover from the HTML interface to the RIA interface, it must be possible to alter the state of the RIA by defining initialization parameters. If a user clicks on a search engine link, he is directed to the HTML document. The changeover to the RIA interface can proceed upon a trigger by the user or automatically. The criterion system presented in this chapter was originally created in order to select an appropriate RIA platform for the SOP 2.0 project mentioned above.

referenCes J2EE Patterns Catalog. (n.d.). Retrieved on March 28, 2008, from http://java.sun.com/blueprints/ patterns/catalog.html About Native, X. M. L. H. T. T. P. (n.d.). Retrieved on March 27, 2008, from http://msdn2.microsoft. com/en-us/library/ms537505.aspx Accessing the HTML DOM from Managed Code. (n.d.). Retrieved on March 28, 2008, from http:// www.silverlight.net/Quickstarts/Dom/DomAccess.aspx Accessing the HTML DOM from Managed Code. (n.d.). Retrieved on March 31, 2008, from http:// www.silverlight.net/Quickstarts/Dom/DomAccess.aspx Adobe AIR Local File System Access. (n.d.). Retrieved on March 28, 2008, from http://labs. adobe.com/wiki/index.php/AIR:Articles:Adobe_ AIR_Local_File_System_Access

153

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

Adobe Flash Player. Features. (n.d.). Retrieved on March 28, 2008, from http://www.adobe.com/ products/flashplayer/productinfo/features/

Coward, D. (2007). What’s new in java SE 7. In JAZOON07. Retrieved from http://www.scribd. com/doc/257998/Whats-in-Java-SE-7

Adobe Flash Player. System Requirements. (n.d.). Retrieved on March 28, 2008, from http://www. adobe.com/products/flashplayer/productinfo/ systemreqs/

Creating Custom Controls for Silverlight. (n.d.). Retrieved on March 28, 2008, from http://www. silverlight.net/Quickstarts/BuildUi/CustomControl.aspx

Adobe-Flex 3: System requirements. (n.d.). Retrieved on March 28, 2008, from http://www. adobe.com/products/flex/systemreqs/

CSS. AJAX, Web 2.0, & Search Engines. (n.d.). Retrieved on March 28, 2008, from http://www. seroundtable.com/archives/006889.html

Adobe/Mozilla Tamarin Project Frequently Asked Questions. (n.d.). Retrieved on March 28, 2008, from http://www.mozilla.org/projects/tamarin/ faq.html#avm2

Datasheet Adobe Flash Player 9. (n.d.). Retrieved on March 28, 2008, from http://www.adobe.com/ products/flashplayer/pdfs/Datasheet_Flash_Player_9_ue.pdf

Assets, E. A. (n.d.). Retrieved on March 28, 2008, from http://www.adobe.com/devnet/flex/ quickstart/embedding_assets/

Duhl, J. (2003). White paper: Rich Internet applications. (Tech. Rep. IDC). Retrieved from http://download.macromedia.com/pub/solutions/ downloads/business/idc_impact_ of_rias.pdf

Browserwatch. (n.d.). Retrieved on March 27, 2008, from http://www.w3b.org/trends/browserwatch.html C# Language Specification (ECMA-334 4th Edition).(n.d.). Retrieved from http://www.ecmainternational.org/publications/files/ECMA-ST/ Ecma-334.pdf Ceruzzi, P. E. (1998). A history of modern computing. The MIT Press. Comet: Low Latency Data for the Browser. (n.d.). Retrieved on March 28, 2008, from http://alex. dojotoolkit.org/?%20p=545 Common Language Infrastructure (CLI). (ECMA335 4th Edition). (n.d.). Retrieved from http:// www.ecma-international.org/publications/files/ ECMA-ST/Ecma-335.pdf Common Language Runtime and Base Class Library in Silverlight. (n.d.). Retrieved on March 28, 2008, from http://msdn2.microsoft.com/en-us/ library/cc221412(vs.95).aspx

154

ECMAScript for XML. (E4X) Specification. (n.d.). ECMA. Retrieved from http://www.ecmainternational.org/publications/files/ECMA-ST/ Ecma-357.pdf ECMAScript Language Specification 3rd Edition. (n.d.). ECMA. Retrieved from http://www.ecmainternational.org/publications/files/ECMA-ST/ Ecma-262.pdf Flash, P. D. K. (n.d.). Retrieved on March 28, 2008, from http://www.adobe.com/products/ flashplayer/download/detection_kit/ Floater, F. (n.d.). Retrieved on March 28, 2008, from http://www.adobe.com/devnet/air/flex/ articles/flickr_floater_04.html Functional programming in the real world. (n.d.). Retrieved on March 28, 2008, from http://homepages.inf.ed.ac.uk/wadler/realworld/ Garret, J. J. (2005). Ajax: A new approach to Web applications. Retrieved on March 28, 2008, from http://www.adaptivepath.com/ideas/essays/ archives/000385.php

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

Get Flash Sites Ranked in Search Engines. (n.d.). Retrieved on March 28, 2008, from http://www. clickz.com/showPage.html?page=3419561 Guthrie, S. (2008). Blog entry: First look at silverlight 2. Retrieved on March 28, 2008, from http:// weblogs.asp.net/scottgu/archive/2008/02/22/ first-look-at-silverlight-2.aspx Hakman, K. (2006). Retrieved on March 21, 2008, from http://www2.sys-con.com/webinararchive. cfm?registered=on&pid=wc_aw6_d1_s3_t2_ hakman Help, F. P. (n.d.). Retrieved on March 28, 2008, from http://www.macromedia.com/support/documentation/en/flashplayer/help/index.html How to: Call Managed Code from JavaScript. (n.d.). Retrieved on March 28, 2008, from http:// www.silverlight.net/quickstarts/Dom/ManagedCodeAccess.aspx Install Badges, A. I. R. (n.d.). Retrieved on March 28, 2008, from http://blogs.adobe.com/simplicity/2007/06/air_install_badges.html Intelli, J. IDEA, JavaScript Editor. (n.d.). Retrieved on March 28, 2008, from http://www.jetbrains. com/idea/features/javascript_editor.html#flex Introduction to HTMLApplications (HTAs). (n.d.). Retrieved on March 28, 2008, from http://msdn2. microsoft.com/en-us/library/ms536496(VS.85). aspx ISO/IEC 23270:2003. (n.d.). Retrieved on March 28, 2008, from http://www.iso.org/iso/ iso_catalogue/catalogue_tc/catalogue_detail. htm?csnumber=36768 Java, F. X. = JavaFX Script. (n.d.). Retrieved on March 28, 2008, from http://weblogs.java.net/ blog/joshy/archive/2007/09/javafx_javafx_s. html Java, F. X. Mobile-Overview. (n.d.). Retrieved on March 28, 2008, from http://www.sun.com/ software/javafx/mobile/index.jsp

Java, F. X. SVG Translator Preview. (n.d.). Retrieved on March 28, 2008, from http://blogs. sun.com/chrisoliver/entry/javafx_svg_translator_preview Java, F. X. Technology FAQs. (n.d.). Retrieved on March 28, 2008, from http://java.sun.com/ javafx/faqs.jsp Java ME at a Glace. (n.d.). Retrieved on March 28, 2008, from http://java.sun.com/javame/index. jsp Java Media Framework, A. P. I. (n.d.). Retrieved on March 28, 2008, from http://java.sun.com/ products/java-media/jmf/ Java. (n.d.). Retrieved on March 28, 2008, from http://developer.apple.com/java/ Java SE Hotspot at a Glace. (n.d.). Retrieved on March 28, 2008, from http://java.sun.com/javase/ technologies/hotspot/ Java-to-Javascript Communication. (n.d.). Retrieved on March 28, 2008, from http://java.sun. com/j2se/1.5.0/docs/guide/plugin/developer_ guide/java_js.html JavaScript-to-Java Communication. (n.d.). Retrieved on March 28, 2008, from http://java.sun. com/j2se/1.5.0/docs/guide/plugin/developer_ guide/js_java.html JavaScript-to-Java Communication (Scripting). (n.d.). Retrieved on March 28, 2008, from http:// java.sun.com/j2se/1.5.0/docs/guide/plugin/developer_guide/js_java.html Kazoun, C., & Lott, J. (2007). Programming flex 2. O’Reilly. Lai, E. (2008). Little demand yet for silverlight programmers. Retrieved on March 30, 2008, from http://www.computerworld.com/action/ article.do?command=viewArticleBasic&article Id=9066838

155

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

LINQ. NET Language-Integrated Query. (n.d.). Retrieved on March 28, 2008, from http://msdn2. microsoft.com/en-us/library/bb308959.aspx

Ruby, P. (n.d.). Retrieved on March 28, 2008, from http://www.ruby-doc.org/docs/ProgrammingRuby/html/tut_modules.html#S2

Lite, A.-F. (n.d.). Retrieved on March 28, 2008, from http://www.adobe.com/products/flashlite/

Ruby-like Mixins in Actionscript 3.0. (n.d.). Retrieved on March 28, 2008, from http://flexonrails. net/? p=73

MacDonald, M. (2007). Pro WPF: Windows presentation foundation in. NET 3.0. Apress. Mahemoff, M. (2006). Ajax design patterns. O’Reilly. Microsoft’s Interoperability Principles and IE8. (n.d.). Retrieved on March 28, 2008, from http:// blogs.msdn.com/ie/archive/2008/03/03/microsoft-s-interoperability-principles-and-ie8.aspx Model, D. O. (DOM). (n.d.). Retrieved on March 28, 2008, from http://www.w3.org/DOM/ MonoTorrent. (n.d.). Retrieved on March 28, 2008, from http://monotorrent.blogspot.com/2007/09/ so-summer-is-finally-at-end.html Moock, C. (2003). What is a flash MX component? O’Reilly Web DevCenter. Moock, C. (2007). Essential Aationscript 3.0. O’Reilly. Nokia to bring Microsoft Silverlight powered experiences to millions of mobile users. (n.d.). Retrieved on March 28, 2008, from http://www. nokia.com/A4136001?newsid=1197788 Norman, J. M. (2005). From Gutenberg to the Internet: A sourcebook on the history of information technology. Norman Publishing. Penetration, A. F. P. V. (n.d.). Retrieved on March 28, 2008, from http://www.adobe.com/products/ player_census/flashplayer/version_penetration. html Penetration, F. P. (n.d.). Retrieved on March 28, 2008, from http://www.adobe.com/products/ player_census/flashplayer/

156

Search enabling Silverlight and AJAX Web applications. (n.d.). Retrieved on March 28, 2008, from http://blogs.msdn.com/jhawk/archive/2007/05/23/searching-enabling-silverlightand-ajax-web-applications.aspx Silverlight, F. A. Q. (n.d.). Retrieved on March 28, 2008, from http://www.microsoft.com/silverlight/ overview/faq.aspx Sun offers JavaFX road map. (n.d.). Retrieved on May 21, 2008, from http://www.infoworld. com/article/08/05/06/Sun-offers-JavaFX-roadmap_1.html Sun starts bidding adieu to mobile-specific Java. (n.d.). Retrieved on March 28, 2008, from http:// www.news.com/8301-13580_3-9800679-39. html Sun’s JavaFX tools to interop with Adobe. (n.d.). Retrieved on March 28, 2008, from http://www. regdeveloper.co.uk/2008/01/24/javafx_tools_ adobe/ System Requirements for JRE 6.0. (n.d.). Retrieved on March 28, 2008, from http://java.com/ en/download/help/6000011000.xml TR10: Offline Web Applications. (n.d.). Retrieved on March 30, 2008, from http://www.technologyreview.com/read_article.aspx?ch=specialsecti ons&sc=emerging08&id=20245 Tamarin Project. (n.d.). Retrieved on March 28, 2008, from http://www.mozilla.org/projects/ tamarin/ The Counter.com. (n.d.). Retrieved on March 27, 2008, from http://www.thecounter.com/ stats/2008/February/browser.php

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

The dom and javascript. (n.d.). Retrieved on March 26, 2008, from http://developer.mozilla.org/en/ docs/The_DOM_and_JavaScript The DOM and JavaScript. (n.d.). Retrieved on March 27, 2008, from http://developer.mozilla. org/en/docs/The_DOM_and_JavaScript The JavaFX Script Programming Language. (n.d.). Retrived on April 9, 2008, from https://openjfx. dev.java.net/JavaFX_Programming_Language. html The XMLHttpRequest Object. (n.d.). Retrieved on March 27, 2008, from http://www.w3.org/TR/ XMLHttpRequest/ Thermo. (n.d.). Retrieved on March 28, 2008, from http://labs.adobe.com/wiki/index.php/Thermo Tim O’Reilly. (2005). What is Web 2.0. Retrieved on March 27, 2008, from http://www.oreillynet. com/pub/a/oreilly/tim/news/2005/09/30/what-isweb-20.html Toolkit, D. (n.d.). Retrieved on March 28, 2008, from https://jdk6.dev.java.net/testDT.html Weber, S., et al. (2008). Workshop on Learning Software Organizations (LSO), Rome, Italy. WHATWG FAQ. (n.d.). Retrieved on March 24, 2008, from http://wiki.whatwg.org/index.php?title=FAQ&oldid=2907#What_are_. E2.80.9CWeb_Applications.E2.80.9D.3F

Key terMs And definitions AJAX: AJAX stands for Asynchronous JavaScript + XML. It is not one technology, but a combination of technologies. These include HTML, Cascading Stylesheets (CSS), Document Object Model (DOM), XML, Extensible Stylesheet Language Transformations (XSLT), XMLHttpRequest, and JavaScript. HTML and CSS are used for presentation. DOM allows

manipulation of the presentation and interaction. XML and XSLT are used for data interchange and manipulation. XMLHttpRequest allows retrieval of data. JavaScript is used to define the underlying logic and interaction of the other technologies. All-Audience Applications: Applications potentially targeting every Internet user. Thus, this type of application has to take care that each potential user can access and use the application, regardless of which browser is installed on his system. If browser plug-ins are needed, only plug-ins with a very high market penetration are sufficient. Click-Wait-and-Refresh-Cycle: This term was coined by Kevin Hakman (2006). It describes the way users interact with traditional Web applications. A user clicks on a button or link, and the request is sent to the server and processed. The user waits until the results are returned to the Web browser, which refreshes the presentation. Rich Internet Application: Applications called RIAs provide a more intuitive, responsive, and effective user experience. This is done by utilizing user interface components and behaviors know from desktop applications. Rich User Experience: The experience of a user using traditional Web applications and websites is characterized by the Click-Wait-andRefresh-Cycle and the available set of user interface components. Thus, a rich experience is built up by adding additional interface components and behaviors and the Click-Wait-and-Refresh-Cycle is avoided by retrieving and presenting data from the server without refreshing the whole page. Web Application: A Web application is an application accessed over the WWW using a Web browser. It is built on Web standards. Additionally, proprietary Web technologies may be used.

endnotes 1 2

http://lobobrowser.org/ http://www.curl.com/

157

An Overview of and Criteria for the Differentiation and Evaluation of RIA Architectures

3

4 5 6

7

8 9

10 11

12 13 14 15

16 17

18 19

20

21

22

23

158

http://www.omnis.net/index.html? detail=overview http://labs.mozilla.com/2007/10/prism/ http://www.aptana.com Widget is the term commonly used to describe user interface components in the context of a JavaScript framework. Browser bugs are a source of problems with HTML and JavaScript development: http:// www. positioniseverything.net http://www.prototypejs.org Meaning: Windows, Mac OS X, and the common Linux distributions. http://jquery.com http://redesign.dojotoolkit.org/jsdoc/dojo/ HEAD/dojo.behavior IE 6 has problems with PNG transparency. http://www.mono-project.com http://www.mono-project.com/Moonlight The CLR is the virtual machine of Microsoft’s .NET framework. It is Microsoft’s implementation of the Common Language Infrastructure. Pronounced ‘Zammel’. Code that is running inside the CLR is called managed code. Play a video while it is still downloading. In combination with the Windows Media services platform See source code of http://www.microsoft. com/silverlight/ Flex uses Cascading Style Sheets (CSS) to style the appearance of components. Mac OS X is not on the list for the SDK, but on the list for Flex Builder 3, and Flex Builder 3 comes with the SDK. Compilers only.

24

25

26

27

28

29

30

31

32

33

34 35 36

37

38 39 40 41

Flex allows managing states, enabling backbutton handling in a way the user would expect from a browser. Allow accessing HTTP and SOAP services. This is not allowed by the compiler in strict mode. But there exists an easy workaround using JavaScript. US, Canada, UK, Germany, France, Japan. China, S. Korea, Russia, India and Taiwan. SVG is not supported by the Flash player, but by the Flex compiler. Link to the Java Web Start file of the Pad: http://download.java.net/general/openjfx/ demos/ javafxpad.jnlp JavaFX preview in NetBeans: http://wiki. netbeans.org/JavaFXPluginPreviewSpecification For Eclipse, an additional plug-in is needed. http://lobobrowser.org/cobra.jsp http://commons.apache.org/ http://java.sun.com/javase/technologies/ desktop/java3d/ An operation is the JavaFX counterpart of Java methods. JavaFX is used to distinguish between operations and functions. In the latest version of the language, there are only functions. US, Canada, UK, France, Germany, Japan http://www.javafx.com http://www.bubblemark.com/ http://java.sun.com/javafx/tutorials/jfx_nb_ getting_started/

159

Chapter 9

The Layered Virtual Reality Commerce System (LaVRCS): Proposing an Immersive Web X.0 Framework for E-Commerce Alan Rea Western Michigan University, USA

AbstrACt In this chapter, the author argues that virtual reality (VR) does have a place in e-commerce as a Web 2.0 application. However, VR is not ready to supplant standard e-commerce Web interfaces with a completely immersive VR environment. Rather, VRCommerce must rely on a mixed platform presentation to accommodate diverse levels of usability, technical feasibility, and user trust. The author proposes that e-commerce sites that want to implement VRCommerce offer at least three layers of interaction: a standard Web interface, embedded VR objects in a Web interface, and semi-immersive VR within an existing Web interface. This system is termed the Layered Virtual Reality Commerce System, or LaVRCS. This proposed LaVRCS framework can work in conjunction with Rich Internet Applications, Webtops, and other Web 2.0 applications to offer another avenue of interaction within the e-commerce realm. With adoption and development, LaVRCS will help propel e-commerce into the Web 3.0 realm and beyond.

introdUCtion Individuals use the Web to search for information, communicate with friends and family, form social networks, and entertain themselves with a plethora of multimedia and interactive applications. Businesses rely on the Web to manage information and extend global communication, as well as market products and services. Although the Web has become integral

within both business and personal contexts, Web 2.0 applications have the ability to extend our computing experiences beyond the standard “point and click” interface to which we have grown accustomed. This is especially important to businesses competing to attract new clients within the ever-changing digital landscape of global e-commerce.

DOI: 10.4018/978-1-60566-384-5.ch009

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

The Layered Virtual Reality Commerce System (LaVRCS)

e-Commerce today Businesses use the Web to market goods and services to people and other businesses. Consumers— businesses and people—increasingly look to the Web to provide choices and the means to make informed purchasing decisions. E-commerce transactions have grown faster than most predictions and continue to grow as more businesses offer and improve online offerings at a rate of over 19 percent growth each year (Loten, 2007). Businesses want to make their Websites easy to use and move people toward intended purchases (Cummins, 2002). Web usability experts work to simplify Web designs and usability so that potential customers find familiar navigation schemes and metaphors (e.g., the shopping cart) to simplify their purchases (van Duyne, et al., 2007). Still, studies examine abandoned shopping carts, unsatisfied e-commerce users, and businesses that have failed online even though they offered quality products or services (Eastlick, et al, 2006; Chen & Rea, 2004). There must remain some elusive criteria that have not yet been met to create an environment in which satisfied Web users explore goods and services and complete e-commerce transactions. Not all businesses can model themselves after Amazon.com and expect the same profits, so we must ask what an e-commerce site can offer to first attract users, then make them comfortable using the site, create trust that the goods or services will be as promised, and ultimately make them secure in their decision to purchase.

e-Commerce tomorrow Although e-commerce sites must first and foremost be secure and offer navigable Web pages and electronic catalogs, they also must allow users to choose how they want to explore the proffered goods and services. This should include an interactive medium so that users can experience

160

wares. The interactive medium most suited to offer another experiential layer is Virtual Reality (VR). Current Web 2.0 implementations, such as Rich Internet Applications (RIA) and Webtops offer another approach to Web application interaction. These feature-rich Web offerings allow users to interact with applications similar to what they are accustomed on their desktops and are increasingly adopted by users (Driver & Rogowski, 2007). Although RIAs and Webtops offer rich interaction, they do not offer the experiential VR layer. But VR is not quite ready to supplant the standard e-commerce Web interface with a completely immersive VR environment; hence the ongoing need for RIAs and Webtops. These accepted Web 2.0 applications will remain for some time and morph into 3D Web applications; however, we should also move to push the user experience into other realms, such as VR. Challenges abound as we attempt to fuse ecommerce and VR. E-commerce must rely on a mixed platform presentation to account for various levels of usability, user trust, and technical feasibility. E-commerce sites that want to implement VR in e-commerce (VRCommerce) must offer at least three layers of interaction: a standard Web interface, embedded VR objects in a Web interface, and semi-immersive VR within an existing Web interface. This system is termed the Layered Virtual Reality Commerce System (LaVRCS). In order to understand how LaVRCS is critical to allowing users to comfortably and effectively access and benefit from e-commerce, we must first define Virtual Reality (VR), then examine crucial implementations for an effective VR ecommerce (VRCommerce) system, and put forth the four levels of VR implementation. With this background we can delve into the challenges facing VRCommerce and the suitability of certain goods and services for e-commerce permutations. From here we will discuss the LaVRCS framework

The Layered Virtual Reality Commerce System (LaVRCS)

and architecture and how it can address these challenges. We will conclude with a discussion of the proposed LaVRCS implementation and its potential current and future permutations.

bACKgroUnd what is virtual reality? Most users are familiar with the graphical user interface (GUI) as the main method to interact with computers. The desktop metaphor, complete with trash cans and recycle bins, helps users take the familiarity of a physical desktop and place it on a computer screen. Users are comfortable with the WIMP (windows, icons, menus, and pointers) interface and this comfort extends onto the Web with Web browsers displaying text and images. A mouse click on a hyperlink allows a user to navigate a 2D Website. Much of this interaction is replicated with current RIA and Webtop applications. We move closer to VR with the 3D Web. However, in VR, methods exist for users to interact with computers beyond the GUI interface: voice recognition, and eye, head, and body tracking devices—all collectively referred to as biocontrollers—are a few. Others include 3D visualizations, biofeedback (e.g., haptics), and instruments that use stereo sound, smell, and even taste. Using human senses and the body to interact with computers brings us into the Virtual Reality realm. Many definitions have been put forth to describe VR. For our purposes a general definition put forth by Barnes (1996)—which is shared by many researchers—can be used: VR is the term used to describe advanced methods of involvement and interaction for humans with a computer-generated graphical (usually 3D) environment. Normally referred to as a VR “world,” this environment is experienced by a

human participant through the use of special VR equipment. Using this definition, we see that VR goes beyond what is currently offered to users on the Web. Even cutting-edge interactive applications such as those offered by Laszlo Systems with its interactive Web 2.0 desktop (Lazslo, 2009) do not offer the experiential component of a VR world.

virtual reality research Researchers have implemented and studied VR is a multitude of areas. VR is used in software engineering (Tecchia, 2006), military applications (Losh, 2006), manufacturing (Kirner & Kirner, 2005), and construction (Lipman & Reed, 2000). VR has also been used to treat phobias (Strickland, et al., 1997) and monitor walking techniques to help people improve ergonomic functionality (Whitman, et al., 2004). Educators and trainers use VR to effectively teach concepts, act out situations, and provide training at a distance (Arns, et al., 2006). Research has even been conducted on how VR can be used with PDAs (Pasman, et al., 2004). However, researchers only recently have conducted studies at how VR can be effectively applied in the e-commerce realm. These studies can be organized around four main areas: 3D object creation, world navigation, agent guidance, and system architectures.

3D Object Creation One of the greatest challenges in VR is making virtual items look and respond as real world items would. Virtual objects need to be imbued with physical properties that will mimic those in the real world. Users should not be able to walk through walls, pick up cars, or fly through the air unless these traits are required.

161

The Layered Virtual Reality Commerce System (LaVRCS)

Pai, et al. (2005) discuss a complex process to translate real-world objects into 3D virtual objects. Using a complex system of robotic measurement devices and cameras, the authors create an object that looks, reacts, feels, and sounds like the original object. In their discussion, they use examples of a stuffed tiger and a clay pot to illustrate the challenges that various types of objects create in a virtual realm. When one touches a stuffed toy, one feels a difference in texture than a clay pot. Their system translates this using haptic feedback. One also hears contact with an item differently. Moreover, poking a pot is a much different experience than poking a stuffed toy. Both the object (pot or toy) and the user react differently to each. The authors admit, though, that the hardware and software necessary to run their system is complex and expensive, they argue that for the interim similar systems could be built and then users could book studio time to use a system, because each system would be capable of many renderings a day. Moreover, they argue that over time the technology will find its way to the desktop computer with portable scanning systems. Whether or not one believes the cost of this technology would eventually reach a point for users to buy, businesses could invest in this technology if they saw a great return. Perhaps other businesses could become scan centers and cater to businesses that need catalogs of 3D objects. LaVRCS has the potential to accomplish this on a smaller scale with simple 3D objects.

World Navigation Creating and using believable 3D objects is crucial to create a believable VR world. However, once the world is created, users must be able to move throughout the world. VR system components, such as Head Mounted Displays (HMDs) and gloves can be used to simulate and direct locomotion. In an e-commerce environment, we must ask what types of locations a user might navigate.

162

A feasible VR application in e-commerce must allow users to work within a context that they are comfortable with yet still allow them to easily navigate the new VR world. One means to accomplish this is to implement familiar worlds and supplement 3D navigation with more familiar 2D maps. Mass & Herzberg (1999) implement this scenario in their VR Mall. Using a Java Applet interface, the VR mall allows users to view a shopping mall complete with storefronts and other items one finds in malls (plants, benches, etc.) while simultaneously charting their location on a 2D color-coded mall map. Chittaro & Ranon (2002) follow a similar approach using a single store by arranging all products on store shelves and labeling the isles much like one would see in the real world. Their AWE3D (Adaptive Web3D Interface) moves a step further in navigational aids by implementing Walking Products that take users to the desired item. These Walking Products are 3D item representations of signs with feet that guide users in the store.

Agent Guidance The AWE3D system’s Walking Products are a good example of agent guidance in VR worlds. Chittaro, et al. (2003) build on the Walking Agent concept to create a humanoid guide. After a discussion of failed VR navigational schemes, the authors present an autonomous agent based on H-anim specifications by the Web3D organization (Web3D, 2009). The specification allows the creation of a humanoid character capable of navigating a VR world as a Virtual Tour Guide. Users can follow the guide using their avatar—a virtual representation of themselves—and find the best path to their destination. Chittaro, et al. (2003) also note that the Virtual Tour Guide becomes a useful tool to help acclimate new users to navigating VR worlds. However, studies have shown that users want to maintain control in e-commerce situations

The Layered Virtual Reality Commerce System (LaVRCS)

whether they are in a VR world or surfing a 2D catalog (Phelps, et al., 2000; Hoffman, et al., 1999). A major part of this control is supplied by allowing the user to make choices and not be led or directed to a certain goal (Gammick & Hodkinson, 2003; Cummins, 2002). Autonomous agents, such as Chittaro & Ranon’s (2002) Walking Agents accomplish this by not appearing unless a user has asked for help, stopping whenever a user stops in the VR world, and disappearing once the user finds the product. Other VR worlds promote choice by offering different avenues to explore the e-commerce space, such as a Web interface and a VR world (Gammick & Hodkinson, 2003). Another means to allow promote user control and increase user comfort levels, and eventually trust, is to implement known interface design and technologies as either an embarkation point to the VR world or as a wrapper around the VR world. Plunging the average user into a new visual paradigm without preparation is not how an e-commerce site can promote repeat customers and referrals.

virtual reality system Architectures No matter what objects, worlds, or agents VR developers employ, they have a variety of hardware and software combinations from which to choose. Of course, the components and virtual worlds must be compatible; for example, one would not place a HMD on a user inside a CAVE environment. We will not explore each hardware and software combination as others have summarized this well (Stansfield, 2005). However, we must classify varying VR levels from the standard desktop interface to the completely immersive (e.g., CAVEs). We do this because successfully implementing VR levels within e-commerce sites depends on the correct balancing of Web and VR to make the largest pool of users the most comfortable with the experience.

As we move from the least immersive to the most immersive, note that system requirements (hardware and software) as well as a user’s technical skill must increase. Financial investment, sometimes by thousands of dollars, increases as well thereby limiting consumer access. Cost cannot be a prohibiting factor when attempting to reach the largest number of computer users.

Entry Level VR System (EVRS) An EVRS uses a personal computer system that is available to most users. A computer with sufficient hardware resources and a current operating system is adequate for EVRS operation. The EVRS will most likely be limited to a Window on the World (WoW) VR experience using the standard monitor, mouse, and keyboard. Open source VR software programs exist that will run on these systems, such as VR Juggler (Cruz-Neira, 2009). Software also exists for a small price. In this category are closed system software products. The immersive, interactive world of Second Life (Linden Labs, 2009) offers users the ability to create customized avatars and environments. Within Second Life, users can interact within communities, conduct business, or simply explore the diverse user worlds. Second Life offers an immersive model that, used effectively, can be used for business (Brandon, 2007). Still, many users are not ready for this immersive environment.

Basic VR System (BVRS) The BVRS uses the same hardware structure as the EVRS with the addition of input devices, such as a VR glove or a haptic or force feedback device, such as Nintendo’s Wii controller. PCs with fast CPU speeds, high-end graphics cards, and highresolution monitors fall into this category. These systems may also include a high fidelity sound system for a surround sound experience. The largest group of users who employ these systems are computer gamers. These systems can

163

The Layered Virtual Reality Commerce System (LaVRCS)

be used for intense experiences in VR games. They have the processing power to run many basic VR worlds and applications. The purchase price for these systems exceeds an average desktop computer and quickly escalates.

Advanced VR System (AVRS) AVRS systems are too expensive for the typical user. These systems require a substantial investment for memory, processing power, software, and hardware needs. Most likely these systems are minicomputers running specialized software and cost many thousands of dollars. Although common high-end hardware architectures and components can be used, most AVRSes employ extremely specialized devices that cost far beyond what an average computer user would purchase.

Immersive VR System (IVRS) IVRS systems are most likely used in research or the entertainment industry. These systems are what most people think of when they imagine VR. Though still not as elaborate as Star Trek’s Holodeck, these systems are not readily available to anyone except government, industry, and academic research labs.

on the roAd to vrCoMMerCe The Advanced and Immersive VR systems require substantial financial investment and, in cases such as the CAVE, substantial physical space. Most users will have neither the funds nor the willingness to commit to these systems. Thus we are left with either the entry-level VR system (EVRS) or the basic VR system (BVRS). Given that both of these are desktop (or perhaps laptop) computer systems, we can assume that most users will have access to either an EVRS or BVRS. Whether a user will have more than a

164

monitor, keyboard, and mouse is questionable. Some game players may have simple HMDs, but these are expensive and somewhat unreliable at the consumer level. Whatever the hardware choice, software for all e-commerce VR worlds must use technologies that already exist on users’ computers (Web browsers, Java) or are easily obtainable (VRML or Web3D plug-ins) at no cost.

suitable e-Commerce venues Before implementing VR into e-commerce, we must be aware that VR is not always necessary. For example, Amazon.com does well without VR because it primarily sells low- to mid-cost packaged goods, such as books, DVDs (movies and shows), and music CDs. Users know what to expect with a book, CD, or DVD format. Amazon. com has also minimized user risk by providing content samples of each. However, other items are not easily suited to a straight 2D e-commerce site. High-touch goods and services such as vacations, homes, and cars are purchased over the Web in far fewer numbers than books. One reason is because of the higher price point, but another is that users want to experience these items before purchasing them. Even with Websites using Flash technology to demonstrate and customize cars (e.g., Saturn.com) or tour homes for sale (e.g., pprmi.com), most users still want to physically examine the items before making a final purchase decision. It is here that RIAs and Webtops also fall short. Studies have shown that although users may research high-touch items on the Internet before buying them, the majority complete the purchase in the real world (Gammick & Hodkinson, 2003). Virtual Reality can assist in both high-touch and packaged goods and services. Packaged items are more easily assimilated into the e-commerce realm via Virtual Reality. Many existing platforms (Chittaro & Ranon, 2002; Mass & Herzberg, 1999) readily lend themselves to packaged goods and

The Layered Virtual Reality Commerce System (LaVRCS)

can be placed on shelves located in VR stores. The challenge, of course, is getting users to frequent these VR stores and malls given navigation challenges and user VR comfort levels. In contrast, high-touch goods are more of a challenge to implement in VRCommerce sites. Gammick & Hodkinson (2003) discuss issues of Website usability, gaining user trust levels, and allowing layers of VR experience. They detail the implementation of a VR e-commerce prototype called Beachtown that allows users to experience a travel destination before booking travel. Unlike other high-touch items (cars and homes), there is no physical means for a person to try the product without a trip to the location. The Beachtown prototype allows users to surf a 2D Website akin to typical tourist Websites. Lists of shops, restaurants, recreation, and hotels can be found and researched. However, users are also able to click on a VRML representation of the Beachtown boardwalk via a VRML plug-in in a Web browser. The researchers found that being able to view even a simulated stroll in a section of Beachtown increased the possibility that users would visit or book a vacation (Gammick & Hodkinson, 2003).

Challenges of vrcommerce Once we decide on a VRCommerce venue, we must consider how realistic it looks, how easy it is to navigate, what kind of guidance we offer users, and what type of computer system users need to interact with our world. Moreover, we must deal with both technical challenges and human factor challenges: learning curves, user acceptance, and trust.

Realism and Latency Researchers point to network latency issues in almost every discussion of VR system prototypes (Jay, et al., 2007; Chim, et. al., 2003; Chittaro, et al., 2003; Mass & Herzberg, 1999). Solutions to

network latency range from layered 3D objects (Chim, et al., 2003), designed to load and cache incrementally to lessen client-side and network requirements, to not sending any 3D data until the user fills out a form requesting certain 3D objects for their associated world (Varlamis, et al., 2004). Some studies simply ignore latency, as researchers are more concerned with other issues. However, no matter how a VR system deals with latency, it will always be a challenge in interactive systems like VRCommerce (Babaioff, et al., 2007). Users connect to the Internet via a variety of means: modems, mobile phones, or broadband connections with high-end workstations. VRCommerce sites must have the flexibility to accommodate every means of access.

Navigation, Learning, and Acceptance Because VRCommerce will not accommodate advanced or immersive VR systems, such as the CAVE, one would assume that using a mouse and keyboard to navigate an image on the screen would be straightforward. Research has shown this not to be the case with one study that documented over 80% of its participants suffering from at least a mild form of Virtual Reality Induced Symptoms and Effects (Wilson, 1997). Moreover, VRISE proved to be dehabilitating in over 5% of the participants. For most users, moving about in a VR world is not second nature and causes some sort of VRISE effect (Knight & Arns, 2006). VR developers have worked to establish more real world metaphors so that users can better identify with the VR world. AWE3D (Chittaro, et al., 2003) places users in a virtual store replete with products arranged on shelves, advertisements along the walls, and customized audio announcing store specials or piping music. AWE3D strengthens this metaphor by adapting a perspective (a view behind a shopping cart) that makes the user more comfortable. Theoretically, the more a user interacts with the world, the more customized and familiar it will become.

165

The Layered Virtual Reality Commerce System (LaVRCS)

Another VR system that emphasizes a real world metaphor is the VR Mall (Mass & Herzberg, 1999). In this system, the authors use a combination of Java and VRML to create an interface that allows merchants to place predefined objects in a space (store). Users can navigate the mall via a Java Applet using a mouse. A 2D maps shows them where they are, limiting confusion and making navigation more manageable. Although the VR mall is an older prototype, it has been used as a template for many newer VR applications in its approach. By integrating 3D with 2D maps, the authors have overcome user resistance to completely immersive VR realms and helped users accept 3D navigation. By only allowing select VR objects to be created, the authors have also simplified the process for merchants who wish to participate but do not have the technical skills.

Guidance and Trust Autonomous agents (Lees, et al., 2007) and Virtual Tour Guides (Zheng, et al., 2005; Chittaro et al., 2003) are effective to guide first-time users through VR sites. However, too much guidance may stop users returning to VRCommerce because part of the VR experience is, in fact, the experience (Cummins, 2002). If users repeatedly follow the same path to purchase an item, then VR is not needed; a bookmarked 2D Web page with the correct product does the job just as well. There must be a solid rationale for a user to want to experience a product or service rather than simply purchase it. An effective VRCommerce site must allow users to choose how they want to reach their destination. In the VR Mall (Mass & Herzberg, 1999) perhaps when a user logs in they are presented with a list of past purchases and allowed to reorder items without entering the VR space. After the transaction is completed perhaps the user has another option to enter the VR Mall. Beach-

166

town offers another approach to guiding users yet giving them the freedom of choice (Gammick & Hodkinson, 2003). Users can navigate a familiar 2D e-commerce site for the entire transaction. Or they can choose to enter a VR world to stroll the Beachtown boardwalk. The Beachtown layered approach to VRCommerce works well because users can choose the medium and their guidance level. The authors also stress that they make sure users have viewed customized items such as surfboards before allowing users to enter into the transaction process. Moreover, the process takes five to six steps before the system asks for personal and credit information. This layered and multi-level transaction builds a relationship between the business and the user before asking the user to supply personal information (Gammick & Hodkinson, 2003). This ultimately builds trust. Wang and Emurian (2005b) produce a detailed interaction paradigm to measure and promote trust in 2D e-commerce sites. Because 2D Websites are the portal to VRCommerce their guidelines are applicable. Any element that can invoke a trusting atmosphere should be incorporated into the overall design to minimize user uncertainty in the new VR environment. More importantly, the authors argue that overall Internet experience is not significant because of a short user learning curve. By extension, one hopes VRCommerce interfaces will ultimately be accepted among the Web-faring public.

overcoming the Challenges with lavrCs Most of the current studies discussed so far focus on one or two of the challenges. However the business realm is more complex. We must work towards a solution that addresses the challenges of realism and latency; navigation, learning, and acceptance; and guidance and trust if a successful VRCommerce system is to be deployed. To ac-

The Layered Virtual Reality Commerce System (LaVRCS)

complish this we need to create a VRCommerce system that can work on a majority of computer systems, has a familiar interface, and can allow users to choose navigation paths and ask for guidance when needed. To this end we propose building the VRCommerce system within an existing 2D e-commerce Web interface because this is the most familiar to the largest customer segment. Moreover, this system must have a multitude of navigational paths and incorporate increasing levels of VR depending on a particular user’s comfort with the system and with VR in general. The Layered Virtual Reality Commerce System (LaVRCS) will permit users to match the level of technology with their comfort level as they navigate the e-commerce site.

LaVRCS System Architecture In order to reach the largest user segment possible, LaVRCS must use web-based VR. Web-based VR can be created with open web-based standards to enable the majority of client system access. By not depending on a particular software implementation, such as Second Life (Linden Labs, 2009), on the client-side beyond downloading a VR browser plug-in, LaVRCS makes VR accessible to almost all users. For those not able (or willing) to use Webbased VR, the system also employs standard Web protocols and Web browsers, thereby allowing all users—to include mobile users and those on non-standard devices—LaVRCS access and interaction. Of course, there is also the possibility to supplant this interaction level with a well-crafted Webtop or 3D Web application, but care must be taken to enable user access.

Client-Side Requirements LaVRCS client-side requirements must be kept to a minimum in order to allow all users access to the site. A user must have a relatively new computer,

an Internet connection and a current Web browser. Although a more powerful computer will access LaVRCS and smoothly utilize all of its features, systems that can run on the majority of multimediaenabled Websites should also do well. Currently the second and third layers of LaVRCS will only accessible via a desktop or notebook computer because of the X3D plug-in requirements. The first layer of LaVRCS can be accessed via any device (e.g., mobile phone) as it is written with open Web standards: XHTML, XML, CSS, JavaScript, and PHP. The Website scales to the device. However, this first layer offers no immersion capability and should, most likely, incorporate RIA functionality to compensate. To access the second and third layer of LaVRCS, users must have an X3D plug-in player. Users can download open source plug-in players, such as Flux and FreeWRL (Web3D, 2009).

Server-Side Requirements In LaVRCS, the server does most of the storage and processing of data, thereby enabling the majority of clients to access the VRCommerce site. This differs from most RIA and Webtop offerings that place more of a burden on the client system. As discussed in the client-side requirements, some devices (e.g., mobile) cannot currently utilize the VR portions of the Website, but can access all e-commerce functions (browsing, buying, etc.). At a minimum the server must run a Web server (e.g., Apache), a database (e.g., MySQL) for tracking items and user movement, PHP, and Java. As an alternative, the server can be built using a java-based server (e.g., Tomcat) with JSP instead of PHP as the server-side scripting language. The database not only stores e-commerce data, such as products and user information, but also data on each 3D object used within the LaVRCS second and third interaction layers. The interactive 3D objects are designed in Google SketchUp (Google, 2009) and then ported

167

The Layered Virtual Reality Commerce System (LaVRCS)

to X3D format. VR interaction components are implemented using the X3D API (Web3D, 2009). As noted in the client-side section, users must download and install a X3D plug-in. Instructions

Figure 1. LaVRCS system architecture

168

on the LaVRCS site will be provided to help users install the plug-in. LaVRCS will eventually include an automatic process for plug-in download and installation (Figure 1).

The Layered Virtual Reality Commerce System (LaVRCS)

LaVRCS will offer customization options for each registered user, such as favorite products, interaction level, etc. All of this information is stored on the server, with the option of a cookie written to the client’s browser to maintain browser variables on subsequent returns. As with most e-commerce sites, clients will need to log in to place an order via secure transactions.

LaVRCS Interaction Layers The success of LaVRCS hinges on its ability to enable user movement throughout chosen layers of the VRCommerce site. These layers permit users to interact with technology at their own comfort level. Inexperienced users can stay at the first layer for as long as they desire. As familiarity and trust build with the VRCommerce site, users may want to experience embedded VR with certain products. Once a user is comfortable with manipulating a single 3D object, she may want to move to the third layer with an immersive VR experience akin to those found in entry and basic level VR systems in order to experience the product within a given VR environment.

first interaction layer In the first interaction layer, LaVRCS will mimic a traditional 2D e-commerce site. It uses a standard Web interface with expected navigational cues. All products and services are available for viewing and purchase at this level. 2D images

accompany product descriptions along with user reviews and recommendations similar to Amazon. com. There is a possibility for future small-scale RIA implementation components. The first interaction layer is designed for users with minimal e-commerce experience and VR comfort. For example, if a user wanted to purchase a new reclining chair, he would search the e-catalog for one with the desired characteristics, read reviews, and then purchase the item. Some users will never leave this interaction layer, some may use it for certain packaged items, and others may forsake it immediately for the second or third layer. The e-commerce site incorporates standard encryption for SSL layers, a valid certificate, and other assurances to promote user trust and insure site credibility (Wang & Emurian, 2005a). Ultimately, this first interaction layer should already be in place using existing research on effective Web design principles. Although no VR interaction is present, this layer is critical for e-commerce success.

second interaction layer In the second interaction layer, users will have the option of manipulating embedded VR objects within the Web interface. Using our reclining chair example, after selecting the chair in the e-catalog, a user could click on a hyperlink marked “Try the Chair.” A small JavaScript window with an embedded VR player would show a 3D image of

Figure 2. 3D VR Chair (Logue, 2009)

169

The Layered Virtual Reality Commerce System (LaVRCS)

the chair (Figure 2). The user could then use his mouse to rotate the chair, recline it, and use its features (e.g., arm rest storage compartment). This 3D image is stored in the e-catalog database, but it is only accessed when the user clicks the link. With the second interaction layer, LaVRCS moves into the realm of the 3D Web application. A registered user will have the option of automatically requesting second layer interaction on the VRCommerce site without using the first interaction layer. If this option is selected in the user profile, all item descriptions will be delivered with text reviews and have an accompanying 3D object embedded in the same Web page. Registered users will be able to turn off this option at any time. This embedded VR object allows regular users to become familiar with manipulating 3D objects via a keyboard and mouse. Users will most likely adopt the second interaction layer for lower-cost, non-packaged items such as appliances, as well as higher cost non-packaged items such as computers and peripherals. Any item purchases that can best be determined in terms of printed specifications, such as speed settings on a blender or RAM and hard drive size, can be effectively supplemented with trying the product in VR. This allows users to experience its features, sounds, and functions without situating it in an entire VR world context.

third interaction layer Once users are familiar with 3D objects, they can choose to move into the third interaction layer. This layer is a semi-immersive (e.g., no HMD) VR WoW world similar to Mass and Herzberg’s (1999) VR Mall. However, LaVRCS will offer an alternative to the complete VR Mall experience. Select products and services will be linked to customized VR worlds in which users can interact with the products in a real world setting, much like Beachtown (Gammick & Hodkinson, 2003). With this interaction layer, LaVRCS moves be-

170

yond RIAs, Webtops, and the 3D Web. Its closest comparison are the complete virtual realms such as Second Life (Linden Labs, 2009). However, unlike Second Life, LaVRCS focuses specifically on the e-commerce environment without the increased socialization aspects or installation of a separate software application. Using the reclining chair example, after finding the chair and viewing it in the 2D e-catalog or as a 3D VR object, the user can situate it in a VR representation of a particular room to see how it functions within the environment. A user could choose a living room, family room, den, etc. and situate the 3D chair object within the room in order to determine its best fit. All other objects, such as couches or end tables, within these rooms can also be moved. In effect, a user can completely re-arrange a room multiple times to determine if the chair fits the room and her overall expectations. LaVRCS could also allow a user to set room size and select generic objects to place in the room to best simulate her actual intended room. A potential feature will enable a user to submit a photo of his room along with dimensions and have it transformed into a 3D model. The user could then move items around the room and virtually situate the chair to determine its best fit within this personalized space. As with the generic room, all objects can be re-arranged multiple times until the user determines the best room arrangement. All customized 3D objects would be stored in the database for later use by the individual. The added feature would require an account to access. Moreover, with the ability to arrange a room in one’s household without the backbreaking labor, customers may simply choose to create virtual representations of many rooms in their home to try out various furniture and decorating combinations. As an added revenue stream, businesses could choose to charge a fee for such a feature. Ultimately, LaVRCS could offer this feature to other e-commerce sites as well.

The Layered Virtual Reality Commerce System (LaVRCS)

fUtUre trends Perhaps at some point in the future, interaction layers in LaVRCS will no longer be needed. As broadband speeds increase and become available to more users, latency issues may become moot. With increased computing power, comes a decreased need to layer and scale 3D objects as researchers such as Chim, et al. (2003) have examined because processors can quickly render images and process VR user commands. Finally, as more users become familiar with multiple computing interfaces, VR may also become commonplace. As technology advances, cost of ownership also drops and more can afford a computer system in their home. Following Moore’s Law, we see dramatic increases in computer power with substantial price decreases. Coupled with affordable, high-speed technology to minimize latency and realism issues, more computer users are becoming familiar with navigating not only 2D environments but also 3D. The influx of RIAs, Webtops, and 3D Web applications will only hasten this acceptance process. Multiple generations are moving to the Web as well. Senior citizens routinely learn to use the Web, video, and other communication devices; youth begin navigating various virtual worlds at an early age with video games. Increasing virtual world adoption, as well as familiarity with navigation and learning in a virtual environment, is reflected in the growth of video game sales with over 12.5 billion dollars spent in 2006 (Video Game Sales, 2007).

growth of virtual worlds As more people learn to navigate and accept these virtual worlds as part of their daily lives, they will begin to trust using them more for other uses. Although still primarily in 2D use, services such as Google Maps are routinely implemented in Websites for directions, as well as used for urban

planning and research (DiSalvo & Vertesi, 2007). With the addition of Google Streetview, more users are transforming a dot on a map to a rudimentary stereoscopic representation of a neighborhood (Zheng & Wang, 2005). Of course, nowhere is trust in Web 2.0 technologies more prevalent than in users sharing their lives, thoughts, and artistic endeavors on sites such as Facebook, MySpace, YouTube, and other social computing Websites. However, virtual worlds extend beyond Web 2.0 and there is still trepidation from users who might share family videos in YouTube, scan want ads on Craigslist, or search for their next home on Zillow, but this is rapidly changing. With over 11.5 million players in MMORPGs such as World of Warcraft (Blizzard Entertainment, 2009), and over two million inhabitants in virtual worlds, such as Second Life (Terdiman, 2007), virtual worlds are quickly becoming an extension of users’ lives. Businesses recognize the importance of virtual worlds with an increasing number of companies setting up offices in Second Life (Brandon, 2007) in order to reach potential customers with demonstrations, technical support, and other services. Dell, CNN, and IBM have set up virtual headquarters in Second Life. Advertising firms, such as Massive, work with companies to place their ads within video games. Market researchers predict companies will spend $800 million in virtual world advertising in 2009 and over $1 billion in 2010 (Digital Media Net, 2005).

fourth interaction layer Just as users have become more comfortable using Web 2.0 applications in their daily lives, they will soon become accustomed to Web 3.0 applications. VRCommerce will play an important role in this evolution. Adding a fourth interaction layer to LaVRCS is part of the framework’s organic growth. With the popularity of virtual worlds, such as Second Life, users will become comfortable with 3D space interaction and navigation.

171

The Layered Virtual Reality Commerce System (LaVRCS)

The LaVRCS fourth interaction layer will allow users to navigate completely within a VR world to experience products and services before making a purchase decision. Based on the Xj3D development browser (Web3D, 2009), the LaVRCS browser will be delivered and work within a Web-based environment. Users will only need a Web browser, a Java Runtime Environment (JRE), and the LaVRCS browser download.

ConClUsion In this paper I have discussed the general concept of Virtual Reality and have examined the challenges Virtual Reality researchers and developers face as they create VR systems and applications, as well as e-commerce sites themselves. Latency, graphical realism, user acceptance, and the navigational issues of VR are coupled with e-commerce challenges of trust and navigation. In order to develop a successful VRCommerce site that overcomes these issues and challenges, I propose a framework for the Layered Virtual Reality Commerce System (LaVRCS) that offers an enabling approach to the greatest number of users. LaVRCS will enable and e-commerce site to implement VRCommerce in a non-threatening manner. It will allow users to explore and sample as much or as little of the VR experience as they desire. Advanced users will return to not only purchase but also to try new products in customized VR worlds. LaVRCS will also allow designers to address the four challenges of VR. Realism and latency issues are addressed. Layers one and two accommodate beginning users with less powerful computers. Advanced users with more powerful systems can access all layers with no latency or realism challenges. LaVRCS will help users learn how to navigate and use VR with its layered approach. Users can become more comfortable with the 3D objects before exploring a complete VR world. Guidance

172

will be offered with instructions at all levels. Eventually, tour guide avatars could be added for users who want them. LaVRCS will work because ultimately it allows users to choose when they want to use VR for e-commerce. Some may never be comfortable using the third layer, but will still return to purchase products and services at the first or second layer. LaVRCS, much like all Web X.0 technology, is ever changing and evolving. A proposed fourth interaction layer that incorporates components of the second and third layers will allow users to choose their immersion layer according to contextual needs.

referenCes Arns, L., Brisbin, M., Foldes, D., & Holland, J. D. (2006, July 30-August 3). Virtual reality for ecosystem dynamics education. ACM SIGGRAPH 2006 Educators Program, SIGGRAPH ‘06 (p. 36), Boston, MA. Babaioff, M., Kleinberg, R., & Papadimitriou, C. H. (2007, June 11-15). Congestion games with malicious players. Proceedings of the 8th ACM Conference on Electronic Commerce, EC ‘07 (pp. 103-112), San Diego, CA. Barnes, M. (1996). Virtual reality and simulation. Proceedings of the Conference on Winter Simulation (pp. 101-110). Blizzard Entertainment. (2008). World of Warcraft Subscriber Base reaches 11.5 million subscribers Worldwide. Retrieved on November 7, 2009, from http://www.blizzard.com/us/press/081121.html Brandon, J. (2007, May 2). The top eight corporate sites in Second Life. Computerworld. Retrieved on March 7, 2008, from http://www.computerworld. com/action/article.do?command=viewArticleBa sic&articleId=9018238

The Layered Virtual Reality Commerce System (LaVRCS)

Chen, K., & Rea, A. (2004). Protecting personal information online: A survey of user privacy concerns and control techniques. Journal of Computer Information Systems, 44(4), 85–92. Chim, J., Lau, R., Leong, H., & Si, A. (2003). CyberWalk: A Web-based distributed virtual walkthrough environment. IEEE Transactions on Multimedia, 5(4), 503–515. Chittaro, L., & Ranon, R. (2002). Dynamic generation of personalized VRML content: A general approach and its application to 3D e-commerce. Proceedings of Web3D 2002 (pp. 145-154), Tempe, AZ. Chittaro, L., Ranon, R., & Leronutti, L. (2003). Guiding visitors of Web3D worlds through automatically generated tours. Proceedings of the Eighth International Conference on 3D Web Technology (pp. 27-38), Saint Malo, France. Cruz-Neira, C. (2009). VR juggler: Open source reality tools [computer software]. Iowa State University Virtual Reality Applications Center. Retrieved from http://www.vrjuggler.org/ Cummins, N. (2002). Integrating e-commerce and games. Personal and Ubiquitous Computing, 6(56), 362–370. doi:10.1007/s007790200041 Digital Media Net. (2005, April 18). Advertising in video games facts and figures released. Retrieved on March 29, 2008, from http://gamedeveloper.digitalmedianet.com/articles/viewarticle.jsp?id=31947 DiSalvo, C., & Vertesi, J. (2007, April 28). Imaging the city: Exploring the practices and technologies of representing the urban environment in HCI. CHI ‘07 Extended Abstracts on Human Factors in Computing Systems, [San Jose, CA.]. CHI, 07, 2829–2832. Driver, E., & Rogowski, R. (2007). RIAs bring people-centered design to information workplaces. Forrester Research. Retrieved on September 14, 2008, from http://www.forrester.com/Research/ Document/Excerpt/0,7211,43452,00.html

Eastlick, M., Lotz, S., & Warrington, P. (2006). Understanding online B-to-C relationships: An integrated model of privacy concerns, trust, and commitment. Journal of Business Research, 59(8), 877–886. doi:10.1016/j.jbusres.2006.02.006 Gammick, J., & Hodkinson, C. (2003). Virtual reality, involvement, and the consumer interface. Journal of End User Computing, 15(4), 78–96. Google. (2009). SketchUp [computer software]. Retrieved from http://www.sketchup.com/ Hoffman, D. L., Novak, T. P., & Peralta, M. (1999). Building consumer trust online. Communications of the ACM, 42(4), 80–85. doi:10.1145/299157.299175 Jay, C., Glencross, M., & Hubbold, R. (2007). Modeling the effects of delayed haptic and visual feedback in a collaborative virtual environment. ACM Transactions on Computer-Human Interaction, 14(2), 8. doi:10.1145/1275511.1275514 Kirner, T. G., & Kirner, C. (2005, April 4-6). Simulation of real-time systems: An object-oriented approach supported by a virtual reality-based tool. Proceedings of the 38th Annual Symposium on Simulation, Annual Simulation Symposium (pp. 188-195). IEEE Computer Society. Knight, M. M., & Arns, L. L. (2006, July 28-29). The relationship among age and other factors on incidence of cybersickness in immersive environment users. Proceedings of the 3rd Symposium on Applied Perception in Graphics and Visualization, APGV ‘06 (Vol. 153, pp. 162-162), Boston. Lazslo Systems. (2009). Retrieved from http:// www.laszlosystems.com/ Lees, M., Logan, B., & Theodoropoulos, G. (2007). Distributed simulation of agent-based systems with HLA. ACM Transactions on Modeling and Computer Simulation, 17(3), 11. doi:10.1145/1243991.1243992

173

The Layered Virtual Reality Commerce System (LaVRCS)

Linden Labs. (2009). Second Life [computer software]. Retrieved from http://secondlife.com/ Lipman, R., & Reed, K. (2000). Using VRML in construction industry applications. Proceedings of the Web3D-VRML 2000 Fifth Symposium on Virtual Reality Modeling Language (pp. 119-124). Logue, P. (2009). 3D VR Chairs [computer software]. Losh, E. (2006, December 4-6). The palace of memory: Virtual tourism and tours of duty in tactical Iraqi and virtual Iraq. Proceedings of the 2006 International Conference on Game Research and Development, ACM International Conference Proceeding Series (Vol. 223, pp. 77-86), Murdoch University, Perth, Australia. Loten, A. (2007, November 19). E-commerce growth continues. Retrieved on March 28, 2008, from http://www.inc.com/news/articles/200711/ ecommerce.html Mass, Y., & Herzberg, A. (1999). VRCommerce-electronic commerce in virtual reality. Proceedings of the First ACM Conference on Electronic Commerce (pp. 103-109). Pai, D. K., van den Doel, K., James, D. L., Lang, J., Lloyd, J. E., Richmond, J. L., & Yau, S. H. (2005, July 31-August 4). Scanning physical interaction behavior of 3D objects. ACM SIGGRAPH 2005 Courses. In J. Fujii (Ed.), SIGGRAPH ‘05 (p. 154). Los Angeles, CA. Pasman, W., Woodward, C., Hakkarainen, M., Honkamaa, P., & Hyväkkä, J. (2004, June 1618). Augmented reality with large 3D models on a PDA: Implementation, performance, and use experiences. Proceedings of the 2004 ACM SIGGRAPH International Conference on Virtual Reality Continuum and Its Applications in Industry, VRCAI ‘04 (pp. 344-351), Singapore.

174

Phelps, J., Nowak, G., & Ferrell, E. (2000). Privacy concerns and consumer willingness to provide personal information. Journal of Public Policy & Marketing, 19(1), 27–41. doi:10.1509/ jppm.19.1.27.16941 Stansfield, S. (2005). An introductory VR course for undergraduates incorporating foundation, experience, and capstone. SIGCSE Bulletin, 37(1), 197–200. doi:10.1145/1047124.1047417 Strickland, D., Hodges, L., North, M., & Weghorst, S. (1997). Overcoming phobias by virtual exposure. Communications of the ACM, 40(8), 34–39. doi:10.1145/257874.257881 Tecchia, F. (2006, November 1-3). Building a complete virtual reality application. Proceedings of the ACM Symposium on Virtual Reality Software and Technology, VRST ‘06 (pp. 383-383), Limassol, Cyprus. Terdiman, D. (2007, January 3). Counting the real “Second Life” population. Retrieved on February 29, 2008, from http://www.news.com/21001043_3-6146943.html van Duyne, D., Landay, J., & Hong, J. (2007). The design of sites: Patterns for creating winning Web sites, 2nd edition. Upper Saddle River, NJ: Prentice Hall. Varlamis, I., Vazirgiannis, M., Lazaridis, I., Papageorgiou, M., & Panayiotopoulos, T. (2004). Distributed virtual reality authoring interfaces for the WWW: The VR-shop case. Multimedia Tools and Applications, 22(1), 5–30. doi:10.1023/ B:MTAP.0000008657.07799.b0 Video Game Sales Post a Record. (2007, January 15). Associated Press. Retrieved on March 3, 2008, from http://www.msnbc.msn.com/id/16597649/ Wang, D., & Emurian, H. (2005a). An overview of online trust: Concepts, elements, and implications. Computers in Human Behavior, 21(1), 105–125. doi:10.1016/j.chb.2003.11.008

The Layered Virtual Reality Commerce System (LaVRCS)

Wang, D., & Emurian, H. (2005b). Trust in ecommerce: Consideration of interface design factors. Journal of Electronic Commerce in Organizations, 3(4), 42–60. Web3D Consortium. (2009). Retrieved from http://web3d.org/ Whitman, L. E., Jorgensen, M., Hathiyari, K., & Malzahn, D. (2004, December 5-8). Virtual reality: Its usefulness for ergonomic analysis. Proceedings of the 36th Conference on Winter Simulation, Winter Simulation Conference (pp. 1740-1745), Washington, D.C. Wi l s o n , J . ( 1 9 9 7 ) . Vi r t u a l e n v i r o n ments and ergonomics: Needs and opportunities. Ergonomics, 40(10), 1057–1077. doi:10.1080/001401397187603 Zheng, J., Yuan, X., & Chee, Y. S. (2005, July 2529). Designing multiparty interaction support in Elva, an embodied tour guide. Proceedings of the Fourth International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS ‘05 (pp. 929-936), The Netherlands. Zheng, J. Y., & Wang, X. (2005, November 6-11). Pervasive views: Area exploration and guidance using extended image media. Proceedings of the 13th Annual ACM international Conference on Multimedia, MULTIMEDIA ‘05 (pp. 986-995), Singapore.

AdditionAl reAding Allison, D., Wills, B., Hodges, F., & Wineman, J. (1996). Interacting with Virtual Gorillas: Investigating the Educational Use of Virtual Reality. The Art and Interdisciplinary Programs of SIGGRAPH ‘96 on SIGGRAPH ‘96 Visual Proceedings, 147.

Arjomandy, S., & Smedley, T. J. (2004). Visual Specification Of Behaviours In VRML Worlds. Proceedings of the Ninth international Conference on 3D Web Technology (Monterey, California, April 5-8), Web3D ‘04, 127-133. Blanchard, A. L., & Markus, M. L. (2004). The Experienced “Sense” of a Virtual Community: Characteristics And Processes. SIGMIS Database, 35(1), 64–79. doi:10.1145/968464.968470 Greenhalgh, C., & Benford, S. (1995). MASSIVE A Collaborative Virtual Environment For Teleconferencing. ACM Transactions on Computer-Human Interaction, 2(3), 239–261. doi:10.1145/210079.210088 Hetherington, R., & Scott, J. (2004). Adding a Fourth Dimension to Three Dimensional Virtual Spaces. Proceedings of the Ninth International Conference on 3D Web Technology, 163-172. Monterey, CA. Hillis, K. (1999). Digital Sensations: Space, Identity, and Embodiment in Virtual Reality. Minneapolis: University of Minnesota Press. Hutchison, A. (2007). Back to the Holodeck: New Life for Virtual Reality? Proceedings of the 2nd International Conference on Digital interactive Media in Entertainment and Arts (Perth, Australia, September 19 - 21), DIMEA ‘07, vol. 274, 98-104. Macredie, R., Taylor, S., Yu, X., & Keeble, R. (1996). Virtual Reality and Simulation: An Overview. Proceedings of the 28th Annual Conference on Winter Simulation, 669-674. Coronado, CA. Middleton, V., McIntyre, R., & O’Keefe, J. (1993). Virtual Reality and Analytical Simulation of the Soldier. Proceedings Of The Conference On Winter Simulation, 1048-1052. Neil, M. J. (1996). Architectural Virtual Reality Applications. SIGGRAPH Comput. Graph., 30(4), 53–54. doi:10.1145/240806.240816

175

The Layered Virtual Reality Commerce System (LaVRCS)

Pesce, M. (2000). The Playful World: How Technology is Transforming Our Imagination. New York: Ballantine Books, 2000. Ramesh, R., & Andrews, D. (1999). Distributed Mission Training: Teams, Virtual Reality,And RealTime Networking. Communications of the ACM, 42(9), 64–67. doi:10.1145/315762.315775 Rheingold, H. (1991). Virtual Reality. Summit Books. New York: Simon and Schuster. Rössler, O. (1998). Endophysics: The World as an Interface. Singapore: World Scientific. Sun, H., Hujun, B., Tong, N. M., & Wu, L. F. (1999). Interactive Task Planning In Virtual Assembly. Proceedings Of The ACM Symposium On Virtual Reality Software And Technology, 174-175. Sutcliffe, A., Gault, B., Fernando, T., & Tan, K. (2006). Investigating Interaction in CAVE Virtual Environments. ACM Transactions on Computer-Human Interaction, 13(2), 235–267. doi:10.1145/1165734.1165738 Thalmann, N., & Thalmann, D. (Eds.). (1994). Artificial Life and Virtual Reality. New York: John Wiley and Sons. Tsvetovatyy, M., Gini, M., Mobasher, B., & Ski, Z. W. (1997). MAGMA: An Agent Based Virtual Market for Electronic Commerce. Applied Artificial Intelligence, 11(6), 501–523. doi:10.1080/088395197118046 Walczak, K., & Cellary, W. (2003). X-VRML for Advanced Virtual Reality Applications. Computer, 36(3), 89–92. doi:10.1109/MC.2003.1185226 Wexelblat, A. (1993). The Reality of Cooperation: Virtual Reality and CSCW. In A. Wexelblat (Ed.), Virtual Reality: Applications and Explorations. Boston, MA: Academic Press Professional, 23-44.

176

Key terMs And definitions 3D Web: This technology delivers threedimensional objects and environments embedded within Web browsers. Using technology specifications, such as those put forth by the Web 3D working group, users can manipulate objects using Web browser plug-in technology. Web-based Virtual Reality is an extension of the 3D Web. CAVE: The recursive acronym stands for Cave Automatic Virtual Environment. A CAVE is a room-sized cube where images are projected on three or all four of the surrounding walls. Images change and move according to the user’s actions within the CAVE. HMD: Head Mounted Display. By wearing a HMD, virtual reality users are presented with 3D stereoscopic images that mimic physical world perception. HMDs are one of the more common VR components used. Rich Internet Application (RIA): RIAs provide limited desktop application functionality within a Web browser. Most RIAs require an additional technology beyond standard XHTML (e.g., Flash). Virtual Reality: A complete simulation of reality. Virtual Reality (VR) can be an exact replica of the real world, a reality that is very different from that which is considered real, or perhaps an intense simulation or a situation (e.g., high cliff used to treat acrophobia). Virtual World: A complete representation of a physical realm. Most likely this world is populated by avatars representing players, as well as virtual representations of virtual world characters (bots). Virtual worlds can mimic environments we are familiar with, or populate worlds with completely different inhabitants and rules of nature (e.g. people can fly). VRCommerce: E-commerce that uses virtual reality environments or 3D objects that allow customers to experience products and services before purchasing them. Popular examples include experiencing travel destinations before purchas-

The Layered Virtual Reality Commerce System (LaVRCS)

ing vacations and test-driving cars before visiting showrooms. VRML: Virtual Reality Markup Language. A set of standards that governs a XML file markup language developed to display 3D interactive vector graphics and virtual environments on the Web. Webtop: A Web desktop. Webtops are RIAs that offer a complete desktop experience delivered within a Web browser. One of the more popular

Webtops is the Google application suite that provides word processing, spreadsheets, and e-mail. Webtops are still rudimentary in their offerings but intend to offer a complete desktop environment. X3D: A successor to VRML. The X3D file specification allows for additional 3D object extensions (e.g., CAD). It also allows for the integration of additional programming languages, such as Java for more expansive virtual environments delivered over the Web for more robust applications.

177

178

Chapter 10

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era Ming-Chien (Mindy) Wu University of Western Sydney, Australia Bhuvan Unhelkar University of Western Sydney & MethodScience.com, Australia

AbstrACt This chapter describes an approach to extending service oriented architecture (SOA) with mobile technologies (MT) resulting in what can be called mobile service oriented architecture (MSOA). Web services (WS) is a popular approach to business applications in the second Web generation (Web 2.0). Mobile technologies (MT) help people reach out and interact with each other anytime and anywhere, transcending time and location boundaries. MSOA brings together MT and WS to create opportunities for offering and consuming services over the wireless networks in Web 2.0 era and beyond. Furthermore, the intelligent convergence of mobile connectivity, network computing, open technology, open identity, and several such emerging technologies pave the way for newer and wider range of service-oriented business opportunities. The authors describe this MSOA model and an approach to its validation through an implementation framework in this chapter.

introdUCtion Mobile Service-Oriented Architecture (MSOA) aims to apply the concept of Web Services (WS) to the rapidly emerging Web 2.0, 3.0 and beyond. Information and communication technologies (ICT), especially the Internet, are developing at a breathtaking speed. The Web, as known in the past, DOI: 10.4018/978-1-60566-384-5.ch010

was a means of communicating messages. This mechanism of communication has now evolved into a means of business collaboration through software applications (Unhelkar, et.al, 2009), resulting in what can be understood as Web 2.0. Web 3.0, however, takes the ability of the web to execute applications even further; it deals with intelligent convergence of connectivity, network computing, open technology, open identity, and several such key emerging technologies.

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

We believe that these two key technologies, when converged, create tremendous opportunities for businesses to offer and consume services independent of location and time across a wide range of networks. This is so because: •



Mobile technologies (MT) include wireless networks, handheld devices and mechanisms to store and present contents to the users in a personalized manner, and Web Services (WS) enable services to be offered across the web by ‘wrapping’ them with commonly understood and standardized interfaces. WS focus on using information, processes and resources that result in an organization’s ability to provide services across the Internet.

Together, the aforementioned two technologies enable businesses to execute complete business transactions (as against mere exchange of data and information through emails). This ability of remote execution of applications opens up opportunities for businesses to collaborate – resulting in them being a part of Web 2.0 and beyond. The Service Oriented Architecture (SOA), when extended with mobility, is of interest to the enterprise architects as well as the business leaders as there are many opportunities resulting from this combination of technologies that did not exist before. MSOA provides the ability for convergence of land-based and mobile connectivity that utilizes the web beyond just a mean of communication. Extending SOA with mobility, as is argued in this chapter, should equip the modern business with ability to incorporate location and time independence in its service offerings. This ‘mobility’ will enable the business to create effective and personalized internal and external mobile business processes. This chapter starts by outlining the research methodology. This is followed by a discussion on the various generations of the web, web services and the SOA. The research project is then divided

into two parts: 1. The model of Mobile Service Oriented Architecture (MSOA) with web service. This initial modeling of MSOA is based on the literature review and research discussions. 2. The implementation framework for enabling such extension and incorporation of mobility in SOA. This implementation framework is based on the case studies by interviewing experienced enterprise architects from the industries. However, the actual implementation of the framework is out of scope for this chapter. This chapter finally concludes with a summary of the MSOA and points out to the future direction of this research project.

reseArCh Methodology The selected methodology for this research is the qualitative research method. This qualitative approach is used to construct the initial model of MSOA and it is made up of literature review, case studies based on interviews and action research studies. The literature review is used to outline the various generations of the Web, understand the meaning of web services and also understand mobility. After the literature review is completed, the initial MSOA model is constructed primarily out of the ensuing research discussions and the initial experimentation. The case studies resulting from the interviews are able to verify the initial MSOA model and also help in creating a complete MSOA implementation framework. Three action research studies are planned for this research project and they will be conducted at the premises of the participated organization to study their MSOA implementations and thereby validating the results to the initial MSOA. First action research implementation methodology is used as an example of this chapter. Eventually, it is hoped, that the resultant MSOA model will be usable across any organization with reduced risks during its implementation.

179

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

web generAtions The Web has been continuously developing and changing. Murugesan (2007a) has discussed the Web’s evolution- past, current and anticipated – which also fits in with the common understanding of the various generations of the Web, referred to as Web 1.0, Web 2.0, Web 3.0 and Web 4.0. The first generation of the Web - Web 1.0 generation - is about connecting information (Info-Centric). The focus of Web 1.0 was primarily on building the web, making it is accessible, and commercializing it for the first time. This generation of Web provides opportunities for people to communicate with other people and businesses using their personal computing devices. The users of this first generation of the Web go through the web sites of businesses to find the information they want and use the email to communicate with people they know. However, the machine and its ability to connect remains at the core of this web generation. The second generation - Web 2.0 generation - is about connecting people (People-Centric) together, participation, interaction and collaboration. The Web in this generation provides the application platform that complements the personal computer used for communication in the earlier generation. The support of an application platform and the ability to create, store and execute applications, has resulted in a technical as well as a social revolution in the use of web (Murugesan, 2007b). The blogs, wikis, social networks, and RSS feeds, as well as the continued growth Web 2.0 applications have been popular and used in past few years. Web Services (WS), referred to earlier, are a successful application of web technology in the generation of Web 2.0. Users can not only get information from the business but also get services and pay for them across the net. Furthermore, people to people interaction moves from communication using emails to creation of social networks and groups resulting in sharing of information, thoughts and knowledge across

180

the web. The third generation - Web 3.0 generation - is about connecting knowledge and applications (Machine-Centric). In this Web 3.0 generation, which we are yet to fully realize, the web is considered as a universal computing grid replacing operating system and hard drive, resulting in large and dynamic groups of machines connected to each other. Web 3.0, a phrase coined by John Markoff (2006), can be considered as a phrase referring to Internet-based technologies and services that emphasize a machine-facilitated understanding of information on the web that would result in a more productive and intuitive user experience. This Web 3.0 is a convergence of several new technologies and service, such as mobility connectivity, network computing, services business, open technologies and open identity. In this generation, people will use the Internet in a more relevant, useful, and enjoyable way; the individuals can reach the web anytime, anywhere to obtain and validate the knowledge they want through their own personalized mobile devices. Finally, the Web 4.0 generation, of which we can only creatively imagine, is about connecting the power of human and machine intelligence in a ubiquitous environment. Web 4.0, known as “Intelligent Web” or “Smart Web”, enables the software agent(s) to reason and communicate with other agents and systems and work collaboratively to accomplish things on user’s behalf (Agent-Centric). Users in this generation will not be categorized only as humans, but they will be intertwined wirelessly with machines in order to interact, reason, and assist each other in ever evolving smart ways.

ws, eA And soA The various generations of Web comprise a suite of technologies which have influenced, and are likely to influence, the way in which enterprises operate and the way their architecture is created.

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

This section looks at the technologies of Web Services (WS), Enterprise Architecture (EA) and Service-Oriented Architecture (SOA). These WS enabled technologies provide collaboration and integration of applications on the Internet. Therefore, we consider these technologies to be era of Web 2.0 generation. Web services are structured within a Service Oriented Architecture which, in turn, provides basis for most modern-day Enterprise Architecture. Marks and Werrel (2003) define Web Service (WS) as “loosely coupled, self-describing services that are accessed programmatically across a distributed network, and exchange data using vendor, platform, and language-neutral protocols.” Web services are configured and deployed across corporate intranets or the Internet using Web Services Description Language (WSDL). WS are self-contained and they describe their offerings in a standardized manner using eXtensible Markup Langauge (XML) so that they can be published, located and invoked across the Internet. EA represents a technology-business philosophy that provides the basis for cooperation between various systems of the organization that may be inside or outside the organizational boundary. SOA also facilitates ability to share data and information with business partners by enabling their applications to ‘service’ with each other. Thus, WS supports enterprise applications, services, related IT service providers as well as deployment of services, applications and processes (Wiehler, 2004). Creating and managing the web service based architecture should result in an infrastructure that would enable enterprises to take their fine-grained services and other information data repositories and compose them into real-time business management information system that make up a comprehensive EA. An enterprise architecture that links together the applications and web services within organization, across enterprise, and across the Internet can be called the Service Oriented Architecture (SOA). The W3C (2004) defines SOA as “A form

of distributed systems architecture. This architecture consists of a set of components which can be invoked, and whose interface descriptions can be published and discovered”. Thus, SOA is an IT architectural approach that increases business agility by aligning IT technologies and services with business goals. SOA enables organizations to establish an environment that uses loosely coupled services to support the requirements of today’s highly competitive businesses. Additionally, in order to increase the ability of the enterprise to serve its customers as well as deal with its business partners in today’s dynamic business environment, there is a need to integrate these IT products and services through a common SOA. A carefully thought out and implemented SOA provides the enterprise with competitive advantage by opening up opportunities to streamline processes, reduce costs, increase customer satisfaction and enable thorough strategic planning (Lan and Unhelkar, 2005). The objective of a successful SOA is to provide real-time responses in both internal business processes and external supplier and customer relationships. Thus, with SOA, business processes could be configured to automatically launch communications with relevant players across the enterprise. Popkin (2007) suggested that SOA is emerging as the popular commercial industry solution for improving collaboration across time, place and platforms. Additionally, Dowell (2007) states that the key to manage the SOA solutions is in understanding, defining, and measuring service level achievement to meet strategic outcomes. McGovern et al(2004) state that SOA provides an important new avenue for the integration of applications. Creating the new applications under SOA offers a significant increase in the qualities of availability, interoperability, maintainability, and reliability of those applications. Butler Group (2004) listed the SOA benefits in their Technology Evaluation and Comparison Report that include: Faster assembly of solutions; Reduction in cost and complexity, with consequent lowering of

181

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

maintenance overhead; Lowered cost of ongoing change; Business interactions distanced from technology constraints; Better enterprise flexibility; The ability to maximize existing IT investments; and a more robust IT environment. However, there is the growing requirements for better integration between systems to support business processes agility, and the needs for realtime and location-independence monitoring of business operations. This need of business agility is leading to the development of a flexible SOA which brings about a synergy between systems, processes, and information that can provide that necessary agility. Services in a flexible SOA can be created, modified, and removed dynamically in almost a real-time manner anywhere by the access provided to the users through any devices. These advantages of SOA, however, need to be considered in collaboration with and as extensions of mobility in order to provide greater advantage to businesses in terms of their agility. .

Mobility And soA Our literature review provided us with the necessary impetus to consider extension and modification of SOA in order to incorporate mobility in it. The need for this extension, as argued earlier, is felt because mobile technologies are now popular and effective technologies in business as well as in enterprise architectures (Unhelkar, 2006, Unhelkar, 2009). Increasing understanding and affordability of mobile technologies, coupled with improving network transmission infrastructures, opens up opportunities for new and innovative business applications. Enterprises seek to capitalize on the emerging MT because mobility can overcome “time and location” boundaries to provide enterprises with the ability to operate effectively, in real-time, and respond quickly to the ever increasing changes in a competitive marketplace (Linthicum, 2000). MT brings mobility connectivity to the individuals and also opens up

182

doors to their access to business independent of location and time (Unhelkar, 2005). Thus, with the integration of mobility in Web, individuals are now able to access their personal resources from anywhere using both fixed as well as wireless networks. The convergence of mobility and the web leads to business applications of a new era. Lee et al. (2004) have defined mobility as the capability of being able to move or be moved easily. Mobility pertains to people’s use of portable and functionally powerful mobile devices that offer the ability to perform a set of application functions overcoming “time and location” boundaries, while also being able to connect to, obtain data from, and provide data to other users, applications and information systems. Such mobility solution offers the best portability, usability, functionality, and connect-ability to the users. The decreasing cost and increasing speed of wireless networks, in addition to the decreasing device costs, are driving the move to wireless communication. Mobile technological developments are further enabling convergence of devices and networks that have resulted in popularity of mobile agents. These mobile agents enable mobile devices to act on behalf of the customer. Thus mobile technologies, including their network and devices, are now an important element in an enterprise’s strategies and, therefore, form a crucial part of SOA. Incorporating mobility in SOA can help real-time information access amongst various systems the deal with production planning and control, inbound and outbound logistics, material flows, monitoring functions, and performance measurements (Rolstadas and Andersen, 2000). A comprehensive MSOA provides an excellent opportunity for creation of an ‘agile’ technical platform that would enable delivery of business services to a “location independent” market. According to Ghanbary (2006), by correct application of MT into the business processes, the business enterprises are likely to gain advantages such as increased profits, satisfied customers and greater customer loyalty. These customer-related

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

advantages will accrue only when the organization investigates its customer behavior in the context of the overall mobile environment. Thus, strategic incorporation of mobility requires organizations to adapt not only technically the mobile features, but also at the same time keep the customer firmly in mind. Umar (2005) states that the Next Generation Enterprises (NGEs) will rely on automation, mobility, real-time business activity monitoring, agility, and self-service over widely distributed operations to conduct business. The value of mobility comes to the enterprise when its SOA is made more effective with the help of mobility. The integration of MT in SOA provides the enterprise with the agility to configure numerous services that can then be offered to its mobile customers. For example, with MSOA, the enterprise can provide 24 hours x 7 days, globalized services including handling product enquires and providing technical support to the customers. MSOA also results in an upgrade to the traditional supply chain resulting in a Mobile Supply Chain Management (M-SCM) (Wu et. al., 2007). Similarly, traditional CRM upgrade to Mobile Customer Relationship Management (M-CRM) and trading procurement to Mobile procurement (M-procurement). These extensions and integrations are the critical points of considering the capabilities of MT and their applications to SOA (Wu, 2007) and they are discussed in greater detail next in order to produce a MSOA model.

MsoA Model in web 2.0 This section outlines the initial MSOA model. This MSOA model incorporates a strategic approach to extending SOA with adoption of MT. This extension to SOA is required, especially in the Web 2.0 environment, wherein the Internet transcends its use as merely a means communication. Web 2.0 and beyond has a need to integrate mobility in its applications to enable their execution in a location and time independent manner.

This is so because that the enterprises face the real-time complicated changed market in the web 2.0 and beyond generation, there is the only way to solve this problem: cooperative the mobility benefit. MEA integrate all the information into a single architecture, once the market change, the information will pass to the company manger in a real-time, therefore, the enterprise could provide the real-time response, strategy, and solution to those changes and their customer. The proposed extension model that would integrate execution capabilities of applications over the web together with mobility is as shown in Figure 1. This extension model shown in Figure 1 is based on our literature review, analysis and discussions of the SOA and Mobile Technology. Thus, this construction of the initial MSOA model is the output of our preliminary work in this domain. A model for MSOA (see Figure 1) shows how web services and mobility affect the overall SOA. There are numerous technologies that used for MSOA, such as wireless hub, application connectivity, data format and transformation, integration modules, support for transactions, enterprise portal, and web service (Finkelstein, 2006). The enterprise repository is a comprehensive system containing all applications and the enterprise model. Users can access the business applications using the Internet though its native application programming interface (API) which itself would be based on the eXtensible Markup Language (XML), web forms, and web service. Once we integrate mobility into the SOA, we find that users (people) can use their mobile, Personal Digital Assistants (PDA) and portable computers (notebook) through their service providers using the wireless access protocol (WAP) to connect to the enterprise repositories and access the enterprise systems. The service publishes details of what it provides, the information to send, and what to expect in return in a registry (which may be public or private). These functions use the WSDL standards, and the directory itself follows the

183

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

Figure 1. A model for MSOA – by applying mobility to SOA. © 2007‐10, MethodScience.com Used with permission.

Universal Description, Discovery, and Integration (UDDI – www.oasis.org) standard. The customer finds out details of the service using the same two core standards (WSDL & UDDI), and then calls, or binds, to the service using SOAP. As a result, the reengineering of business processes actually results in what can be called serviceoriented process orchestration. Such serviceoriented model of enterprise system architecture is based on the electronic cooperation between the various enterprise information systems (such as SCM and CRM systems) together with mobile strategy management. In such cooperation, user have mobile devices through the mobile portal to access web service and pass through enterprise bus or middleware to connect the enterprise system architecture in the MSOA environment. The need to consider mobile networks in this evolving MSOA model cannot be over emphasized. The ability to provide excellence in realtime communication between the business and

184

consumer is enhanced through mobile networks. A mobile application that is using the WS to transmit their data is classified as Mobile Web Services (MWS). According to Pashtan (2005), mobile terminals and mobile services are an integral part of the extended web that includes the wireless domains that facilitate automated interoperation between terminal and network services. Mobile networks are mainly grouped in to two categories: short-range and long-range. Short-range technology, such as Bluetooth, is now used into most of mobile devices functions and processes. Bluetooth technology enables easy synchronization between a personal computer (PC) server and one or more other mobile terminals (Buttery, and Sago, 2004). This synchronization has been particularly successful in cooperative applications and providing access to MSOA. Long-range networks include cellular networks and WiMax. Long range networks help in effective integration of enterprise information, application, processes and systems.

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

Furthermore, intra-organizational MSOA application integration can use Wireless Local Area Network (WLAN) technology to provide employees access to the enterprise system anytime, anywhere. Significant progress is achieved as a result of end-to-end secured transactions, “alwayson” applications and location-based services in GPRS networks, flexible broadband services with WLAN environment at high-demand venues such as airports, hotels, university campus, or exhibition centers, ad hoc networks; and finally, rich mobile multimedia devices in Web 2.0 and beyond generation. Voice over IP (VoIP) is yet another technology that rides on the back of IP connectivity and that helps in extending globalization of business with SOA. VoIP overcomes costs, times and integration issues across different geographical and time zones. In addition, Global Positioning System (GPS) devices and Radio Frequency Identification (RFID) tags and readers are already used in SCM systems to improve delivery service and tracking production location (Hurster et. al., 2006). RFID Technology helps highly location-based tracking, reduces the cost and risks, and also improves the efficiency and effectiveness of MSOA. Thus, mobility and WS come together in the MSOA model in order to provide pervasive, simple, and platform-neutral interaction between the users and the business, resulting in greater opportunities for various mobile devices and applications to interact with each other. MSOA brings about not only internal integration but, through its extendibility, also offers greater efficiency to its external suppliers, customers and other trading partners over the mobile network and Internet. As discussed by Ghanbary and Unhelkar (2007), those users could connect to the Collaborative Web Based System (CWBS) of WS or MWS and requests to register in the system. The CWBS prompts the appropriate member to carry out the relationship of multiple organizations which could collaborate with each other not necessarily known to each other. This CWBS of

MSOA would extend enterprise social networks, including not only clients also suppliers, and even if collaborative multiple enterprise work together in Web 2.0, 3.0 and beyond generation. MSOA not only enables the enterprise to present a unified view of the system to their suppliers and clients, but also improves quality by reducing errors by eliminating duplication of data entry. When either the customers or the sales personnel enter data related to a service or product, that data is directly entered through a mobile device in the enterprise system architecture. Thus there is no duplication of data entry that improves efficiency and reduces costs and errors. Extending SOA with mobility is complimented by ensuring that the business processes of an organization are reengineered to cater to the mobile data-entry points. Hoque (2000) states that a good EA needs to take into account the following: agility, interoperability, reusable assets, ownership, scalability, and cycle time. Sharif et al. (2004) declared that two key issues within EA research today are: “evaluation of business models which can influence EA” and “implementation EA within the organization for assessing the impact”. The time and location independence of mobility open up tremendous opportunities for organizations to offer integrated services to their clients and partners, which result to MSOA. Thus, MSOA connect existing and new systems to enable collaborative operation within the entire organization in real-time – providing new and improved services without location and time limitations. However, SOA with mobility has the challenges of security, privacy, computing power and usability. In particular, in the past, security concerns were the main factor inhibiting the widespread deployment of web services. SOA also requires security levels to control that the users who can use which service for what purpose (Nand, 2006). Enterprises now expect the use of MSOA over the mobile Internet to be based on secure foundations as well. Thus, this research continues to focus, amongst other things, on the security aspect of MSOA. In fact, our investiga-

185

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

tions lead us to incorporate CLEW (Closed Loop Environment for Wireless), a mobile technology designed by Alacrity Technology, a Canberra based company, in our model. This technology improves the security of mobile technologies and the proposed research project is aimed at incorporated CLEW-based security in MSOA. Thus, our MSOA model proposed here considers the advantages and challenges of mobile web services that would provide benefits to organizations beyond its boundaries in a collaborative manner.

MsoA iMPleMentAtion frAMeworK As mentioned upfront in our research methodology, at this stage of this research, we have conducted interviews of experienced enterprise architects, business analysts, Chief Information Officer (CIO) and IT executives to provide us with valuable input and help in constructing and fine-tuning our MSOA implementation framework (Wu and Unhelkar, 2008). The experiences of these experts indicates that in order to reach a clear vision of MSOA and to build the many services that support MSOA, enterprises need to understand the human, system, process, and technology aspects of the MSOA. Thus, for example, creating the center of excellence or similar cross-functional group to provide resources and guidance, to serve as a repository for best-practice information, and to operate tools that support the MSOA implementation is the critical factor for success (Swenson, 2007). Figure 2 indicates an MSOA implementation framework. This framework is based numerous interviews conducted by the lead author of this chapter and subsequent analysis of the results. The initial framework has also been described by Curtis and Wu (2009) and in its initial phase, this framework focuses on how teams can be organized for effective MSOA implementation. The framework shown in Figure 2 is divided into

186

two cross-service architect teams: the first team is the business architects team, whose responsibility is to analyze the business systems, process, information, and people structure of MSOA; the second team is mobile technology architects team, whose responsibility is to support mobile technology solution to meet the requirements from business architects team. The primary goal for these two teams is to understand how people work, who owns what responsibilities, and which interdependencies link business processes and technology resources. Based on expert advice and discussions with subsequent interviewees, we discovered a need to divide the team of Business architects in the framework further into 6 groups that would cover internal and hybrid organizations. The team of internal organization has 4 groups which are distribution and marketing groups, financial groups, operations groups, also product and placement groups. Team of hybrid organization has 2 groups, SCM and CRM groups. The goal for business team in this MSOA implementation framework is to discuss and agree on the business elements of an application. The architects of all groups in a business team have to determine their department direction, describe the core business processes, define the department services, declares the applications requests, priorities features, and most importantly to meet the users and services requirements within business strategy and support the department objectives. One interviewee, who was the CIO of his organization, pointed out that Mobile Technology architects team should be divided into 3 groups, which are analysis and planning group, implement and training group, and also support and maintain group. The business architects provide the mobile business requirements within the constraints of the technology and the mobile technology team works out how to implement mobile services. The aim of the mobile technology team is to discuss and agree on how to manage the technological underpinnings and support the business in its effort to become agile.

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

Figure 2. MSOA implementation framework based on the MSOA model outlined. © 2007−10, MethodScience.com Used with permission.

The architects of analysis and planning group firstly analyze the requirements of business from the mobile technology and its infrastructure. The architects investigate the technical complexity of implementing mobile services. Another interviewee, an IT Executive manager, interviewee provided some questions need to be figure out in this stage, such as: •





Should mobile technology within MSOA be integrated with the supporting hardware, software and database within the structure? Is there enough expertise within the enterprise to implement MSOA or should some of the implementation work be outsourced? If outsourced, how this knowledge will be transferred to support and maintain the group?

Answering all these kinds of questions after investigation from the architects, the analysis team

helps understand how the MSOA can be utilized fully by the enterprise. This can then be followed by project planning and creation of task list for implementation. Once the MSOA implementation plan created, the models and corresponding documents are passed to the implementation and training team. The implementation team follows the plan and also provides service and mobility solution to business architect team. There is substantial training needed in MSOA implementations as both technical and business personnel need to change the way they approach their work. For example, business agility becomes a real possibility with MSOA, but the need to configure and validate business processes is paramount. Furthermore, from a user’s viewpoint, there is a need to provide a help desk, desktop support, production support, systems team, computer operators, also support and maintain architect group with appropriate consulting and training. The support and maintenance team needs to be closely supported in a MSOA implementation as it has to take over from where

187

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

Table 1. DMERA migration plan table (Wu and Unhelkar, 2008) Current state “as-is”

Desired Target state “To-be”

Details of Documentation

Description

Key factors

Reserve assessment indicator

Platform and interoperability diagrams from existing enterprise information systems

Analysis

Assessment indicator result

Target MSOA opportunities

Which IT and MT infrastructures want to be existed and extended with current SOA

Design

Construct current SOA

Construct Target MSOA

MSOA diagrams show how new MSOA can be matched target MSOA

Implementation plan

Target MSOA opportunities result

Target MSOA implementation plan

MSOA migration progresses of data, process, system, people implementation

the implementation is completed. The support and maintenance issues in MSOA become even more important if the project is outsourced. This MSOA model is created through interviews by the researcher as a part of a doctoral research study, and it is validated by an action research within a software development company (Company D) in Australia. The purpose of this action research is to apply to this company’s projects repository, in order to create a “Company D Mobile Enterprise Reference Architecture” (DMERA). Such DMERA provide more mobility opportunities to the enterprise in terms of improving the effectiveness and efficiency of architectural work at Company D. The DMERA Transition Road Map shows the clear process migration plan from current different generation information systems to the enterprise target dream architecture- DMERA. This table has been used to DMERA implementation of Company D, which is included “Current” and “Target”, “As-is” and “To-be” to follow the information system lifecycles including description, analysis, design, and implementation plan of current state of enterprise, and desired target state MSOA of enterprise. Following Table 1 is the designed migration plan table of DMERA implementation for this research. Firstly, the research team interviews the company to understand the key factors of people, process, technology, data, and system of current EIS; analysis and document them into the current state description section. From those documents,

188

finding the gap and drawing the diagrams to show the reserve assessment indicator for the target state. After that, setting up the meeting with SOA building team of the company to those diagrams and those assessment indicator results. Moreover, discuss in the meeting to analysis and decide which IT and MT infrastructures could be extended with current SOA as target MSOA opportunities into their company. After the meeting, the research team construct the current EA diagrams, also the target MSOA diagram to shows how the MT could be adapted to new MSOA, and makes the new MSOA be matched the enterprise expected extension. Furthermore, these diagrams have to be modified through many times meeting with SOA building team to have final decision which MT application opportunities they would like to integrate into their MSOA and have a plan schedule to show how the implementation processes and timeline should be achieved. After the research team and SOA building team finalize the implementation plan schedule, the implementation plan tables of people, processes, technology, data and systems have been completed as well. At the last stage, the research team has construct the comprehensive MSOA implementation table to prevent duplicates implementation process between different factors implementation table, also effective the implementation processes and reduce the implementation schedule time.

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

ConClUsion & fUtUre direCtions This chapter outlined the importance of MSOA model as a mean of identifying integration opportunities and providing opportunities to integrate various Mobile applications and technologies within the enterprise. Moreover, this chapter also provided an overview of web generation, web service, EA, SOA, MT, mobility effect SOA and implementation framework of MSOA. We argued that MT needs to be integrated with the overall SOA and the business processes of the enterprise. Such mobile integration would result in a MSOA model which would enable the enterprise to conduct business independent of location and time boundaries in the Web 2.0 generation. Thus, an integrated MSOA is a powerful tool to help manage the enterprise’ operation and better CRM as well.

referenCes W3C: Web Services Architecture. (2004, February 11). W3C working group, note. Retrieved from http://www.w3.org/TR/2004/NOTEwsarch-20040211/ Butler Group. (2004, February). Technology evaluation and comparison report-enterprise architecture: An end-to-end approach for realigning IT with business aims. Butler Group. Buttery, S., & Sago, F. A. (2004). Future application of Bluetooth. In Mobile and wireless communications: Key technologies and future application. British Telecommunications Plc: The IEE. Curtis, D., & Wu, M. (2009). Investigation into the impact of integration of mobile technology applications into enterprise architecture. In B. Unhelkar (Ed.), Handbook of research in mobile business, 2nd edition: Technical, methodological, and social perspectives. Hershey, PA: IGI Global.

Dowell, S. J. (2007). Enterprise architecture within the service-oriented enterprise. In P. Saha (Ed.), Handbook of enterprise systems architecture in practice. Hershey, PA: IGI Global. Finkelstein, C. (2006). Enterprise architecture for integration: Rapid delivery methods and technologies. USA: Artech House. Ghanbary, A. (2006). Evaluation of mobile technologies in the context of their applications, limitations, and transformation. In B. Unhelkar (Ed.), Mobile business: Technological, methodological, and social perspectives. Hershey, PA: Idea Group Publishing. Ghanbary, A., & Unhelkar, B. (2007, May 1923). Technical and logical issues arising from collaboration across multiple organisations. Proceedings of IRMA Conference, IRMA 2007, Vancouver, Canada. Hoque, F. (2000). E-enterprise: Business models, architecture, and components. Cambridge University Press. Hurster, W., Fuychtuller, H., & Fischer, T. (2006). Mobile batch tracking: A breakthrough in supply chain management. In B. Unhelkar (Ed.), Handbook of research in mobile business: Technical, methodological, and social perspectives. Hershey, PA: IGI Global. Irani, Z., Themistocleous, M., & Love, P. E. D. (2003). The impact of enterprise application integration on information system lifecycles. [Elsevier Science]. Information & Management, 41, 177–187. doi:10.1016/S0378-7206(03)00046-6 Krafzig, D., Banke, K., & Slama, D. (2005). Enterprise SOA: Service-oriented architecture best practices. Pearson Education, Inc. Lan, Y., & Unhelkar, B. (2005). Global enterprise transitions: Managing the process. Hershey, PA: IGI Global.

189

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

Lee, V., Schneider, H., & Schell, R. (2004). Mobile applications: Architecture, design, and development. Hewlett-Packard Development Company L.P., publishing by Pearson Education as Prentice Hall Professional Technical Reference. Linthicum, D. S. (2000). Enterprise application integration. Addison-Wesley Information Technology Series. Markoff, J. (2006, November 12). Entrepreneurs see a Web guided by common sense. The New York Times. Marks, E. A., & Werrel, M. J. (2003). Executive’s guide to Web services. Hoboken, NJ: John Wiley & Sons, Inc. McGovern, J., Ambler, S. W., Stevens, M. E., Linn, J., Sharan, V., & Jo, E. K. (2004). A practical guide to enterprise architecture. Pearson Education, Inc. Murugesan, S. (2007a). Get ready to embrace Web 3.0–business intelligence advisory service. Cutter Executive Report, 7(8). Murugesan, S. (2007b). Business uses of Web 2.0: Potential and prospects. Cutter Business-IT Strategies Executive Report, 10(1). Nand, S. (2006). Developing a theory of portable public key infrastructure (PORTABLEPKI) for mobile business security. In B. Unhelkar (Ed.), Handbook of research in mobile business: Technical, methodological, and social perspectives. Hershey, PA: IGI Global. Pashtan, A. (2005). Mobile Web services. UK: Cambridge University Press. Popkin, J. (2007). Leveraging the value proposition of SOA: How enterprise architecture helps organizations analyze and develop their service strategy. Telelogic. Ramakrisham, K. R., Bhattar, R. K., Dasgupta, K. S., & Palsule, V. S. (2006). Review of wireless technologies and generations. In B. Unhelkar (Ed.), Handbook of research in mobile business: Technical, methodological, and social perspectives. Hershey, PA: IGI Global. 190

Rolstadas, A., & Andersen, B. (2000). Enterprise modeling-improving global industrial competitiveness. Kluwer Academic publishers. SAP AG. (2005). Creating an enterprise services architecture road map. SAP Group. Sharif, A. M., Elliman, T., Love, P. E. D., & Badii, A. (2004). Integrating the IS with the enterprise: Key EAI research challenges. The Journal of Enterprise Information Management, 17(2), 64–170. Swenson, K. (2007). The key to SOA governance: Understanding the essence of business. Cutter IT Journal, 20(6), 17–22. Umar, A. (2005). IT infrastructure to enable next generation enterprises. Information Systems Frontiers, 7(3). ISSN:1387-3326. Unhelkar, B. (2005). Transitioning to a mobile enterprise: A three-dimensional framework. Cutter IT Journal, 18(8). Cutter Information LLC. Unhelkar, B. (2006). Handbook of research in mobile business: Technical, methodological, and social perspectives, 1st edition. Hershey, PA: IGI Global. Unhelkar, B. (2009). Mobile enterprise transition and management. NY: Taylor and Francis. Unhelkar, B., Ghanbary, A., & Younessi, H. (2009). Electronic collaboration and organizational synergy. Hershey, PA: IGI Global. Unhelkar, B., Wu, M., & Ghanbary, A. (2008). Integrating mobile technologies in enterprise architecture with a focus on global supply chain management systems. In M. S. Raisinghani (Ed.), Handbook of research on global information technology (pp. 499-518). Hershey, PA: IGI Global. ISBN: 978-1-59904-876-5. Wiehler, G. (2004). Mobility, security, and Web services: Technologies and service-oriented architectures for a new era of IT solutions. Publicis KommunikationsAgentur GmbH, GWA, Erlangen.

Mobile Service Oriented Architecture (MSOA) for Businesses in the Web 2.0 Era

Wu, M. (2007, December 4). Australian Conference on Information Systems, (ACIS) 2007, Doctoral Consortium Paper Extending Enterprise Architecture with Mobility to Create Mobile Enterprise Architecture (M-EA), Toowoomba, Q.L.D. Wu, M., & Unhelkar, B. (2008, May 11-14). Extending enterprise architecture with mobility. 2008 IEEE 67th Vehicular Technology Conference, Singapore.

Key terMs And definitions Enterprise Architecture (EA): Represents a technology-business philosophy that provides the basis for cooperation between various systems of the organization that may be inside or outside the organizational boundary.

Mobile Service Oriented Architecture (MSOA): An approach to extending Service Oriented Architecture (SOA) with Mobile Technologies (MT) Mobile Technology (MT): Include wireless networks, handheld devices and mechanisms to store and present contents MSOA Implementation Framework: The framework focuses on how teams can be organized for effective MSOA implementation Service Oriented Architecture (SOA): an enterprise architecture that links together the applications and web services within organization, across enterprise, and across the Internet Web Service: loosely coupled, self-describing services that are accessed programmatically across a distributed network, and exchange data using vendor, platform, and language-neutral protocols

191

192

Chapter 11

Towards Web 3.0: A Unifying Architecture for Next Generation Web Applications Tzanetos Pomonis University of Patras, Greece Dimitrios A. Koutsomitropoulos University of Patras, Greece Sotiris P. Christodoulou University of Patras, Greece Theodore S. Papatheodorou University of Patras, Greece

AbstrACt While the term Web 2.0 is used to describe the current trend in the use of Web technologies, the term Web 3.0 is used to describe the next generation Web, which will combine Semantic Web technologies, Web 2.0 principles, and artiicial intelligence. Towards this perspective, in this work we introduce a 3-tier architecture for Web applications that will it into the Web 3.0 deinition. We present the fundamental features of this architecture, its components, and their interaction, as well as the current technological limitations. Furthermore, some indicative application scenarios are outlined in order to illustrate the features of the proposed architecture. The aim of this architecture is to be a step towards supporting the development of intelligent Semantic Web applications of the near future, as well as supporting the user collaboration and community-driven evolution of these applications.

introdUCtion Current trends in Web research and development seem to revolve around two major technological pillars: Social-driven applications, a main component in the Web 2.0 domain, and the Semantic Web. It

is our firm belief that web semantics and Web 2.0 are complementary visions about the near future of the Web, rather than in competition: surely they can learn from each other in order to overcome their drawbacks, in a way that enables forthcoming web applications to combine Web 2.0 principles, especially those that focus on usability, community and

DOI: 10.4018/978-1-60566-384-5.ch011

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Towards Web 3.0

collaboration, with the powerful Semantic Web infrastructure, which facilitates the information sharing among applications. Recently, the term Web 3.0 is used to describe the long-term future of the web (Lassila, 2007; Hendler, 2008). Web 3.0 will surely incorporate semantic web and Web 2.0 principles, but researchers believe that it will also include some more sophisticated concepts like artificial intelligence on the web. Towards this direction, in this work we propose a 3-tier architecture for web applications that will fit into the Web 3.0, the next generation web. At the lower layer of the architecture, we introduce and describe an advanced semantic knowledge base infrastructure that can support integration of multiple disparate data sources, without requiring a concrete underlying semantic structure. In addition, the upper layers of the architecture provide greater flexibility in the user interactions with the underlying ontological data model. As a result, it supports user collaboration and communitydriven evolution of the next generation web applications. This architecture gives the developers the ability to build complicated web applications which combine the philosophy of Web 2.0 applications, and the powerful technical infrastructure of the Semantic Web, supported by applying Artificial Intelligence principles on the Web. Furthermore, this architecture is well suited for supporting enhanced Knowledge Systems with advanced knowledge discovery characteristics, towards the future implementation of an Internet-scale Knowledge System. For example, the proposed architecture could be used to enrich current wiki applications towards next generation semantic wiki platforms that will mash-up scattered data sources and provide intelligent search capabilities. The following text is organized in five sections. In section 2 we start by providing some broad definitions and discussing the concepts of Semantic Web and Web 2.0. Furthermore, we discuss related work and the theoretical background of the research area. In section 3, we describe in

detail the proposed architecture, its components, its fundamental features and the current technological limitations. In section 4, we outline some indicative application scenarios in order to illustrate the features of the proposed architecture and prove that it can be applied today and support modern web applications. Finally, we discuss future work and summarize our conclusions.

bACKgroUnd As Semantic Web and Web 2.0 were firstly introduced separately by groups with completely contrary beliefs on the evolution of World Wide Web, and even targeting different audiences, there has been a common perception that both are competing approaches for organizing and emerging the Web. The Semantic Web, outlined by Berners-Lee (2001), becomes a revolutionary technological approach for organizing and exchanging information in a cross-application dimension. Strongly supported by World Wide Web Consortium and powered by heavy academic and enterprise research, Semantic Web can demonstrate standardized and well-defined approaches in language description, such as RDF (Manola, 2004), RDF(S) (Brickley, 2004) and Web Ontology Language OWL (Smith, 2004), as well as research background in ontology engineering and modeling tools, from SHOE (Heflin, 1998) to Protégé (Knublauch, 2004). Semantic Web is powered by a strong AI background through its foundation on the Description Logics (DL) formalism (Baader, 2007). DL languages have become in recent years a wellstudied formalism, originating from Semantic Networks and Frames and, as such, they have been extensively used in formal Semantic Web specifications and tools. These languages are of variable expressive strength which comes with the cost of increased computational complexity. Therefore, current research in this area is focused on efficient and

193

Towards Web 3.0

advanced algorithms and procedures that would provide intelligent querying capabilities for the real word Web, based on DL descriptions and possibly subsets of and reductions from them that may exhibit more satisfying computational properties (Grau, 2008). One main reason for transforming the current Web to a Semantic Web is the ability to deduce new, un-expressed information that is only implied by existing descriptions. If the Web is to be considered as a huge, distributed knowledge base, then well-known AI techniques, at least for the part with sound foundations in logic, can be utilized in order to form the basis for intelligent negotiation and discovery on the Semantic Web. Such techniques may include for example deductive query answering and inference-based reasoning (Luke, 1996; Berners-Lee, 2001). On the other hand, the Web 2.0 term, introduced by Tim O’Reilly (2005), represents a widely spread trend of adopting certain technologies and approaches in web development, targeting more flexible and user friendly applications, and easier distributed collaboration. The usability aspect is met by Rich Internet Applications (RIA) (Loosley, 2006) and especially Asynchronous JavaScript and XML (AJAX), which support the creation of responsive user interfaces as well as more interactive browsing experience. Collaboration conveniences come through the creation of virtual online communities of users that contribute effort and data to a common cause, achieving better results than each individual could do on his own. Finally there is a greater flexibility in data handling, enabling the development of hybrid web applications, called Mash-ups, which combine discrete data sources and services from different sites in order to provide a unified and enriched result. Therefore, the Semantic Web can provide a rich and powerful technical infrastructure for any kind of web application, while the paradigm of Web 2.0 applications can be used to provide useful guidelines, focusing on usability and collaboration. Thus, the Semantic Web and Web 2.0 principles

194

can be combined as complementary approaches to provide more efficient web applications. Such applications could be thought to be part of next generation’s web and seem to fall under the term Web 3.0 (Hendler, 2008), which lately is sort of “talk of the town” (Lassila, 2007). In this context, there are several approaches; from developing AJAX tools for the Semantic Web (Oren, 2006) and studying the combination of ontologies and taxonomies (Mika, 2005), up to the proposition of sophisticated hybrid architectures, combining both of these technologies (Ankolekar, 2007). All of the above are of great use in any datahandling web application, and where there is need for a knowledge system. Especially for next generation knowledge systems that try to benefit from Web 2.0 approaches and collaborative development in order to build, or more precisely grow, Internet-scale knowledge systems (Tenenbaum, 2006).

ProPosed ArChiteCtUre In this section we propose an architecture for web applications, which provides developers the ability to structure complicated web applications, which combine the vision of Web 2.0 and the rich technical infrastructure of the Semantic Web, supported by applying Artificial Intelligence principles. Such applications could be next generation semantic wikis, intelligent mash-ups, semantic portals and in general any data-handling web application that intends to provide semantic information combined with advanced intelligent querying capabilities. The information of these applications could be delivered by two main ways: i. ii.

Directly to end users through the web-based interface of a stand-alone application To other programs or services, that act as intermediaries with third-party web

Towards Web 3.0

applications, by interacting with the API of our semantic infrastructure to retrieve precisely the information they need. A conformant implementation may follow the traditional 3-tier model, which lately (Hendler, 2008) is commonly used to support web 3.0 applications, with an important variation: Where a database server would be typically used, we now use a knowledge base system, since a traditional DBMS lacks the necessary features and functions for managing and utilizing ontological knowledge. Note that each of the three layers may be physically located on different computer systems. The proposed architecture is presented in Figure 1. In fact, from the designer’s point of view, our architecture could be decentralized in at least two ways: i.

ii.

The Semantic Web knowledge bases that data is extracted from, could be both logically and physically distributed (in the case of OWL, this can be accommodated by the owl:import directive) and in such case, an application has to provide for their integration. This is necessary, since Web Ontologies are expected and already tend to be developed in parts and fragments, each addressing a specific view of knowledge. Therefore, it is evident that their combination and alignment could provide richer descriptions and more powerful inferences. The layers of the application could also be distributed both at the logical and the physical level: the front-end layer, the application logic layer and the knowledge management layer.

Such a truly decentralized architecture, in accordance to the traditional 3-tier paradigm, is not yet possible with the majority of the current stateof-the-art and highly expressive inference engines, due to limitations of their interface capabilities,

as described later. On the other hand, such an approach can have a more substantial contribution to the utilization of semantic information by users and applications by eliciting more obvious value from ontological data (Hendler, 2008). The lower part of the proposed 3-tier architecture is a knowledge management layer (or system) which integrates and administers data sources that may be disparate in nature: ontology documents, metadata, feeds and other information with underlying semantic structure of variable density, from semantic data to plain text (zero density). As a result, this layer acts as a semantic mash-up that aligns information to a common, mediating ontology (the core ontology); at the same time this layer performs the low-level reasoning functions that are required in order to deduce implied information. Such an implementation can load Semantic Web Knowledge bases (OWL documents) that are available either on the local file system, or on the Internet. A temporary copy of every document is stored locally and is then loaded by the knowledge base server (an inference engine like RACER). RACER (Haarslev, 2003) can create and store in memory an internal model for each ontology it classifies. Classification takes place once for each ontology, during its initial loading. User requests, queries, additions and other interventions to the ontological model are being interpreted through the application logic layer. This is responsible for the ontological information loading, proper rendering / presentation of it to the user and the decomposition of the user requests to low-level functions of the knowledge management system. Ontological data and reasoning results (Koutsomitropoulos, 2005) are fetched by interacting with knowledge management system, which could physically be located in another machine, e.g. over the TCP/IP protocol. In case of using RACER, this interaction is greatly facilitated through the JRacer API. The application logic can be implemented using the Java programming

195

Towards Web 3.0

Figure 1. The proposed 3-tier architecture

language, as well as JSP, JavaBeans and Java Servlets technologies. Tomcat can be used as an application server. Individual users or users being part of communities may interact with the underlying knowledge base through the front-end layer of the architecture on a reciprocal basis: this means that they are not confined to the mere ingestion of data sources; rather, they are also enabled to fully interact with them, by adding, commenting and

196

incrementing on the underlying ontological data model (user additions). On a standalone web application scenario, this is accommodated through web pages, either static ((X)HTML), or dynamic implementing rich interfaces, where, for example, the user experience is enhanced by the AJAX paradigm in JSP pages rendered by the browser. However, web services, programs/scripts and other interoperability interfaces may also interact as clients with the front-end layer. Communication

Towards Web 3.0

with the application layer can be conducted over the HTTP protocol, using forms.

semantic Mash-Up A conformant application should be able to handle information originating from a number of sources and organized with different levels of semantic density. For the purpose of our work, semantic density of ontology-forming information can be defined as the extent to which the intended models of the ontology can capture the domain conceptualization. This definition is consistent to the definition of ontology, as introduced by Guarino (1998). Roughly, we can distinguish among three types of such information: i.

ii.

iii.

Information that is already adequately described and its semantics expressed in machine readable ways. Ideally, this kind of information is serialized in web ontology languages, such as OWL (or it is trivial to do so). Information that is organized as a flat aggregation of annotations, as it is most often the case with metadata schemata. Such a schema may imply an underlying semantic model but this is not adequately captured. However each annotation has distinguished semantic interpretation from each other. Information that is being given as simple, unorganized text, in the form of natural language. Semantics that may be hidden in such descriptions are not expressed in any way.

In the first two cases, we can employ a technique known as semantic profiling (Koutsomitropoulos, 2007) in order to intensify the semantic density of information. This in turn would increase the expressivity of descriptions leading to the ability to process and respond to more powerful, inference-based queries. Even in the case where

annotations are flatly organized as metadata elements, we can construct a fully-structured ontology model out of them, enriched with new constructs specific to our application or constructs that capture relations already implied in the schema; then, we can align available descriptions to our new ontology by using an automated translation process (e.g. based on XSLT), requiring no enduser intervention. The third case is the trickiest one, since it offers no starting point to be based upon. Necessarily, it would require some form of natural language processing (NLP) (Alani, 2003) in order to identify, for example, keywords that may reveal the subject classification of the textual description. Such keywords can then be mapped to an existing ontology, such as WordNet (http://wordnet.princeton.edu/), in order to extract semantic relations among them and populate, to a limited extend, our common mediating ontology.

Advanced interconnectivity features The front-end layer of the proposed architecture can support stand-alone web applications that provide an enhanced user experience which is accommodated through rich interfaces. Targeting a Rich Internet Application (RIA) (Loosley, 2006), where a web application has the features and functionality of traditional desktop applications, using advanced Web 2.0 approaches, like the AJAX technique, where the necessary processing for the user interface is typically transferred to the web client, but the bulk of the data is kept back on the application server. However, there is prediction for additional interconnection features. A variety of interoperability interfaces may also interact as clients with the front-end layer of the architecture. For example, a conformant web application can facilitate third-party developers integrating its freely distributed semantic information into their web sites, by providing direct, high-level access to the data of its knowledge base, through its API.

197

Towards Web 3.0

Third-party web services can reach a high level of interoperability through the architecture’s API to provide interconnection to third-party web applications using web service specific techniques, e.g. communicating using XML messages that follow the SOAP standard. In addition, programs or scripts (written in any language) that conform to the API of our architecture can have access to the data of the knowledge base. Finally, a thirdparty developer can use web scraping techniques to extract content from any website over HTTP for the purpose of transforming that content into another format suitable for use in his web application.

Community interaction Collaboration conveniences are essential features of this architecture. In order to achieve better results in growing and supporting a conformant application, users should be allowed to contribute effort and data. In this way user information can contribute to the population of the application’s ontology schema. There may be cases however, where the alternation of the ontological schema itself may be desirable. For example, administrators and power users should be able to define new ontology classes or properties and these definitions are incorporated or imported in the central ontology. Of course such alternations are to be done in an incremental-only way, since the knowledge on the Semantic Web is inherently monotonic. Moreover, one has to be careful making these additions, in order to avoid redundancy, i.e. multiple equivalent descriptions that are being repeated. To this end, the frequent classification and consistency checks on the ontology may be helpful, since completely identical descriptions can be identified through reasoning.

198

technological limitations The best choice for the underlying formalism for our methodology is to use at least OWL DL, since this OWL dialect offers a satisfactory expressivity level, adequate to powerful inferences (Horrocks, 2003). However, the majority of the current stateof-the-art and highly expressive inference engines lack in fully supporting the specific requirements of our architecture. FaCT++ and Pellet are currently the only two DL-based engines that appear to fully support the decidable subset of OWL. However they only support DIG 1.1, which is insufficient for full OWL DL support (Dickinson, 2004), a fact that mostly drives the upcoming 2.0 specification. DIG 1.1 communication takes place over HTTP and there is no other TCP/IP-like connectivity support; in the case where a tool or application needs to utilize these reasoners, one may use a programmatic API (e.g. Jena or the Manchester API) that interfaces these reasoners as direct inmemory implementations (Horridge, 2007). This approach may have the advantage of reducing the message-passing load of the DIG protocol, but surely is insufficient for developing truly decentralized Web applications and services for the Semantic Web. As DIG 2.0 specification that would solve the aforementioned problems is currently in flux, these reasoners cannot be used in developing a distributed web service for Semantic Web Knowledge Discovery that would fully support OWL DL. In such a case we should opt for RACER as a DL-based reasoning back-end. RACER used to be dominant in terms of expressivity and interface abilities among DL-reasoners, when Pellet was not even existent. Now, RACER, being freely available for non-commercial purposes, is the only free engine, with expressive strength closest to OWL DL that exposes/maintains an independent, full-featured, IP-compatible communication interface.

Towards Web 3.0

indiCAtive APPliCAtion sCenArios In this section we outline some indicative application scenarios in order to illustrate the features of the proposed architecture and prove that it can be applied today and support modern web applications.

developing the Application Let’s consider a web developer/engineer deciding to use our architecture, in order to develop and run a semantic wiki specializing in Cultural Heritage. The first thing he has to do is to design the proper ontology, based on OWL, in order to completely describe the desired information that is to be presented through his site. This can be information about monuments, historical artifacts, ancient manuscripts, or even modern bibliography about cultural heritage. For this particular domain, a good starting point may be the CIDOC Conceptual Reference Model, a recent ISO standard (Crofts, 2003). The next step is to decide whether he is going to use only locally created and stored information, as of a usual semantic wiki, or he is also going to gather information through the web. In the latter case he can search for sites with similar content and categorize them based on the density of their underlying semantic structure. Afterwards he has to map this information to his own ontology, either by using semantically enhanced application profiles (Koutsomitropoulos, 2007), for information with notable semantic structure, or by using natural language processing techniques, e.g. (Alani, 2003), for simple, unorganized text, like the one he can get from Wikipedia (http:// www.wikepedia.org) articles. Now he is ready to pick up the suitable software components for his conformant implementation of our architecture that will support his application. Such a combination could be: RACER as

the inference engine, Tomcat as the application server and Java server-side technologies, e.g. JSP, JavaBeans and Java Servlets. Finally, he develops a user-friendly and highly interactive front-end for his application, using for instance JSP supported by AJAX techniques. His main focus should remain to provide his end-users with web application modules (components) that make easier, not only making intelligent queries and entering new information in the wiki’s knowledge base, but also fully interacting with data sources, by adding, commenting and incrementing on the underlying ontological data model. As a result of the described bottom-up development procedure, the desired application is up and running. Based on the development infrastructure, the information mediated through this semantic wiki can be then collaboratively manipulated and enriched by its target users.

intelligent Querying One of the main advantages of such a web application is to make possible for end-user to submit intelligent queries. Take, for example, the case where, in the underlying ontology, there is the expression that a sword is made of iron. The OWL description for this would be:





The expression describing that iron is metal would be:

199

Towards Web 3.0



One can now retrieve every metal item, via the following expression.



A more complex example is the notion of “co-author”, which is of great use in the context of applications that host resources developed collaboratively (e.g. books, research papers in repositories, digital libraries items, wikis etc.). This particular relation in not often explicitly captured in metadata, for example the DC does not provide any field for this. A co-author relation, which is held among authors, is implied by author relations that exist between authors and items. In particular, consider an author A. This author is in “co-author” relationship with all other authors that are in “author” relationship with the items that A has authored. This kind of relation is a typical example of the need for role-chains that are accommodated by OWL 1.1 (not even OWL DL). In Description Logics syntax (Baader, 2007), the notion of “coauthor” can be described as:

incrementing the ontological data Model Now let’s imagine a user of the above semantic wiki. He has spent some time using the application and has become familiar enough with entering and editing content about the topics related to this cultural heritage wiki. As he is really interested in the cultural domain, he notices that although this wiki is filled with a large amount of information about monuments, historical artifacts, manuscripts, and literature in general, it lacks specific information about paintings, although they are strongly considered to be a discrete field of interest in cultural heritage. Furthermore, he notices that the specific painting part was not taken into consideration during the initial design procedure of the application’s underlying ontological data model, so he is not able to enter information about paintings in this wiki. As he has now become an experienced user, sort of power user, of this application, he is aware of all its potential. This one is not a simple semantic wiki where users are confined to the mere insertion of data, but provides its users with advanced web application modules (components) in order to facilitate them to fully interact with its infrastructure. As a result, he decides to take advantage of this feature and enrich the underlying ontology, by including the required schema for painters. For example he can define the class of “Painters” as a “Person” who has “performed” at least one “Painting_Event”. Note also that the class “Person” and the property “performed” refer to another ontological schema, namely CIDOCCRM:

author– ∘ author ⊑ co_author where – stands for inverse relation, ∘ for role composition, and ⊑ is the sub-property relation. Notice also that this kind of relation cannot be (easily) described even in traditional DBMSs.

200



Towards Web 3.0





in a suitable ontology, creating a semantic mashup. Thus, a common user of this portal, not only has all the information he needs in a single site, but additionally he could benefit from advanced features of semantic personalization (Tziviskou, 2007; Ankolekar, 2006) and intelligent querying support.

As a result this semantic wiki is now ready to receive information also for paintings and painters.

fUtUre reseArCh direCtions

Acting as a semantic Proxy While this semantic wiki works fine as a stand alone web application, which provides its users with comprehensive information for cultural heritage topics, it has also a lot to offer to other web applications. This application is based on an open architecture, whose content and especially its underlying ontological data model are freely available, and therefore it could act as a proxy of semantically structured data. Thus every developer who wants to build an informational site for cultural heritage does not have to collect it over the Web and map it to a new ontology. All he has to do is to use this application’s advanced interconnection features, e.g. its API, to have a unique repository of the desired information, and use it as is or further map this wiki’s ontological model to his own one, a procedure which becomes trivial.

Regarding the future work, it will include both implementations and research work that can be summarized in the following points: •







Specify, design and develop indicative web applications based on our architecture, in order to demonstrate, study and evaluate its features and potentials. Make these pilot web applications available and encourage users to participate, comment and enrich underlying ontological model. Study and evaluate the user collaborations and the community-driven evolution of the applications. Investigate analytically current web technologies in order to decide which ones are best it into our architecture. Get feedback from other researchers and web developers on our proposed architecture and modify or enrich it.

ConClUsion other indicative Applications Another indicative application could be a semantic movie portal. In such a portal, information for movies could be collected from Internet Movie Database (http://www.imdb.com) using ordinary web scraping techniques, while information about respective DVD releases could be collected from the Amazon website (http://www.amazon.com) using its API. All this information could be unified

In this work we have shown that Semantic Web and Web 2.0 can be complementary visions about the future of the Web, rather than in competition. This was done by the proposition of a unifying architecture, which can be used to support any data-handling web application. Such applications could combine the philosophy of Web 2.0 applications, and the powerful technical infrastructure of the Semantic Web, supported by applying

201

Towards Web 3.0

Artificial Intelligence principles on the Web. Applications with such features are considered to be the next generation web applications, or Web 3.0 applications. Semantics and knowledge-discovery capabilities play a key role in this unifying architecture. We recognize, from a methodological point of view, reasoning and inferences as prominent features in Semantic Web scenarios that are necessary in order to enable intelligent services. Therefore, the lower part of the proposed 3-tier architecture is a knowledge management layer, where a database server is typically used in other architectures. This layer can support the integration of multiple disparate data sources, without requiring a concrete underlying semantic structure. User requests, queries, additions and other interventions to the ontological model are being interpreted through the application logic layer. Finally, the front-end layer of the architecture supports on one hand the rich interaction with users (and communities), and on the other hand the interoperability with other web applications through web services or other programs. Overall, the proposed architecture is a step towards supporting the development of intelligent semantic web applications of the near future as well as supporting the user collaboration and community-driven evolution of these applications.

referenCes Alani, H., Kim, S., Millard, D. E., Weal, M. J., Hall, W., Lewis, P. H., & Shadbolt, N. R. (2003). Automated ontology-based knowledge extraction from Web documents. IEEE Intelligent Systems, 18(1), 14–21. doi:10.1109/MIS.2003.1179189

202

Ankolekar, A., Krötzsch, M., Tran, T., & Vrandecic, D. (2007). The two cultures: Mashing up Web 2.0 and the Semantic Web. In WWW ‘07: Proceedings of the 16th International Conference on World Wide Web (pp. 825-834). New York: ACM Press. Ankolekar, A., & Vrandecic, D. (2006). Personalizing Web surfing with semantically enriched personal profiles. In M. Bouzid & N. Henze (Ed.), Proc. Semantic Web Personalization Workshop. Budva, Montenegro. Baader, F., & Nutt, W. (2007). Basic description logics. In F. Baader, D. Calvanese, D. McGuinness, D. Nardi & P. F. Patel-Schneider (Eds.), The description logics handbook (2nd ed.). Cambridge. Berners-Lee, T., Hendler, J., & Lassila, O. (2001). The Semantic Web. Scientific American, 5. Brickley, D., & Guha, R. V. (2004). RDF vocabulary description language 1.0: RDF schema. W3C Recommendation. Retrieved on February 10, 2004, from http://www.w3.org/TR/rdf-schema/ Crofts, N., Doerr, M., & Gill, T. (2003). The CIDOC conceptual reference model: A standard for communicating cultural contents. Cultivate Interactive, 9. Retrieved from http://www.cultivate-int. org/issue9/chios/ Dickinson, I. (2004). Implementation experience with the DIG 1.1 specification (Tech. Rep. HPL-2004-85). Bristol: Hewlett Packard, Digital Media Sys. Labs. Grau, B. C., Motik, B., Wu, Z., Fokoue, A., & Lutz, C. (2008). OWL 2 Web ontology language: Profiles. W3C Working Draft. Retrieved from http://www.w3.org/TR/owl2-profiles/ Guarino, N. (1998). Formal ontology and information systems. In N. Guarino (Ed.), Formal ontology in information systems. Proceedings of FOIS’98 (pp. 3-15). IOS Press.

Towards Web 3.0

Haarslev, V., & Möller, R. (2003). Racer: A core inference engine for the Semantic Web. In Proc. of the 2nd Int. Workshop on Evaluation of Ontologybased Tools (EON2003) (pp. 27-36), Florida. Heflin, J., Hendler, J., & Luke, S. (1998). Reading between the lines: Using SHOE to discover implicit knowledge from the Web. In AI and Information Integration: Papers from the 1998 Workshop (pp. 51-57). AAAI Press. Hendler, J. (2008). Web 3.0: Chicken farms on the Semantic Web. Computer, 41(1), 106–108. doi:10.1109/MC.2008.34 Horridge, M., Bechhofer, S., & Noppens, O. (2007). Igniting the OWL 1.1 touch paper: The OWL API. In Proc. of the OWL Experiences and Directions Workshop (OWLED’07), Innsbruck, Austria. Horrocks, I., Patel-Schneider, P. F., & van Harmelen, F. (2003). From SHIQ and RDF to OWL: The making of a Web ontology language. Journal of Web Semantics, 1(1), 7–26. doi:10.1016/j.websem.2003.07.001

Lassila, O., & Hendler, J. (2007). Embracing “Web 3.0”. IEEE Internet Computing, 11(3), 90–93. doi:10.1109/MIC.2007.52 Loosley, C. (2006). Rich Internet applications: Design, measurement, and management challenges. White Paper. Keynote Systems. Luke, S., Spector, L., & Rager, D. (1996). Ontology-based knowledge discovery on the World Wide Web. In A. Franz & H. Kitano (Eds.), Working notes of the Workshop on Internet-Based Information Systems at the 13th National Conference on Artificial Intelligence (AAAI ‘96) (pp. 96-102). AAAI Press. Manola, F., & Miller, E. (2004). Resource description framework (RDF) primer. W3C Recommendation. Retrieved on February 10, 2004, from http:// www.w3.org/TR/rdf-primer/ Mika, P. (2005). Ontologies are us: A unified model of social networks and semantics. In Proc. 4th International Semantic Web Conferences (ISWC05) (pp. 522–536). Galway, Ireland.

Knublauch, H., Fergerson, R. W., Noy, N. F., & Musen, M. A. (2004). The protégé OWL plugin: An open development environment for Semantic Web applications. In Proc. 3rd International Semantic Web Conference (ISWC04). Springer.

O’Reilly, T. (2005). What is Web 2.0–design patterns and business models for the next generation of software. Retrieved on September 30, 2005, from http://www.oreillynet.com/pub/a/oreilly/ tim/news/2005/09/30/what-is-web-20.html

Koutsomitropoulos, D. A., Fragakis, M. F., & Papatheodorou, T. S. (2005). A methodology for conducting knowledge discovery on the Semantic Web. In Proc. of 16th ACM Conference on Hypertext and Hypermedia (Hypertext 2005), International Workshop on Adaptive and Personalized Semantic Web. Salzburg, Austria.

Oren, E., Delbru, R., & Decker, S. (2006). Extending faceted navigation for rdf data. In I. Cruz & S. Decker (Ed.), Proc. 5th International Semantic Web Conference (ISWC06) (pp. 559–572), Athens, Greece.

Koutsomitropoulos, D. A., Paloukis, G. E., & Papatheodorou, T. S. (2007). Ontology-based knowledge acquisition through semantic profiling. An application to the cultural heritage domain. In Proc. of the 2nd International Conference on Metadata and Semantics Research (MTSR 2007). Corfu, Greece. CD-ROM.

Smith, M. K., Welty, C., & McGuinness, D. (2004). OWL Web ontology language guide. W3C Recommendation. Retrieved on February 10, 2004, from ttp://www.w3.org/TR/owl-guide/ Tenenbaum, J. M. (2006). AI meets Web 2.0: Building the Web of tomorrow, today. AI Magazine, 27(4), 47–68.

203

Towards Web 3.0

Tziviskou, C., & Brambilla, M. (2007). Semantic personalization of Web portal contents. In WWW ‘07: Proceedings of the 16th International Conference on World Wide Web (pp. 1245-1246). New York: ACM Press.

Key terMs And definitions 3-Tier Architecture: 3-tier architecture is a client-server architecture in which the user interface, functional process logic (“business rules”), computer data storage and data access are developed and maintained as independent modules, most often on separate platforms. Knowledge System: A knowledge system (a.k.a. knowledge-based system) is a program for extending and/or querying a knowledge base. A knowledge base is a collection of knowledge expressed using some formal knowledge representation language. Mash-Up: A mash-up is a web application that combines data from more than one source into a single integrated tool. Ontology: An ontology is a formal representation of a set of concepts within a domain and the

204

relationships between those concepts. It is used to reason about the properties of that domain, and may be used to define the domain. Ontologies are used as a form of knowledge representation about the world or some part of it. Semantic Web: The Semantic Web is an evolving extension of the World Wide Web in which the semantics of information and services on the web is defined, making it possible for the web to understand and satisfy the requests of people and machines to use the web content. It derives from W3C director Tim Berners-Lee’s vision of the Web as a universal medium for data, information, and knowledge exchange. Web 2.0: Web 2.0 is a term describing the trend in the use of World Wide Web technology and web design that aims to enhance creativity, information sharing, and, most notably, collaboration among users. Web 3.0: Web 3.0 is a term used to describe the future of the World Wide Web. Following the introduction of the phrase “Web 2.0” as a description of the recent evolution of the Web, many technologists, journalists, and industry leaders have used the term “Web 3.0” to hypothesize about a future wave of Internet innovation.

Section 4

Information Search, Bookmarking, and Tagging

206

Chapter 12

Web 2.0—Social Bookmarking: An Overview of Folksonomies Richard Derham University of Canterbury, New Zealand Annette Mills University of Canterbury, New Zealand

AbstrACt Folksonomies is a relatively new concept and, as yet, it has not been widely studied in academic circles. In practice, folksonomies have therefore outpaced academic research in inding solutions to the problems facing them. The goal of this chapter is to bring together the current literature on folksonomies and explore avenues for future work. Hence, this chapter will examine what are folksonomies, what they are/can be used for, and explore their beneits and challenges using real world examples from systems such as Delicious and Flickr. The chapter also overviews some of the current research and suggests avenues for further work.

introdUCtion The World Wide Web (WWW) has been growing at a phenomenal rate over the last decade as more and more resources of diverse types are added to the Internet daily. While many sites such as ecommerce sites tend to rely on web analytics and various usability features to elevate them in the search lists, information sites especially those created by amateurs and other information-oriented content (e.g. images, music, video) are sometimes more difficult to locate and index. Social bookmarking DOI: 10.4018/978-1-60566-384-5.ch012

systems address this gap by providing a method that enables users to create and apply bookmarks (tags) to information content that they want to retrieve at a later stage. Consisting of freely chosen keywords, these tags can then be organized, managed, indexed and shared with others for later retrieval of the content. Also referred to as collaborative tagging, social indexing, and social tagging, social bookmarking is gaining popularity. This is due in part to the inadequacies of taxonomies for indexing and retrieving the vast amount of content now available on the web and elsewhere. Most bookmarking services are free; some also provide free storage for users (e.g.

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Web 2.0—Social Bookmarking

Flickr for digitized images). Advances in social software applications are also enabling better indexing and sharing of content, especially content that would normally be overlooked or ranked low by search engines. Although the basic concept of user-defined tagging itself is not new, emergent forms of social bookmarking such as folksonomies that have come about with Web 2.0 are relatively new concepts. Deriving from the activity of social bookmarking, folksonomies comprise freely chosen tags or keywords used by individuals to classify content for later retrieval of that content, and sharing the content with others. So although these userassigned tags are often created for personal use in most cases they are made public that is, available to others so they can locate and retrieve the same or related content. This allows the sharing of content with others interested in the topic area and the forming of communities of people with similar interests. This openness and sharing enables the social aspect of bookmarking. Folksonomies are gaining popularity as they become more widely used across various social software applications. This meteoric rise in popularity is largely attributed to developments and trends in Web 2.0 in areas such as technology/ software development, information retrieval, and collaboration among users. Although the mechanisms and subject matter for social tagging may vary across systems, the collaborative open nature of the folksonomy tends to be shared by most systems. Given the popularity of collaborative systems and the services that support and enable these forms of user-driven tagging (e.g. Delicious for bookmarks, Flickr for digitized images, Connotea and CiteULike for bibliographic data), it is becoming increasingly important for practice (and hence researchers) to address the many problems and issues that relate to folksonomies, social tagging, information retrieval and individual and communal behaviors. These social tagging systems may also afford multiple benefits to organizations enabling and supporting

informal networks within firms for resource and knowledge management, information sharing and retrieval, social networking and expert discovery (Damianos et al., 2007). In practice, folksonomies have outpaced academic research in finding solutions to the problems facing them. As academic interest in folksonomies increases, researchers are largely focusing on issues linked to information retrieval such as addressing the ambiguities that can arise when individuals use different tags to refer to the same content, and extracting semantic structures from folksonomies (Mika, 2007; Spiteri, 2007). While there has been some success in these areas, little has been done to understand the motivations and behaviors of those engaged in these social communities (Marlow et al., 2006) or explore the value of folksonomies in business and social settings (Damianos et al., 2007). Hence, the goal of this chapter is to bring together the current literature on folksonomies and explore avenues for future work, particularly as these relate to the behavioral and social dimensions of folksonomies. The chapter will therefore explain what are folksonomies, what they are used for, and outline the advantages and challenges of folksonomies. Solutions to some of the challenges of folksonomies are also examined, as well as avenues for further work in information retrieval. There are also many opportunities for researchers to explore the motivations and behaviors of the online communities that form around these tags.

bACKgroUnd what is a folksonomy? A folksonomy is the result of personal free tagging of information and objects (i.e. anything with a URL) for one’s own retrieval. The tagging is done in a social environment and is usually shared and open to others (Vander Wal, 2007). A folksonomy therefore arises from the act of tagging by the

207

Web 2.0—Social Bookmarking

person who is also consuming the information. A folksonomy consists of freely chosen tags or keywords that are selected and used by an individual to classify information and objects (e.g., a webpage in Delicious or a video in YouTube) and bookmark the item. The tags are freeform and as such are whatever the creator/saver of the bookmark wants them to be. A folksonomy therefore allows users to create and assign tags to content, using these to index the content so that the tagger can retrieve the item when needed and others interested in the topic can use the tag to find the item. Folksonomies are often a result of ‘collaborative tagging’ (Macgregor & McCulloch, 2006) as both sharing and tagging contributes to the form of the folksonomy. However, not all systems making use of tags refer to these as folksonomies. For instance, Gmail also uses tags but these tags are referred to as a personomy as they are for the use of just one person (Hotho et al 2006a). The term ‘folksonomy’ is a portmanteau word derived from blending the words ‘folk’ and ‘taxonomy’. The word ‘folk’ is used because it is created by the people rather than by experts and, ‘taxonomy’ because it represents a conceptual indexing system for categorizing data (Hotho et al., 2006a). Invention of the term, ‘folksonomy’, is attributed to Thomas Vander Wal who used it in a posting to an information architecture mailing list in 2004 (Vander Wal, 2007). The term folksonomy is considered by some as slightly inaccurate and misleading as while taxonomies are hierarchical classification systems, folksonomies are non-hierarchical categorizing systems (Paolillo & Penumarthy, 2007) that rely on an emergent semantic derived from the convergence of language rather than a formalized semantic (Hotho et al., 2006a). Hence the term ‘social tagging system’ may be a more accurate term to describe the phenomenon. However ‘folksonomy’, despite its flaws, is the term that has become synonymous with this indexing method.

208

Folksonomies is one of the phenomena that have emerged with the Web 2.0 era. Harnessing the Internet as the technology platform, folksonomies exhibit key attributes of a Web 2.0 technology such as openness, community, and interaction. They are widely used by various social software applications and services including well-known social bookmarking systems such as Delicious (formerly del.icio.us) and media sharing sites such as Flickr and YouTube. The rapid uptake of folksonomies has been largely related to the trend towards amateurs publishing content on the web through different social software applications (e.g. digital images in Flickr, videos on YouTube, blog postings, etc). As amateur-created content increases, the creators and users of this content are developing and expanding the classification schemes for that content. Folksonomies therefore represent a trade-off between a traditional classification system (taxonomy) and having no classification at all.

Key Aspects: resource Categorization, tags and Users. Folksonomies essentially consist of three dimensions: resources, tags and users. A key aspect of folksonomies is the categorization of information resources. Categorization involves taking ideas and grouping them into categories in order to understand them, and differentiate between them. Folksonomies provide a way of categorizing information (using tags) that differs from more formal methods of indexing such as taxonomies. For example, taxonomies are typically formal, ontological top-down classification systems prepared by experts (such as librarians). By contrast, folksonomies are informal, bottom up non-hierarchical systems that comprise of tags which are created (and extended) by anyone who wishes to contribute to the system (Hotho et al., 2006a). Hence, while taxonomies have semantic structures built into them by design, such patterns emerge over time in folksonomies. Folksonomies

Web 2.0—Social Bookmarking

therefore develop their structure organically as they are contributed to by more and more users, and convergence in the use of language is achieved (Hotho et al., 2006a; Rattenbury et al., 2007). Folksonomies are therefore more easily extended than taxonomies which would require extensive reworking of the structure to include new terms. This openness and extendibility permits folksonomies to be more responsive to change than taxonomies. Folksonomies therefore differ in a number of ways when compared with taxonomies, and it is these differences that have lead to the many challenges associated with folksonomies. Tags are also central to the development of a folksonomy. Based on their function, Golder and Huberman (2006) identified seven different types of tags in use: • • • • •







Identifying what it is about (e.g. ‘Beijing Olympics’) Identifying who it is about (e.g. ‘Winston Churchill’) Identifying what it is (e.g. ‘photo’) Identifying who owns it (e.g. ‘government’). Reining categories are tags that provide additional information supporting other tags; often round numbers (e.g. 25, 100) might be used. Reining categories tend to be much more individual and of little or no use to another user. Identifying qualities or characteristics. These tags link particular attributes (e.g. ‘funny’) to the content. These provide the tagger’s opinion of the material. Self reference tags are another highly individual type that relates the bookmark to the tagger. For example, the tag ‘mystuff’ would signify that the bookmarked material is the tagger’s own. Task organizing tag is also an individual tag. The classic example is the ‘toread’ tag, which is an instruction to the tagger to read this piece at a later time.

Since the Golder and Huberman (2006) paper, a newer tag - the geotag - has been gaining prominence since late 2007. The geotag allows users to add geographical or location metadata (e.g. longitude and latitude coordinates, place names) to content. When tagging, users will often apply multiple tags to a particular content. Research examining the trends in tagging show these users will often apply a more general tag that is in widespread use first, in their list of tags followed by more specific tags (Golder & Huberman, 2006; Hotho et al., 2006b). Folksonomies not only provide a means for categorizing content using unstructured tags, it also enables sharing of categorized data with others. The shared nature of a tag is an essential feature of a folksonomy, as it allows others to see how a tag is used and to view or use the tags that others have created. As a social phenomenon, a well-developed folksonomy derives from a continuous loop of use (indexing), search and retrieval, and examination and feedback that enables a community of users to shape the folksonomy (i.e. its vocabulary, meaning and use), encourage useful tags, and remove useless tags. Although a folksonomy arises initially from the personal tagging of web-based content, the social aspect of its use and evolution distinguishes it from other types of free-form tagging. Users therefore play a significant role in the development of the folksonomy, being motivated to tag content for organizational and/or social reasons (Marlow et al., 2006). Users who tag content for organizational reasons will use tagging as a means of managing content, while those who tag for social reasons will use tags for communication - to converse with others about specific content or to express themselves or their opinions. In addition to these motivations, Marlow et al (2006) suggested six incentives for tagging that may influence the utility and use of tags:

209

Web 2.0—Social Bookmarking













Future retrieval. This is tagging in order to come back to the bookmarked item later (e.g. using a tag such as ‘toread’ or ‘blogthis’). Contribution and sharing. This is tagging in order to provide data that will be useful for others inding or assessing the item. Attract attention. Here users apply popular tags to an item so that it can be seen by a large number of people. However in some cases (e.g. tag spamming) the tags used may bear no relation at all to the content. Opinion expression. This is tagging an item to convey a value judgment about the item bookmarked. Play and competition. Here tagging is based not on the content itself but rather as a form of entertainment. For instance, a user might wish to make a particular tag the biggest in their tag cloud. Self presentation. This is tagging a resource in order to leave one’s mark on the particular resource. For example, concert footage on YouTube could be tagged ‘there’.

folksonomies in Action Although the concept of shared online bookmarking has been around since the late 1990s, social bookmarking and development of folksonomies began to gain momentum around 2004 fueled in part by competitiveness of and advances in social bookmarking services. Several players gained significant prominence over the new few years. These include general bookmarking services such as Delicious, Furl and Magnolia; digitized image services such as Flickr and YouTube; reference management services such as Connotea and CiteULike; and enterprise social bookmarking services such as IBMs Lotus Connections Dogear and Connectbeam. Using examples, the next section overviews key features of different types of social bookmarking services.

210

Delicious (http://Delicious.com/) is one of the early pioneers of social tagging with over 3 million registered users and 100 million unique URLs bookmarked. It is a free open-ended social bookmarking website that enables users to bookmark and manage items on the web for their own use and to share these with other users. The system uses a non-hierarchical system of freely chosen one-word descriptors (tags) for tagging items. For example, if the user searches for the tag “elephant”, they are taken to a webpage that lists all the items tagged ‘elephant’. Users can then view a particular item or they can view all the bookmarks saved by another person using the ‘elephant”. The ‘taxonomy’ part of the folksonomy is satisfied as this is a classification system, and the ‘folk’ part of folksonomy is satisfied because other users can find pages that have been bookmarked by searching or browsing on a particular tag or user (Golder & Huberman, 2006). To enable the collective aspect of social book marking, Delicious will show the most popular bookmarks (see Figure 1) as well as recently added items and the tags assigned to them. By browsing the various pages, users can get a sense of what other people are interested in and find or form communities of people with common interests. Flickr (www.flickr.com) is a well-known digitized image and video hosting service. In 2007, Flickr claimed to have more than 40 million visits per month, 3 billion photos stored and 2-3 million new photos uploaded each day (Auchard, 2007). Like other bookmarking services (e.g. Delicious) Flickr allows users to tag their own content and that of others; photos can also be made public or private. User can also find similarly themed content using a particular tag. For example, entering the tag ‘africa’ will take the user to photos (or videos) tagged with ‘africa’. However, unlike general bookmarking services such as Delicious, Flickr (and also YouTube) only allows users to tag content that has been uploaded to the site; users therefore cannot tag content that is held

Web 2.0—Social Bookmarking

Figure 1. Delicious “in action” – Popular Bookmarks (Source: http://delicious.com/.) Reproduced with permission of Yahoo! Inc. ©2009 Yahoo! Inc. DELICIOUS and the DELICIOUS logo are registered trademarks of Yahoo! Inc.

elsewhere. Although Flickr also allows users to tag the content of others and to add comments, most of the content are tagged by the content owner rather than tagged by other users (Marlow et al., 2006). Flickr also allows users to geotag content. Geotagging in Flickr is done using machine tags. These are a particular type of tag with syntactic content in the form, namespace:predicate=value; this allows the storage of extra information about the tag. For example, using machine tags a photo could be geotagged with two tags, geo:long=123.456 and geo:lat=123.456, representing the longitude and latitude coordinates linked to that image. Users can then search by over 100,000 place names to find content of interest (Auchard, 2007). Reference management sites such as Connotea (www.connotea.org) are not as popular as services such as Delicious or Flickr, in part because they

are more specialized. Connotea is a ‘free to use’, open-source online reference management and bookmarking service created by the Nature Publishing Group. It is primarily aimed at researchers and allows them to track and share scholarly articles and references online. Bibliographic information can be held as public, private or shared with a particular group, and can be imported from/exported to desktop reference management systems using various file formats such as RIS, EndNote, and BibTeX (Connotea, 2008). Like Delicious, Connotea allows users to tag websites. It also permits online storage for references and bookmarks, simple non-hierarchical organization of bookmarks, opening the list to others, the auto-discovery of bibliographical information, and geotagging. Connotea also provides RSS feeds for new content and it allows users to add and view comments linked to articles (Connotea, 2008).

211

Web 2.0—Social Bookmarking

Dogear is an enterprise social bookmarking system developed by IBM. Developed to support information management and sharing in corporations and large organizations, Dogear allows users within the organization to bookmark pages on the company intranet. The system provides the typical benefits of social bookmarking (e.g. information retrieval for the individual and information sharing between users), allowing users to tag and share intranet resources. Unlike public services such as Delicious, Dogear requires users to be authenticated against a company directory and to use their real names (and not pseudonyms). Use of real names helps organizations to build a directory of expertise as users can find individuals’ contact information by drilling through from tags. For example, a user can refer to the ‘HTML’ tag to see who has been bookmarking pages on this topic and then approach that person for help with web development (Millen et. al., 2005).

Figure 2. Locating folksonomies in research

212

CUrrent stAte of folKsonoMies Folksonomies as a type of computer-mediated collaboration sits at the intersection of two fields in information systems: social computing and information retrieval. Folksonomies contain key features from both domains (See Figure 2). For example, while it inherits its tagging and metadata aspects from information retrieval, its sharing and communication aspects are derived from social (or collaborative) computing. In fact, as will be shown later in this chapter, this situation leads to a certain amount of tension between two fundamental aspects of folksonomies – tagging and collaboration. Folksonomies have several advantages over hierarchical categorization systems, including low barriers to participation, providing feedback and immediate benefit to users, and making information retrieval of certain types of content/objects easier. On the other hand, several issues also arise. For example, folksonomies cannot distinguish be-

Web 2.0—Social Bookmarking

tween polysemes, are blind to synonyms, contain basic level variation between users, and suffer from entropy and noise. There is also a fundamental tension since users have both highly individual and highly social uses for the folksonomy. These issues are outlined in the next sections.

Advantages of folksonomies Folksonomies have come to prominence because they hold several advantages over other methods of characterizing metadata. These include low barriers to participation, immediate benefit to users, immediate feedback, open-endedness, browsing and unanticipated uses (Mathes, 2004; Wu, Zhang et al., 2006)

Low Barriers to Participation The most significant feature is that folksonomies do not require specialized knowledge about the system in order to use it. This means that anyone can use it without needing training (Wu, Zhang et al., 2006). By contrast, in the case of information professionals (such as librarians) maintaining a taxonomy, the overhead for participation could involve several long and expensive years of training.

Immediate Benefit to Users Related to the advantage of low barriers to entry is the immediate benefit a folksonomy brings to users (Hotho et al., 2006b). For example, users who tag photographs in Flickr, or employees who tag content on a corporate intranet will derive benefit from their efforts the very next time they want to retrieve any of these items. Additionally, they can benefit from the classification work that others have undertaken in the meantime. Over time folksonomies enable individuals to manage information resources, to locate experts, and to form and build communities of users with similar interests.

Feedback Once a tag is assigned to an item, it is possible to immediately see what items other users have assigned that tag to. This feedback allows users see whether their understanding of the tag aligns with other users; it therefore provides an opportunity for the user to change the tag or enhance the entry with additional tags (Mathes, 2004; Wu, Zhang et al., 2006)

Open-Endedness Folksonomies are inherently open-ended as any keyword a user wants to use is permissible. Folksonomies can therefore respond quickly to changes and innovations (Wu, Zubair et al., 2006). This aspect compares favorably to the more time intensive and hence more costly operation of changing a taxonomy.

Browsing and Finding Folksonomies are also useful when browsing for information as their use of hyperlinks attached to the tags can help increase the efficiency and effectiveness of user browsing (Mathes, 2004). Using folksonomies, a user can move easily from one tag (e.g. ‘Lamborghini’) to a different yet related tag (e.g. ‘Ferrari’). Browsing can therefore expose users to a wider set of results than keyword searching and take them into slightly different areas which may have the advantage of increasing breadth of knowledge. Users can also discover others who share similar interests and thus acquire links to additional resources pertaining to their own interests (Wu, Zubair, et al., 2006). A key element that distinguishes browsing via the pathways of a folksonomy from that of traditional taxonomies or ontologies is that folksonomies allow users to traverse paths (or ‘desire-lines’) that reflect their choices, such as the terms used or the level of precision desired (Merholz, 2004). By contrast, taxonomies provide

213

Web 2.0—Social Bookmarking

pathways that are defined by professionals of creators of the content rather than by the users of the content (Mathes, 2004).

Unanticipated Uses Folksonomies also lend themselves to being used in new ways by individuals or small groups of users. Wu, Zubair et al. (2006) describe the ‘long tail’ of extremely low frequency tags being used in idiosyncratic ways that would never be included in a more formally constructed system. For example, the tag ‘sometaithurts’ (so meta it hurts) is applied to a photo of someone using Flickr (Mathes, 2004). Other users can join in a type of conversation or community by applying this tag to their photograph.

ChAllenges Although folksonomies have several advantages, there are elements that inhibit their usability and usefulness in different situations. While many of these relate to problems of language arising from the use of an unrestrained vocabulary, others relate to anti-social behaviors such as spamming. Key issues therefore include ambiguity problems related to the use of polysemes, homonyms and synonyms, basic level variation, entropy, the use of spaces and multiple words, as well as tag spamming.

Ambiguities in language Tag ambiguities arise from many sources including polysemes, homonyms and synonyms, plurals, different languages, spaces and multiple words.

Polysemes Polysemes are words that have multiple related meanings, while homonyms have multiple different (unrelated) meanings. One example of a

214

polysemy is the word ‘foot’ which could refer to the part of the leg of a vertebrate below the ankle, an organ that an invertebrate uses to move itself along, the lowest part of something (e.g. the foot of the hill), or the lower part of the stem of a plant. In all of these senses the word ‘foot’ relates to the base, bottom or end of something. While the exact meaning of a polysemous word is not usually ambiguous in its context (e.g. ‘the house at the foot of the hills’ or ‘the foot soldier’) the less specific the search context, the greater the ambiguity introduced into the search and hence the search results. In a more formal setting, taxonomies could be crafted to work around polysemes but this is not possible in a folksonomy.

Homonyms Homonyms are also an issue in folksonomies although less so. For example, the tag ‘starwars’ might apply to the science fiction franchise and movie series created by George Lucas and to a missile defense initiative conceived by Ronald Reagan. This ambiguity can be resolved in a search of the tags by specifying additional terms such as ‘starwars NOT movie’ or ‘starwars NOT missile’, to eliminate the unwanted homonym. However, this would only work if the tags attached to the material contained all of the specified words.

Synonyms Synonyms occur where two or more words have the same (or nearly the same), meaning. Synonyms can pose a greater problem than a polysemy or homonym as a user can never be sure which synonyms have been used or they may not be aware of certain synonyms (Golder & Huberman, 2006; Guy & Tonkin, 2006). The problem of synonymy is intensified in a collaborative free-form environment. For example, a ‘pupil’ may also be referred to as a ‘student’, ‘scholar’, ‘learner’, ‘novice’, ‘beginner’ or by the less common term, ‘neophyte’. Even an individual tagger may apply

Web 2.0—Social Bookmarking

these tags inconsistently. Hence, it may be difficult for a searcher to determine whether they have retrieved all the relevant items (or even the most relevant items).

Plurals and Different Languages Related to the problem of synonymy is the use of plurals and different languages in tagging. The use of plurals as a tag contributes to the problem of synonymy as a tag and its plural (e.g. rabbit and rabbits) may be considered two completely different terms in a folksonomy. Hence, a search for one will not return the other unless the retrieval system has the capability to include related searches (e.g. Delicious). Similarly, although most systems exhibit a bias towards the English language, in keeping with the openness of social tagging, different languages are also used to tag content (e.g. ‘elephant’ vs ‘elefante’ vs ‘elefant’), further compounding this problem of synonymy. The use of unqualified acronyms, slang terms, jargon and neologisms also contribute to the synonymy problem (Spiteri, 2007).

Entropy Entropy in a folksonomy is caused by an over abundance of idiosyncratic tags which are meaningful and useful to only one user. Since the tags are less descriptive to others, search results using these tags are likely to be less useful and full of noise (Wu, Zubair, et al., 2006).

Spaces and Multiple Words Most folksonomies are designed for tags to be single words without spaces (Mathes, 2004). For example, in Delicious a tag of John Key would be regarded as two separate tags, ‘John’ and ‘Key’, since the space character is used as a separator. Users get around this limitation by removing the spaces altogether (e.g. johnkey) and using upper case letters (i.e. ‘JohnKey’) or other characters

(e.g. john-key) to distinguish words. However, the system used by an individual is not necessarily intuitive for others, nor might it be used consistently. It may therefore be difficult for the tagger or other users to identify the search term. Such practices can also compound the entropy problem.

basic level variation Basic level variation relates to where in a hierarchy an individual places a particular item, that is, where they select as the base level for classifying content (Golder & Huberman, 2006). For example, a user who wishes to tag a page about a mallard, may select the more general (superordinate) level of ‘bird’ as their basic level, while another user may select the more specific (subordinate) level of ‘duck’ as their basic level, and still another user may use the even more specific term ‘mallard’ to classify the item. Any subsequent search for ‘bird’ would not return a (different) resource that had been tagged with only ‘duck’ or ‘mallard’ even though such terms may be subsumed within the higher level term ‘bird’. Basic level variations occur for many reasons such as the level of expertise of the individual defining the tag, the degree to which the level of specificity is important to the individual, and sensemaking (Golder & Huberman, 2006).

Individual vs. Collective Motivations In the world of folksonomies, there is a fundamental tension between two different uses being made of the same systems and of tags. While some users tag for the purpose of their own information retrieval, others tag content so it can be found by others (Hammond et al., 2005). As all tagging collectively contributes to the classification system, collaborative systems will come to consist of tags whose meanings and uses are widely agreed on as well as idiosyncratic (personal use) tags (Golder & Huberman, 2006). Since systems such as Flickr

215

Web 2.0—Social Bookmarking

and Delicious are used by individuals to tag content for their personal use, such users will want to use terms that are meaningful to themselves. At the same time these systems are useful when shared. Hence, it may be that an individual’s motivation will determine whether the tags that are chosen tend to reflect the collaborative aspect of the folksonomies or are geared more towards building a personal but meaningful retrieval system. Both goals are not mutually exclusive and may be necessary for explaining folksonomies (Mathes, 2004).

that relate to information retrieval as well as those that arise in the social domain of tag use. In practice, collaborative services that make use of folksonomies (e.g., Delicious) have focused primarily on issues related to information retrieval by introducing features aimed at addressing some of the organizational weaknesses associated with folksonomies. A key feature is the tag cloud which is commonly seen alongside the folksonomy. Other initiatives include listing related tags, tag bundles, and tag descriptions.

tag Clouds Tag Spamming Like search engine spamming where content is created to mislead search engines, tagging systems are also susceptible to spam in the form of tag spamming. Tag spamming is observable when popular and sometimes non-related tags are assigned to resources – these tags are deliberately designed or chosen to attract users to the link or mislead or confuse users (Koutrika et al., 2007). Koutrika et al. identifies several examples of tag spam. For example, a malicious user may tag several photos with a particular tag so that it appears on the ‘list’ of popular tags. Others may tag content that a user does not want to view with a commonly used tag to deceive them into retrieving the item. Tag spam could also take the form of a company tagging several pages (except that of its competitor) with a tag such as ‘buy cars’ so that users cannot easily find the competitor’s site. Multiple tags might also be assigned to a particular content to increase its visibility.

Addressing the ChAllenges For folksonomies to work well and overcome limitations such as those that arise from ambiguities in language and basic level variations, a shared understanding of tags over time needs to occur. To achieve this, attention needs to be paid to issues

216

Tag clouds are weighted lists designed to provide a visual guide to user-generated tags that have been used and their relative popularity (Macgregor & McCulloch, 2006). Each tag on the tag cloud is hyperlinked to the items bookmarked with that tag. Two types of features make a tag cloud: text features and word placement (Rivadeneira et al., 2007). Text features include font weight, the degree of boldness of a font, font size and font color. Word placement features include sorting, clustering, grouping tags together to signify some meaning, and spatial layout. Tags can appear in any order (e.g. alphabetical, random, sorted by weight). One popular form of the tag cloud is an alphabetical list of tags with a font size that is proportional to the use of the tag (see Figure 3). Tag clouds can assist users with various tasks (Rivadeneira et al., 2007). For example, tag clouds can help users search for particular content by locating a desired term within the cloud. The users can also browse content by looking at the tag cloud with no particular term in mind. Users can also examine tag clouds to form an impression of the underlying data set being represented, or use a tag cloud to help recognize particular content. For example, if a user wants to identify a particular individual, they can examine the tag clouds for characteristics that may help identify the individual.

Web 2.0—Social Bookmarking

Figure 3. Tag clouds (Source: http://Delicious.com/) Reproduced with permission of Yahoo! Inc. ©2009 Yahoo! Inc. DELICIOUS and the DELICIOUS logo are registered trademarks of Yahoo! Inc.

related tags Services such as Delicious also provide a list of related tags using an algorithm that determines the other tags that most often accompany a given tag. This allows the user to browse a related tag and find bookmarks that have been tagged with a synonym or other related words. For example, if the user selects the tag ‘CSS’, Delicious will return a list of the related tags ‘webdesign’, ‘design’, ‘web’, ‘javascript’, ‘webdev’, ‘html’, ‘tutorial’, ‘tutorials’, ‘reference’, ‘inspiration’ and ‘tips’ (See Figure 4). Clicking on the + sign to the left of a related tag will display bookmarks that are tagged with both tags. For example, if a user is viewing the tag “CSS’ and clicks on the + sign next to the related tag ‘HTML’, Delicious will display bookmarks tagged with both ‘CSS’ and ‘HTML’. Where multiple tags are used to tag content, related tags can help users overcome some of the problems linked to ambiguity in language and basic level variations, by providing links to content that is associated with the search terms

being used. Where tags are unrelated to other tags, finding the content by persons other than the tagger becomes very difficult. Although the extent to which unrelated tags occur is not entirely known, it is not unheard of. For example, Tag Patterns reported over 167,800 tags being tracked (Tag Patterns, 2008)

tag bundles Tag bundles are essentially tags for tags. For example, as the number of tags used increases, users may choose to group related tags into a tag bundle. This introduces a level of hierarchy into the tag structures and can help users with organizing and searching their tags. For example, in Delicious, a user might place the tags ‘css’, ‘html’, ‘javascript’ and ‘ajax’ in a bundle called ‘web’ (See Figure 5).

tag descriptions Tag descriptions also try to resolve problems of tag ambiguity (e.g. polysemy, homonymy) by

217

Web 2.0—Social Bookmarking

Figure 4. Related tags for ‘CSS’ (Source: http://Delicious.com/) Reproduced with permission of Yahoo! Inc. ©2009 Yahoo! Inc. DELICIOUS and the DELICIOUS logo are registered trademarks of Yahoo! Inc.

Figure 5. Tag Bundle for ‘web’ (Source: http:// Delicious.com/) Reproduced with permission of Yahoo! Inc. ©2009 Yahoo! Inc. DELICIOUS and the DELICIOUS logo are registered trademarks of Yahoo! Inc.

including a description of how the user uses the tag in question (See Figure 6). Although features such as tag clouds, related tags and the availability of immediate feedback when a tag is assigned can help alleviate problems such as synonymy, unless taggers can agree on a common set of search terms such problems are likely to remain significant.

information retrieval

CUrrent reseArCh In academic circles, researchers are attending to the issues and challenges of social tagging. Of the two streams – information retrieval and social computing – work related to information retrieval, ontologies, and tag spamming is receiving the far greater attention (Gruber, 2007; Hassan-Montero & Herrero-Solana, 2006; Hotho et al., 2006a; 2006b; Koutrika et al., 2007; Krause et al., 2008; Xu et al., 2006).

218

In information retrieval, there is a stream of research focused on the structure and form of folksonomies particularly as this relates to language. For example, Spiteri (2007) examined tag structures observing that while tags largely corresponded with standards for controlled language (e.g. use of single nouns, alphabetic characters and recognized spelling) there were potential problems pertaining to the use of singular versus plural terms, multi-term tagging, unqualified abbreviations/acronyms, and count nouns. Guy and Tonkin (2006) also examined the form of tags focusing on the tag literacy of users – their findings showed that a high percentage of tags were misspelt, not in a form that could be decoded by their multilingual dictionary, or were composed of multiple words, or a combination of languages. They therefore suggested a number of mechanisms aimed at improving tag quality (e.g. spell checking, suggesting synonyms, etc.). To address

Web 2.0—Social Bookmarking

Figure 6. Description of ‘CSS’(Source: http:// Delicious.com/) Reproduced with permission of Yahoo! Inc. ©2009 Yahoo! Inc. DELICIOUS and the DELICIOUS logo are registered trademarks of Yahoo! Inc.

the issues that arise with tag ambiguity, HassanMontero and Herrero-Solana (2006) suggested an alternative tag cloud that uses an algorithm to strip away highly idiosyncratic tags (such as ‘toread’ and ‘cool’) and obvious synonyms. The remaining tags are then presented with related tags clustered together, whilst retaining the convention that more common tags have a larger font size. Xu, et al., (2006) also explored a number of tactics aimed at improving the efficacy of tagging. This included setting criteria that define good tagging systems, including algorithms for tag suggestions, and introducing authority (or reputation) scores to combat spam. Other researchers have also explored the value and usefulness of folksonomies for searching and organising resources (Damianos et al., 2007; Heymann et al., 2008; Sinclair & Cardew-Hall, 2008). For example, Sinclair and Cardew-Hall (2008) looked at the usefulness of tag clouds for information-seeking. Their study showed tag clouds were preferred when the information-seeking task was more general while search interfaces were preferred when the information-seeking task was more specific. Heymann et al. (2008) looked at whether data provided through social bookmarking can enhance web searching. After examining the characteristics of over 40 million bookmarks on Delicious, they concluded that social bookmarking can yield some search data not currently provided by other sources. Morrison (2008) examined the efficacy of folksonomy-based retrieval systems compared with search engines and subject directories, concluding that folksonomies have the potential to improve the performance of web searches. Damianos et al. (2007) explored

the adoption and usefulness of folksonomies in corporate settings. They concluded that while folksonomies could not replace more formal structures, they were useful for organizing and making transparent various repositories and collaborative spaces located in corporate intranets. The benefits to organizations included information sharing and dissemination, encouraging information discovery, supporting communities and social networks and locating experts.

ontologies and folksonomies The topic of ontologies is also at the forefront of research on folksonomies (Gruber, 2007; Heymann & Garcia-Molina, 2006; Mika, 2007; Van Damme, et al., 2007). Unlike traditional ways of organizing information objects which rely on welldefined pre-specified classification systems (e.g. use of simple controlled vocabularies, taxonomies, or fully developed ontologies), folksonomies tend to have flatter, organic structures which emerge (and change) over time as individuals manage their information needs (Hotho et al., 2006a). The uncontrolled nature of folksonomies often results in inconsistencies and redundancies that complicate information retrieval. Yet, research suggests that these interactions and uses of folksonomies can be leveraged to generate and maintain ontologies and other useful metadata (Gruber, 2007; Van Damme et al 2007). For example, Mika (2007) devised and used a tripartite Actor-Concept-Instance model to determine the ontology of a set of Delicious tag clusters. Focusing on structures and hierarchies, Heymann and Garcia-Molina (2006) proposed an algorithm that

219

Web 2.0—Social Bookmarking

converts tags into a navigable hierarchy of tags. Van Damme et al. (2007) advocated an approach that combines lexical, semantic and social data to derive and maintain actual ontologies from folksonomies while Gruber (2007) explored the development of a “common ontology of tagging” (i.e. TagOntology).

tag spamming Finally, although widely publicized occurrences of tag spamming are few, as tagging becomes more prevalent tag spam could become a serious problem for service providers and users. Researchers are therefore exploring methods to counter such occurrences (Koutrika et al., 2007; Xu et al., 2006). For example, using an experimental framework for modeling tagging systems and tagging behaviors, Xu et al. (2006) proposed a set of countermeasures based on reputation scores to address spamming behavior. On the other hand, Koutrika et al., (2007) devised and assessed the impact of a proposed ranking algorithm on tag spam. Service providers such as Flickr and Delicious are also taking steps to mitigate tag spam, but in some cases these actions could negatively impact legitimate users. For example, users trying to circumnavigate the issues associated with synonymy may assign multiple tags to their content only to be caught out by various anti-spamming mechanisms.

fUtUre reseArCh The examples from practice and research suggest that while some of the challenges in folksonomies are being addressed, there remains a vast untapped opportunity for further work. For example, there is still debate in the popular press as to whether folksonomies really do work. User-driven tagging has the advantage of giving users the power to organize content in a way that is meaningful to them. It can also improve search accuracy

220

by making more visible, content that would not otherwise be found. User-driven tagging might therefore be one way of overcoming the search issues that are typically associated with amateurcreated content and information-only sites. On the other hand, the unfettered use of user-developed tags has the potential to exacerbate attempts at organizing and locating content aimed at communal usage. Not only are different users likely to use different tags, there are issues with the misuse of tags. These situations therefore provide opportunities for further research in information retrieval, computer-mediated collaboration, and social computing. Folksonomies permit the sharing of resources in a social setting and typically comprise three elements: resources, tags, and users (Hotho et al., 2006b; Marlow et al., 2006). Although each of these areas has been studied independently, Marlow et al. (2006) suggest that a unified usertag-resource approach might provide insights into areas such as information retrieval, organization and discovery, spam filtering, and identifying emerging trends and topics within communities. Given the placement of folksonomies at the boundary of information retrieval and social computing this tripartite approach to studying folksonomies makes sense, and underpins the suggestions that follow. In the area of information retrieval, researchers should continue to strive for better ways of managing and navigating folksonomies. For example, since the wisdom of crowds is already successful in folksonomies for tagging content it is likely that it might also be successful in identifying synonyms. Hence, one possible avenue is to investigate systems will permit users to suggest synonyms that can be bundled together. Improvements can also be made by using systems that provide simple error-checking or that provide tag suggestions (Guy & Tonkin, 2006). There are also opportunities for future research to address more effective ways of organizing folksonomies, and to explore and extract their underlying structure to

Web 2.0—Social Bookmarking

enable higher quality searches. The development and use of trust or reputation scores may also help address anti-social behaviors and improve the quality of folksonomies through community self-regulation. While the aspect of information retrieval is receiving attention in academic circles as evidenced by the growing number of conference papers in the subject area, far less attention has been paid to the user-side of folksonomies. Again bearing in mind the tripartite relationship between user, resources and tag (Marlow et al., 2006), there are numerous research opportunities related to computer-mediated collaborative systems as well social computing to explore the motivations and behaviors of individuals and communities using these systems. Such research is likely to draw on and engage multiple disciplines including information systems, sociology, marketing, psychology, library science and organization science. For example, future research might examine how individuals derive and use folksonomies, and how communities act to form and regulate folksonomies (e.g. to elevate the use of particular terms and shut out or marginalize other terms). Indeed over time, certain aspects of a folksonomy can begin to exhibit a certain degree of stability. Golder and Huberman (2006) suggest this may be attributed to factors such as imitation behaviors or the emergence of shared knowledge, but these conjectures are yet to be explored empirically. At the individual level, researchers can also examine the factors that influence people to engage in these communities of knowledge. More specifically future research can explore the factors that impact the use and diffusion of folksonomies, as the more people use these social resource sharing systems, the greater the value of the system itself and the greater the benefit to those engaged with the system. Examining the motivations and incentives of users is another rich area for future research (Marlow et al., 2006; Paolillo & Penumarthy, 2007). Here researchers might consider distinguishing different roles such as the author/

creator roles from other users, or the impact of system attributes (e.g. tagging rights, tagging support, social connectivity etc) on tagging behaviors. Examining the vocabulary used and the materials aligned with these tags can also provide unique insights into sensemaking and the development of the language of a folksonomy as well as insights into the interests and thinking of individuals and communities. Finally, reference to the domain of social computing (Parameswaran & Whinston, 2007) will also reveal research directions and possible theories that can be linked to the domain of folksonomies. For example, the tension between individualism and personal expression and philanthropic activities designed to elevate online sharing and its social benefits are likely to offer rich insights into the formation and behaviors of social systems. Several theoretical frameworks may also provide useful perspectives on folksonomies. These include social capital theory, social network theory, diffusion of innovations theory, theory of planned behavior, theory of sensemaking, actor-network theory, and motivation theories such as incentive theory and affective-arousal theory.

ConClUsion In summary, a folksonomy constitutes a practice and method of collaboratively creating and managing tags to annotate and categorize content and build collective knowledge. Folksonomies have several advantages over hierarchical categorization systems such as taxonomies including low barriers to participation, providing feedback and immediate benefit to users, and making information retrieval easier. On the other hand, several issues also arise. For example, folksonomies cannot distinguish between polysemes, are blind to synonyms, contain basic level variation between users, and suffer from entropy and noise. There is also a fundamental tension since users have both highly individual and highly social uses for

221

Web 2.0—Social Bookmarking

the folksonomy. This chapter therefore explores a number of initiatives in practice and in research to address some of the ambiguities and organizational issues that arise with folksonomies. The study also suggests avenues for future research as these relate to information retrieval, computer-mediated collaboration and social computing.

referenCes Auchard, E. (2007). Flickr to map the world’s latest photo. Retrieved on September 14, 2008, from http://www.reuters.com/articlePrint?article Id=USHO94233920071119 Connotea. (2008). Connotea beginner’s guide. Retrieved on September 13, 2008, from http:// www.connotea.org/guide Damianos, L. E., Cuomo, D., Griffith, J., Hirst, D. M., & Smallwood, J. (2007). Exploring the adoption, utility, and social influences of social bookmarking in a corporate environment. Paper presented at the Hawaii International Conference on System Sciences, Waikoloa, HI. Golder, S., & Huberman, B. A. (2006). Usage patterns of collaborative tagging systems. Journal of Information Science, 32(2), 198–208. doi:10.1177/0165551506062337 Gruber, T. (2007). Ontology of folksonomy: A mash-up of apples and oranges. International Journal on Semantic Web and Information Systems, 3(1), 1–11. Guy, M., & Tonkin, E. (2006). Folksonomies: Tidying up tags. D-Lib Magazine, 12(1). Retrieved on September 13, 2008, from http://www.dlib.org/ dlib/january06/guy/01guy.html Hammond, T., Hannay, T., Lund, B., & Scott, J. (2005). Social bookmarking tools (I). D-Lib Magazine, 11(4). Retrieved on September 13, 2008, from http://www.dlib.org/dlib/april05/ hammond/04hammond.html

222

Hassan-Montero, Y., & Herrero-Solana, V. (2006). Improving tag-clouds as visual information retrieval interfaces. Paper presented at the International Conference on Multidisciplinary Information Sciences and Technologies, Barcelona, Spain. Heymann, P., & Garcia-Molina, H. (2006). Collaborative creation of communal hierarchical taxonomies in social tagging systems. (Tech. Rep. 2006-10). Retrieved on September 13, 2008, from http://dbpubs.stanford.edu:8090/pub/2006-10 Heymann, P., Koutrika, G., & Garcia-Molina, H. (2008). Can social bookmarking improve Web search? Paper presented at the International Conference on Web Search and Web Data Mining, Palo Alto, CA. Hotho, A., Jaschke, R., Schmitz, C., & Stumme, G. (2006a). Information retrieval in folksonomies: Search and ranking. In Y. Sure & J. Domingue (Eds.), The Semantic Web: Research and applications, vol. 4011/2006. European Semantic Web Conference (pp. 411-426). Heidelberg: Springer-Verlag. Retrieved on September 13, 2008, from http://www.springerlink.com/content/ r8313654k80v7231/ Hotho, A., Jaschke, R., Schmitz, C., & Stumme, G. (2006b). Trend detection in folksonomies. In Y. S. Avrithis, Y. Kompatsiaris, S. Staab & N. E. O’Connor (Eds.), Semantic multimedia, vol. 4306/2006. International Conference on Semantics and Digital Media Technologies (pp. 56-70). Heidelberg: Springer-Verlag. Retrieved on September 13, 2008, from http://www.springerlink. com/content/mgx4406381n10668/ Koutrika, G., Effendi, A., Gyongyi, Z., Heymann, P., & Garcia-Molina, H. (2007). Combating spam in tagging systems. Paper presented at the International Workshop on Adversarial Information Retrieval on the Web, Alberta, Canada.

Web 2.0—Social Bookmarking

Krause, B., Schmitz, C., Hotho, A., & Stumme, G. (2008). The anti-social tagger–detecting spam in social bookmarking systems. Paper presented at the Workshop on Adversarial Information Retrieval on the Web, Beijing, China. Macgregor, G., & McCulloch, E. (2006). Collaborative tagging as a knowledge organisation and resource discovery tool. Library Review, 55(5), 291–300. doi:10.1108/00242530610667558 Marlow, C., Naaman, M., Boyd, D., & Davis, M. (2006). HT06, tagging paper, taxonomy, Flickr, academic article, to read. Paper presented at the Conference on Hypertext and Hypermedia, Odense, Denmark. Mathes, A. (2004). Folksonomies-cooperative classification and communication through shared metadata. Retrieved on March 10, 2008, from http://www.adammathes.com/academic/computermediated-communication/folksonomies.html Merholz, P. (2004). Metadata for the masses. Retrieved on September 13, 2008, from http://www. adaptivepath.com/ideas/essays/archives/000361. php Mika, P. (2007). Ontologies are us: A unified model of social networks and semantics. Web Semantics: Science . Services and Agents on the World Wide Web, 5(1), 5–15. doi:10.1016/j.websem.2006.11.002 Millen, D., Feinberg, J., & Kerr, B. (2005). Social bookmarking in the enterprise. Social Computing, 3(9), 1–7. Morrison, P. J. (2008). Tagging and searching: Search retrieval effectiveness of folksonomies on the World Wide Web. Information Processing & Management, 44(4), 1562–1579. doi:10.1016/j. ipm.2007.12.010 Paolillo, J. C., & Penumarthy, S. (2007). The social structure of tagging Internet video on del.icio.us. Paper presented at the Hawaii International Conference on System Sciences, Waikoloa, HI.

Parameswaran, M., & Whinston, A. B. (2007). Research issues in social computing. Journal of the Association for Information Systems, 8(6), 336–350. Rattenbury, T., Good, N., & Naaman, M. (2007). Towards automatic extraction of event and place semantics from Flickr tags. Paper presented at the International ACM SIGIR Conference on Research and Development in Information Retrieval, Amsterdam, The Netherlands. Rivadeneira, A. W., Gruen, D. M., Muller, M. J., & Millen, D. R. (2007). Getting our head in the clouds: Toward evaluation studies of tagclouds. Paper presented at the SIGCHI Conference on Human Factors in Computing Systems, San Jose, CA. Sinclair, J., & Cardew-Hall, M. (2008). The folksonomy tag cloud: When is it useful? Journal of Information Science, 34(1), 15–29. doi:10.1177/0165551506078083 Spiteri, L. F. (2007). The structure and form of folksonomy tags: The road to the public library catalog. Information Technology and Libraries, 26(3), 13–25. Van Damme, C., Hepp, M., & Siorpaes, K. (2007). FolksOntology: An integrated approach for turning folksonomies into ontologies. European Semantic Web Conference: Bridging the Gap between Semantic Web and Web 2.0 (pp.71-84), Innsbruck, Austria. Retrieved on September 13, 2008, from http://www.heppnetz.de/files/vandammeheppsiorpaes-folksontology-semnet2007-crc.pdf Vander Wal, T. (2007). Folksonomy coinage and definition. Retrieved on September 9, 2008, from http://vanderwal.net/folksonomy.html Wu, H., Zubair, M., & Maly, K. (2006). Harvesting social knowledge from folksonomies. Paper presented at the Conference on Hypertext and Hypermedia, Odense, Denmark.

223

Web 2.0—Social Bookmarking

Wu, X., Zhang, L., & Yu, Y. (2006). Exploring social annotations for the Semantic Web. Paper presented at the International Conference on World Wide Web Conference, Edinburgh, Scotland. Xu, Z., Fu, Y., Mao, J., & Su, D. (2006). Towards the Semantic Web: Collaborative tag suggestions. Paper presented at the Collaborative Web Tagging Workshop at WWW2006, Edinburgh, Scotland.

Key terMs And definitions Basic Level Variation: Terms that describe an item vary along a continuum ranging from the very specific to the very general. For example, a particular sea-creature could be described by the very specific term “Hammerhead”, or the very general term “Fish”, or the intermediate term “Shark” Folksonomy: Folksonomy is the result of personal free tagging of information and objects (i.e. anything with a URL) for one’s own retrieval. The

224

tagging is done in a social environment (usually shared and open to others). Folksonomy is created from the act of tagging by the person consuming the information. (Vander Wal, 2007) Polyseme: A word or phrase with multiple, related meanings. For example, window can refer to both a hole in a wall that allows light in and a pane of glass filling such a hole Social Computing: An area of information technology that is concerned with the intersection of social behavior and computational systems. Synonym: Different words with identical or at least similar meanings. For example, lorry and truck Tag: A tag is a keyword assigned to a piece of information (e.g. a website, a picture, or video clip), describing the item and enabling keywordbased classification and search of information. A type of metadata Tag Cloud: A visual depiction of content tags used on a website. Tags are typically listed alphabetically, and tag frequency is shown with font size or color

225

Chapter 13

Social Semantic Bookmarking with SOBOLEO Valentin Zacharias FZI Research Center for Information Technology, Germany Simone Braun FZI Research Center for Information Technology, Germany Andreas Schmidt FZI Research Center for Information Technology, Germany

AbstrACt The novel paradigm of social semantic bookmarking combines the positive aspects of semantic annotation with those of social bookmarking and tagging while avoiding their respective drawbacks; drawbacks such as the lacking semantic precision of tags or the cumbersome maintenance of ontologies. Social semantic bookmarking tools allow for the annotation of Internet resources based on an ontology and the integrated maintenance of the ontology by the same people that use it. This chapter motivates social semantic bookmarking by examining the respective problems of tag based bookmarking and semantic annotation. Social semantic bookmarking is then introduced and explained using the SOBOLEO application as an example. It also gives an overview of existing applications implementing this new paradigm and makes predictions about its movement into the mainstream and remaining research challenges.

introdUCtion An important challenge for today’s internet users is the recovery of internet resources that they had once found interesting and useful; as well as the discovery of new interesting information. Social bookmarking systems (such as delicious1) can aid in these tasks by supporting users in the collection, management and sharing of bookmarks; DOI: 10.4018/978-1-60566-384-5.ch013

i.e. references to such resources and information. For organization, navigations and searching these systems utilize tags. Tags are arbitrary keywords that are used by the users to further describe the internet resources in order to aid their retrieval. Tags are renowned for their flexibility and ease of use, because just any tag can be used and there is no overhead for vocabulary management. However, this missing structure is also the root cause for a number of problems plaguing tagging and hampering tag-based retrieval:

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Social Semantic Bookmarking with SOBOLEO

problems such as typos, tags on different levels of abstraction, or synonyms. Replacing tags with annotations based on a controlled vocabulary or ontology can help alleviate these problems. Systems that use ontologies as source for annotating internet resources are, however, also not without their problems. For one they are often cumbersome to use; but more importantly, they view ontology creation as a process separate from its use; a process performed by people different from those that use it. Another problem is that these systems often assume that the ontology stays unchanged for prolonged periods of time and requires only occasional updates. All this leads to unsatisfied users being confronted with out-of-date, incomplete, inaccurate and incomprehensive ontologies that they cannot easily use for annotation; this problem is particular acute in fast changing domains (Hepp, 2007). The novel paradigm of Social Semantic Bookmarking combines the positive aspects of semantic annotation with those of social bookmarking while avoiding their respective drawbacks. Social semantic bookmarking tools allow for the annotation of internet resources with respect to an ontology and the integrated maintenance of the ontology by the same people that use it. Through the use of stateof-the-art web technologies such as bookmarklets and AJAX (e.g., for autocomplete functionality), these systems make ontology-based annotation of web documents as simple as tagging. Through easy-to-use, lightweight web ontology editors that are integrated into the system, the barrier between ontology creation and use is removed; users who annotate with the help of the ontology are the same who continuously evolve this ontology. Because internet resources are annotated with concepts (and not keywords), the problems of homonyms, synonyms etc. are avoided. We present Social Semantic Bookmarking using the example of our system SOBOLEO (SOcial BOokmarking and Lightweight Engineering of Ontologies) – a system combining the above mentioned features with an innovative

226

search engine and functionality supporting the discovery of experts on specific topics based on their interaction with the system. We also shortly discuss other social semantic bookmarking systems such as Bibsonomy, int.ere.st, GroupMe!, Fuzzy, and Annotea. Finally, we sketch the trends that shape the future of social bookmarking – one of the most visible and best known developments of the Web 2.0 world.

bACKgroUnd: (lingUistiC) tAgging vs. seMAntiC AnnotAtion (linguistic) tagging and its Problems Social bookmarking systems allow their users to annotate bookmarks with several arbitrary tags they find most suitable for describing them. In this way – in contrast to the traditional folder structure like browser favorites – users can organize their bookmarks according to more than one category. This facilitates the organization, navigation and search in the bookmark collection. These systems make collecting bookmarks a social experience by allowing the users to share their bookmarks with others. Furthermore, not only are the bookmarks visible to other users but also the tags used to describe them. That means, you can share your own tags and use the other users’ ones. You can see which tags and annotated resources you have in common with other users or what they annotated with the same tags. In this way, you can find people with similar interests and discover new interesting resources. Social bookmarking systems give users the possibility to have their own view on the resources and to express their opinion or present themselves without any restriction (cf. Marlow 2006). The users do not have to learn complex and predefined schemata or syntax, and problems of controlled vocabularies can be avoided (cf. Mcgregor 2006).

Social Semantic Bookmarking with SOBOLEO

At first glance, this seems to result in a chaotic collection of bookmarks. However, first studies have shown that among the users a common understanding and vocabulary, also known as folksonomy, emerges from the tags and the process of tagging (cf. Golder 2006, Marlow 2006, Sen 2006). Features increasing awareness for already used tags like tag clouds or tag recommendation further support this effect positively. However, folksonomies have only very limited structure. Their missing semantic precision hampers efficient search and retrieval support, in particular in complex domains, because of problems like the following (cf. Golder 2006, Guy 2006): •







(Mis-)Spelling: The most obvious problem is that tags are simply misspelled or written in different ways because of occurring plurals, abbreviations or compound words, e.g. ‘spagetti’ vs. ‘spaghetti’, ‘noodle’ vs. ‘noodles’, or ‘spaghettiCarbonara’ vs. ‘spaghetti_carbonara’. Multilingualism: Tags only relate to one language. That means, especially in Europe with many different languages, users have to annotate a resource with many tags in different languages, e.g. with ‘pasta’, ‘noodles’, and ‘Nudeln’, in order to ensure that other users will ind it later on (e.g. to promote their own great spaghetti recipe). Polysemy: Tags can have several similar meanings. This leads to search results with low precision because of irrelevant resources; e.g. with the tag ‘pasta’ the users can think of a dish that contains pasta as its main ingredient or of the aliment itself as shaped and dried dough made from lour and water and sometimes egg. Homonymy: The problem of homonymy is comparable to the problem of polysemy. However, in this case, one tag can have several totally different meanings. This also





leads to irrelevant results as all resources that relate to these different meanings are annotated with the same tag. For instance the word ‘noodle’ can have the meaning of an aliment but also of a swearword for a human head. Synonymy: Resources are not found because they are annotated with another tag with the same meaning, e.g. with the tag ‘vermicellini’ instead of ‘spaghettoni’. Similar to mulitlingualism, the users have to annotate the resources with many synonymous tags in order to ensure the retrieval by other users. Mismatch of abstraction level: Also a typical search problem emerges because tags are speciied on different abstraction levels, i.e. either too broad or too narrow. This problem, also known as the “basic level phenomenon” (Tanaka 1991), can be traced back to different intentions and expertise levels of the users. For instance, one user tags a resource on the basic level with ‘spaghetti’, another with ‘noodles’ and a third differentiates ‘spaghetti’ from ‘bigoli’ (thicker spaghetti) and ‘vermicelli’ (thinner spaghetti). A resource annotated with ‘spaghetti’, however, cannot be found with the search term ‘pasta’.

semantic Annotation and its Problems These problems of (linguistic) tagging-based approaches can be addressed by relying on controlled vocabularies. These approaches restrict the index terms that can be assigned to information resources to a controlled set of terms. With these approaches users cannot use just any term to annotate a resource, but are restricted to the controlled vocabulary. Using controlled vocabularies has a number of advantages (Macgregor 2006):

227

Social Semantic Bookmarking with SOBOLEO













It controls the use of (near-) synonyms by establishing the term that is to be used to represent a word. It discriminates between homonyms, i.e. it enforces that every term used has only one well deined meaning. It controls lexical anomalies such as grammatical variations or the use of terms without relevance for the information retrieval task (such as leading articles or prepositions) A structured vocabulary also facilitates the use of codes and notations that are mnemonic, predictable and language independent. In physical environments a controlled vocabulary facilitates the iling, storage and organization of resources. It may point the user to closely related, more suitable terms by indicating the presence of broader, narrower or related terms.

Recent years saw the rise of semantic annotation approaches that rely on a semantically described controlled vocabulary, i.e. instead of terms in a controlled vocabulary these approaches use concepts whose relations are represented in some machine understandable form. These approaches also rely on the use of some standardized formal language for representing the ontology, such as RDF (Manola 2004), SKOS (Miles 2008) or one of the OWL languages (Dean 2004). These semantic annotation approaches have a number of potential benefits: •



228

Better Retrieval: The formally represented relations between the concepts in the ontology can be used to offer superior browse or query facilities. In the case where a powerful language like OWL is used, queries may even be answered using reasoning algorithms. Better Use of Annotation: The availability of machine understandable context for







the used annotation terms can be utilized to make better use of the annotation; e.g. information that some annotations represent geographic locations for which a latitude and longitude is known can be used to show the annotated document in a map or to make them available based on the users current location. Better Quality Assurance: The information contained in the ontology about concepts used for annotation can enable checks on whether an annotation is likely to make sense; this can help to catch errors early. Also changes in the ontology can be checked whether they violate its integrity. Better (Semantic Web) Integration: The ontology that is used in the annotation is usually assumed to be also used in other systems and the common usage of the ontology can enable the integration of data created and managed in these diverse systems. Another related aspect is, that semantically annotated data can become part of the Semantic Web and then Semantic Web aware agents and applications can make use of it. Better Support of Vocabulary Management: Through the use of standardized languages to represent the ontologies, these approaches can rely on a landscape of tools that is available to create, manage and evolve these ontologies.

Semantic annotation approaches can be roughly split into two categories: on the one hand approaches that mostly rely on automatic annotation and on the other hand those that rely on manual annotation. Automatic approaches use machine learning techniques to automatically create annotations for documents based on a training set. The best known examples for the mostly automatic approach are the KIM platform (Popov, 2003), SemTag/Seeker (Dill, 2003) and MnM (Vargas-Vega, 2002). Manual approaches

Social Semantic Bookmarking with SOBOLEO

support a user in creating annotations with respect to an ontology, the most famous of which are the Ont-O-Mat system (Handschuh, 2002) and Annotea (Koivunen, 2006). Semantic annotation approaches, however, have not found widespread adoption yet; to a large extend because of their inherent limitation rooted in their perspective on the annotation process (i.e., the use of the ontology) and the creation of the ontology as two separate processes, performed by a different set of people and the latter being done by dedicated knowledge engineering specialists. However, separating the use and the creation of the ontology and involving ontology engineering specialists is causing a number of problems: •





High Cost: Knowledge engineers are highly paid specialists, and their effort comprises not only the actual implementation of the domain ontology, but also learning about and understanding the domain of interest. While in many Web 2.0 scenarios a large amount of work is done for free by users interested in the result, this is unlikely to work when knowledge engineers with little innate interest in the domain in question are involved. Domain Errors: Knowledge engineers are specialists for the domain of knowledge formalization – not for the domain that is being formalized. For this reason they will not have an understanding of the domain comparable to that of domain experts, this limited understanding may cause errors in the resulting ontology (Barker 2004). Heavyweight Process and Upfront Investment: Because annotation cannot start without an available ontology, there needs to be an upfront investment to inance the development of this ontology, which includes a systematic requirements elicitation phase. During the usage phase of the ontology, there also needs to be a accompanying process to collect newly emerging





requirements, bugs and other change requests and to implement them into a newer version of the ontology. High Time Lag: There will always be some time lag between the emergence of a new concept and the time when it is included in the ontology and can eventually be used. This time lag is relatively large, when the users of the ontology cannot make the change themselves but must rely on knowledge engineers understanding the requirement, implementing it and inally rolling out the new version of the ontology. In fast moving domains this time lag can quickly get so big that the ontology as a whole becomes unusable (Hepp 2007). Low Appropriateness and Understandability: An ontology is appropriate for a task if it enables the users to reach their goals more quickly. However, having different people using and developing the ontology makes reaching appropriateness of the ontology much harder. A particular challenge is to ensure that the ontology is at the right level of abstraction to be understood by the domain experts.

soCiAl seMAntiC booKMArKing In the previous sections we have seen that (linguistic) tagging approaches, while popular, struggle with problems such as polysemy, multilingualism or abstraction level mismatches. At the other end many current semantic annotation approaches struggle (like most approaches build on controlled vocabularies of some kind) with the problem of timely updates and appropriateness of the controlled vocabulary as well as affordable creation. Social Semantic Bookmarking now combines the benefits of tagging with semantic annotation in order to address their respective weaknesses. Social semantic bookmarking systems allow for the annotation of resources (e.g. web pages,

229

Social Semantic Bookmarking with SOBOLEO

documents) with concepts whose definition and description also evolves collaboratively within the same system. Similar to tagging approaches, they allow for creating new concepts whenever a need arises. Unlike these approaches, concepts can have powerful descriptions and can be interlinked; for example allowing the system to understand that ‘swimming bath’ and ‘swimming pool’ are synonyms for the same concept. These powerful concept descriptions are similar to those used in traditional semantic annotation, but social semantic bookmarking allows for adding and changing concepts permanently and easily at the time the concepts are used. The SOBOLEO2 system (Zacharias 2007) is a particular social semantic bookmarking system that will be used to further illustrate this approach in this section. SOBOLEO is based on AJAX technology and works in most current browsers – thus does not require any local installation. It consists of four application parts: an editor for the modification of the shared ontology, a tool for the annotation of internet resources, a semantic search engine for the annotated internet resources and an ontology browser for navigating the ontology and the bookmark collection. SOBOLEO’s functionality and the concept of social semantic bookmarking will be further described with an example of a user who annotates an internet resource with a new concept ‘sphaghetti’, then adds some information about this new concept. A different user will then search for the annotated resource at a different level of abstraction and find it using the semantic search feature.

Annotation The annotation process starts when a user finds an interesting resource that he/she wants to add to the shared repository. In this example a user discovers a tasty pasta recipe. In order to annotate the document the user clicks on a bookmarklet in

230

his/her browser, which opens up the small dialog window shown in Figure 1. The user can annotate the web document using any of the concepts already known to the system and is supported by auto completion in doing that. Here the user also adds a new concept named ‘Spaghetti’ – adding a concept to the ontology is seamlessly done by simply entering a term that is not yet known to the system. Once the user clicks save, the system stores the URL of the document with all assigned concepts, any new concepts are also added to the shared ontology of the repository. In this way, the users themselves can incrementally extend and immediately adapt the ontology to their actual use. Besides, the SOBOLEO system crawls the content of the annotated web page that is added to a full text index associated with the repository.

ontology editing Each user of SOBOLEO belongs to a user group that has a shared repository containing the annotations and the ontology. Such a user group consists of people working on the same topic, such as a department in a large company or a special interest group/community spanning continents. The ontology in the shared repository is represented using a subset of the SKOS standard; it allows for concepts with a preferred label, a description, and any number of alternative labels. It also allows for the following relations between concepts: broader concept, narrower concept, and related concept. The ontology in this shared repository is edited using the AJAX editor shown in Figure 2. The editor is a collaborative realtime AJAX editor; i.e., it can be used by multiple persons simultaneously in their respective browsers with edits showing up for the others in realtime. In the example the user opens the editor to add more information about the new ‘Spaghetti’ concept that was automatically added when the user saved his/her annotation in the previous step.

Social Semantic Bookmarking with SOBOLEO

Figure 1. Annotating a webpage with concepts from the ontology and new terms

Such automatically added concepts are collected under the special concept called “prototypical concepts” where the users can consolidate and place them within the ontology later. Therefore, the user first uses the mouse to drag the ‘Spaghetti’ concept onto the ‘Pasta’ concept, quickly establishing the relation that ‘Spaghetti’ is a narrower concept than ‘Pasta’. He/she also adds a short description to ‘Sphaghetti’ and ‘Spaghetto’ as synonym. These modifications to the ontology are immediately visible and effective for the whole system; e.g. for the auto complete support for the annotation or the semantic search (see below).

browsing the repository Browsing the repository is the most common approach to retrieving information from a shared repository. With a browsing interface users can navigate to the concepts they are interested in and see the resources annotated with these. The browser interface also gives the chance to change any of the annotations. In SOBOLEO and social semantic bookmarking the user can also see the ontology and use the taxonomic structure for navigation. Figure 3 shows the browsing interface for the new ‘Spaghetti’ concept. The interface shows the concept name, its different alternative labels, and its description. Also shown are the most recently annotated documents (with links to

231

Social Semantic Bookmarking with SOBOLEO

Figure 2. Collaborative realtime ontology editing

change the annotation) and the relations to other concepts allowing for navigating there.

semantic search In addition to the browse interface the ontology is also used to enable semantic search. The semantic search in SOBOLEO combines semantic search utilizing the concept labels and their broadernarrower relations with a full text search over all annotated resources. The semantic search engine also offers query refinement and relaxation functionality. In the example, a different user is interested in finding a recipe including noodles, garlic and basil and enters these words as search term. The

232

semantic search recognizes that ‘noodles’ is a synonym for pasta and that spaghetti is a special kind of pasta. The search engine further finds that garlic refers to another concept and then that the annotation described earlier combines not only spaghetti and pasta as annotation but also includes basil in the sites content – hence this page is returned as a first result. The result is shown in Figure 4. Please note that neither a full text engine (because ‘noodles’ is not written on the page), nor a social tagging system (because neither noodles nor basil is a tag), nor a pure semantic search engine (because basil is not annotated) could make a comparable ranking of the result.

Social Semantic Bookmarking with SOBOLEO

Figure 3. Browsing interface for navigating to concepts and annotated resources

Advantages and Challenges of social semantic bookmarking After the description of the SOBOLEO system we can now revisit the respective weaknesses of semantic annotation and tagging to show how they are tackled by the novel paradigm of Social Semantic Bookmarking. Five problems were identified for linguistic tagging and are tackled by social semantic bookmarking in the following way: •

(Mis-)Spelling: The ontology controls lexical anomalies such as grammatical variations or the use of terms without relevance for the information retrieval task (e.g. by recording ‘spagetti’ as an alternative label of ‘spaghetti’). It also offers a tool to rein







in misspelling e.g. through auto-complete or tag recommendation with respect to the ontology. Multilingualism: Concepts in a Social Semantic Bookmarking system can have names in multiple languages. Each user can then see the name in his/her language, but all still refer to the same concept. Polysemy: Concepts are independent of their names and multiple concepts can have the same name. There can be two concepts both with the name pasta, one referring to pasta-ingredient and one to pasta-recipe. Homonymy: Again different concepts can be used to represent the different meanings of a word. In search, for example, the system can then use this to ask a disambiguation question to the user.

233

Social Semantic Bookmarking with SOBOLEO

Figure 4. Result of the semantic search





Synonymy: Concepts can have multiple names; it’s no problem, that a concept has the names of ‘vermicellini’ and ‘spaghettoni’. Mismatch of abstraction level: The ontology mediates between different levels of abstraction. In the example above the page was annotated with ‘spaghetti’, while the user searched with the more general ‘noodles’ – the background knowledge that spaghetti is a kind of noodle was used to still retrieve this result.

The problems of most existing semantic annotation approaches, stemming from the separation of creation and use of the ontology are also avoided by social semantic bookmarking. The following problems were identified above: •

234

High Cost of Knowledge Engineers: The ontology is maintained by the users, no knowledge engineers are needed or their role is greatly diminished.









Domain Errors: Errors due to a lack of domain knowledge by the knowledge engineers also do not affect social semantic bookmarking, because domain experts make the changes themselves. Heavyweight Process and Upfront Investment: The ontology is created incrementally during the use of the system; no up-front investment in knowledge engineering is needed. High Time Lag: Changes to the ontology can be done immediately and are also immediately visible during the use of the system. Low Appropriateness and Understandability: The users of the system can permanently and immediately adapt the ontology to their use.

At the same time, however, it should not be concealed that social semantic bookmarking also introduces one new challenge that is not faced by linguistic tagging or previous annotation

Social Semantic Bookmarking with SOBOLEO

approaches: domain experts must be enabled to jointly create a working ontology without becoming knowledge engineers themselves. Tools engineered for this task can aid domain experts, but nevertheless it is still an open question how well this will work in the end.

to create broader/narrower relations between tags. However, tag relationships are only local, i.e., each user can (and has to) maintain his/her own relationships and cannot profit from others’ contributions in that respect.

int.ere.st relAted worK There are a number of other approaches that follow a similar direction often presented as semantic tagging. These include Bibsonomy, Int.ere.st, GroupMe, Fuzzzy, and Annotea, which we will describe and compare in the following.

bibsonomy Bibsonomy (Hotho 2006) is a system for the management of bookmarks of internet resources and publication entries. Bibsonomy offers functionality similar to that of well-known social bookmarking services but specifically tailored towards academics – e.g., it offers sophisticated support for uploading and exporting bibliographic information. At its core, Bibsonomy has a functionality very similar to social bookmarking services, but additionally offers users the possibility to create broader/narrower relations between tags. However, tag relationships are only local, i.e., each user can (and has to) maintain its own relationships and cannot profit from others’ contributions in that respect.is a system for the management of bookmarks of internet resources and publication entries. Bibsonomy is maintained by the Knowledge and Data Engineering Group of the University of Kassel. Bibsonomy offers functionality similar to that of well-known social bookmarking services but specifically tailored towards academics – e.g., it offers sophisticated support for uploading and exporting bibliographic information. At its core, Bibsonomy has a functionality very similar to social bookmarking services, but additionally offers users the possibility

Int.ere.st (Kim 2007) is a system concentrating on the transferability of tags and tagged resources between systems. Int.ere.st is created by the Digital Enterprise Research Institute, Galway and the Biomedical Knowledge Engineering of Seoul National University, Korea. Its functionality centers on making uploading and exporting tagging data simple and to allow for creating relations between tags (potentially coming from different systems).

groupMe! GroupMe (Abel 2007) attempts to bridge the gap between the Semantic Web and Web2.0 with an RDF based social bookmarking application. GroupMe! is developed by the Semantic Web Group at the University of Hannover in Germany. The main unique functionality of GroupMe! is the extension of the tagging idea with the concept of ‘groups’: all annotated resources can be organized into groups and these form another level of information that can be used for browsing and search.

fuzzzy Fuzzzy (Lachica 2007) is a system for managing bookmarks of internet resources and ISBN numbers. Fuzzzy is developed within the PhD project of Roy Lachica at the University of Oslo. It is based on Topic Maps technology and besides parent/child and horizontal tag relations the users can choose of 22 specific predefined association types to link tags. Another main concept is voting for gardening and maintenance: the users can vote

235

Social Semantic Bookmarking with SOBOLEO

on bookmarks, tags a bookmark is annotated with, relations between tags, and users. •

Annotea Annotea (Koivunen 2006) is a metadata standard for semantic web annotations, it is implemented in a number of tagging tools and server applications. Annotea and its implementations have been developed by the W3C. Annotea differs from other approaches to social tagging in its emphasis on standards on decentrality, that it has sharing of bookmarks among services build in from ground up.

Comparison To give a comprehensive overview of the respective strength and weaknesses of the approaches shortly introduced above, the table below details the main discriminating features among the applications, including SOBOLEO. The features used for the comparison are the following: •











236

Public: Whether the application has a public installation that can be used by any user. Full Text Indexing: Whether the application stores the text content of the annotated resources and uses it to facilitate search. Import/Export Formats: All tools discussed have some means to import or export the bookmarks, this row details which formats are used. Synonyms: Whether the application supports a notion of two natural language terms representing the same thing. Other Relations: The relations between tags/concepts that are supported by the applications, other than synonyms. Shared Relation Editing: Whether relations between tags exist only for one user or whether they are shared, i.e. in some systems the relation between tags is only

visible to one user. Other users would need to create the same relation again. Open Source: Whether the source code of the applications is available as open source.

As a general conclusion, there is a big interest to extend social bookmarking in the direction of more semantics and in particular to tackle the problem how tagging data can be exchanged between systems, however, at the same time the table shows that there still is considerable disagreement about what are the most important features and – even more crucially – what are suitable formats to exchange the tagging data. Without an agreement in this domain, the promise of exchanging tagging data can obviously not be achieved. It is also interesting to see that the majority of the approaches still restricts the editing of relations between tags to only the private space and/or do not allow for a real community driven evolution of the semantic model.

fUtUre trends Based on this state of the art, it is obvious that there is a huge potential for future development of Social Semantic Bookmarking approaches as part of the developments towards a Web 3.0 as a user-centered semantic web, combining usercentered Web 2.0 approaches, and the potentials of semantic technologies. In particular, we observe four major trends •

The incremental and stepwise adoption of social semantic bookmarking ideas by current social bookmarking providers. This can be seen to some extend already today in the adoption of ‘tag bundles’ (a simpliied kind of superconcept) by del. icio.us or the user machine tags by lickr3 (tags from a reserved namespace that are used to, for example, display pictures on a

Social Semantic Bookmarking with SOBOLEO

Table 1. Comparison of social semantic bookmarking tools Public







Full Text Indexing

Import/ Export Formats

Synonyms

Other Relations

Shared Relation Editing

Open Source

Bibsonomy

Yes

No

XML, RSS, BURST, SWRC, Bibtex

No

Broader/ Narrower

No

No

Int.ere.st

No

No

SCOT, SIOC, FOAF

Yes

Identical

No

No

GroupMe!

Yes

No

RSS, FOAF

No

Group

Yes

No

Fuzzzy

Yes

Yes

XTM, RSS

Yes

Broader/ Narrower, Specific association types

Yes

No

Annotea

No

No

Annotea

Yes

Broader/ Narrower

No

yes

SOBOLEO

No

yes

SKOS, RSS

Yes

Broader/ Narrower, Related

Yes

No

DC,

map). Slowly there will also be an emerging consensus on the formats best suited for the exchange of tagging data. In academia the extension of social semantic bookmarking into the direction of more formality (e.g., enabling users to state information such as that two concepts are disjunct) and the attempt to merge it into a general collaborative editing of a knowledge store and its fusion with semantic wiki approaches. There will also be an increased interest to combine semantic tagging with machine learning approaches. The movement of social (semantic) tagging approaches into the enterprise and, in this context, the combination of traditional semantic tagging approaches with social semantic tagging. These approaches will use a core of a centrally deined vocabulary that is not editable and a fringe that is evolved by the community. The extension of tagging approaches beyond the domain of internet resources

and iles, applications and appliances allowing tagging of locations in the ‘real world’ or people in the context of competence management (Braun 2008, Farell 2006).

ConClUsion Social semantic bookmarking allows a group of users to collaboratively create and evolve an index of resources together with the powerful semantic vocabulary used to organize it. Social semantic bookmarking promises better retrieval, better use of annotation, better integration of the repository with semantic web infrastructure etc. while avoiding the problems commonly associated with semantic annotation approaches – such as a high initial cost to build ontologies. Parts of the vision of social semantic bookmarking are already realized today in applications widely used, and evaluation studies like (Braun 2007a) confirm that users appreciate the new paradigm. We are confident that more will follow

237

Social Semantic Bookmarking with SOBOLEO

within the next years. The future will also see social semantic bookmarking being used in the enterprise and researchers fusing it with semantic wiki approaches and extending them with more formal semantics. But for that, we also need a better understanding of the emergence and evolution of ontologies as part of everyday collaborative activities and appropriate models and support mechanisms. Promising research approaches include the ontology maturing process (Braun 2007b), which is further explored as part of the Integrating Project MATURE4 .

referenCes Abel, F., Henze, F. M., Krause, D., Plappert, D., & Siehndel, P. (2007). Group me! Where Semantic Web meets Web 2.0. In J. Golbeck & P. Mika (Eds.), Proceeding of the 5th Semantic Web Challenge, 6th International Semantic Web Conference (ISWC 2007). CEUR Workshop Proceedings (Vol. 295). Barker, K., Chaudhri, V. K., Char, S. Y., Clark, P., Fan, J., Israel, D., et al. (2004). A questionanswering system for AP chemistry: Assessing KR&R technologies. In D. Dubois, C. Welty & M.-A. Wiliams (Eds.), Proceedings of the 9th International Conference on Principles of Knowledge Representation and Reasoning (pp. 488-497). Menlo Park: AAAI Press Braun, S., & Schmidt, A. (2008). People tagging & ontology maturing: Towards collaborative competence management. In P. Hassanaly, A. Ramrajsingh, D. Randall, P. Salembier & M. Tixier (Eds.), 8th International Conference on the Design of Cooperative Systems (COOP 2008) (pp. 231-241). Aix-en-Provence: Institut d’Etudes Politiques d’Aix-en-Provence

238

Braun, S., Schmidt, A., Walter, A., Nagypal, G., & Zacharias, V. (2007a). The ontology maturing approach to collaborative and work-integrated ontology development: Evaluation results and future directions. In L. Liming Chen, P. CudréMauroux, P. Haase, A. Hotho & E. Ong (Eds.), Emergent Semantics and Ontology Evolution 2007. Proceedings of the First International Workshop on Emergent Semantics and Ontology Evolution (ESOE-2007), 6th International Semantic Web Conference (ISWC 2007) (pp. 5-18). CEUR Workshop Proceedings (Vol. 292). Braun, S., Schmidt, A., Walter, A., Nagypal, G., & Zacharias, V. (2007b). Ontology maturing: A collaborative Web 2.0 approach to ontology engineering. In N. Noy, H. Alani, G. Stumme, P. Mika, Y. Sure & D. Vrandecic (Eds.), Proceedings of the Workshop on Social and Collaborative Construction of Structured Knowledge (CKC), 16th International World Wide Web Conference (WWW 2007). CEUR Workshop Proceedings (Vol. 273). Dean, M., & Schreiber, G. (2004, February 10). OWL Web ontology language reference. W3C Recommendation. Dill, S., Eiron, N., Gipson, D., Gruhl, D., & Guha, R. (2003). SemTag and seeker: Bootstrapping the Semantic Web via automatic semantic annotation. In Proceedings of the 12th International World Wide Web Conference (WWW 2003) (pp. 178-186). New York: ACM Press. Farrell, S., & Lau, T. (2006). Fringe contacts: People-tagging for the enterprise. In Proceedings of the Collaborative Web Tagging Workshop, 15th International World Wide Web Conference (WWW 2006). Handschuh, S., & Staab, S. (2002). Authoring and annotation of Web pages in cream. In D. Lassner, D. De Roure & A. Iyengar (Eds.), Proceedings of the 11th International World Wide Web Conference (WWW 2002) (pp. 178-186). New York: ACM Press

Social Semantic Bookmarking with SOBOLEO

Hepp, M. (2007). Possible ontologies: How reality constraints building relevant ontologies. IEEE Internet Computing, 11(1), 90–96. doi:10.1109/ MIC.2007.20 Hotho, A., Jäschke, R., Schmitz, C., & Stumme, G. (2006). BibSonomy: A social bookmark and publication sharing system. In A. de Moor, S. Polovina & H. Delugach (Eds.), Proceedings of the 1st Conceptual Structures Tool Interopability Workshop (CS-TIW 2006)(pp. 87-102), Aalborg, Aalborg University Press Kim, H. L., Yang, S.-K., Song, S.-J., Breslin, J. G., & Kim, H.-G. (2007). Tag mediated society with SCOT ontology. In J. Golbeck & P. Mika (Eds.), Proceeding of the 5th Semantic Web Challenge, 6th International Semantic Web Conference (ISWC 2007). CEUR Workshop Proceedings (Vol. 295). Koivunen, M.-R. (2006). Semantic authoring by tagging with annotea social bookmarks and topics. In K. Möller, A. de Waard, S. Cayzer, M.-R. Koivunen, M. Sintek & S. Handschuh (Eds.), Proceedings of the Semantic Authoring and Annotation Workshop (SAAW), 5th International Semantic Web Conference (ISWC 2006). CEUR Workshop Proceedings (Vol. 209). Lachica, R., & Karabeg, D. (2007). Metadata creation in socio-semantic tagging systems: Towards holistic knowledge creation and interchange. In L. Maicher & L. M. Garshol (Eds.), Scaling topic maps. Topic Maps Research and Applications (TMRA 2007) (LNCS 4999, pp. 160-171). Berlin: Springer. Macgregor, G., & McCulloch, E. (2006). Collaborative tagging as a knowledge organisation and resource discovery tool. Library Review, 55(5), 291–300. doi:10.1108/00242530610667558 Manola, F., & Miller, E. (2004, February 10). RDF primer. W3C Recommendation.

Marlow, C., Naaman, N., Boyd, D., & Davis, M. (2006). Position paper, tagging, taxonomy, flickr, article, to read. In Proceedings of the Collaborative Web Tagging Workshop, 15th International World Wide Web Conference (WWW 2006). Miles, A., & Bechhofer, S. (2008, January 25). SKOS simple knowledge organization system reference. W3C Working Draft. Popov, B., Kiryakov, A., Kirilov, A., Manov, D., Orgnyanoff, D., & Goranov, M. (2003). KIMsemantic annotation platform. In D. Fensel, K. Sycara & J. Mylopoulos (Eds.), Proceedings of the 2nd International Semantic Web Conference (ISWC 2003) (LNCS 2870, pp.834-849). Berlin: Springer. Sen, S., Lam, S. K., Rashid, A. M., Cosley, D., Frankowski, D., Osterhouse, J., et al. (2006). Tagging, communities, vocabulary, evolution. In P. Hinds & D. Martin (Eds.), Proceedings of Proceedings of the 2006 20thAanniversary Conference on Computer Supported Cooperative Work (CSCW 2006) (pp. 181-190). New York: ACM Press. Vargas-Vera, M., Motta, E., Dominique, J., Lanzoni, M., Stutt, A., & Ciravegna, F. (2002). MnM: Ontology driven semi-automatic and automatoc support for semantic markup. In A. Gómez-Pérez & V. R. Benjamins (Eds.), Proceedings of the 13th International Conference on Knowledge Engineering and Management (EKAW 2002) (pp. 371-391). New York: ACM Press. Zacharias, V., & Braun, S. (2007). SOBOLEO– social bookmarking and lightweight ontology engineering. In N. Noy, H. Alani, G. Stumme, P. Mika, Y. Sure & D. Vrandecic (Eds), Proceedings of the Workshop on Social and Collaborative Construction of Structured Knowledge (CKC), 16th International World Wide Web Conference (WWW 2007). CEUR Workshop Proceedings (Vol. 273).

239

Social Semantic Bookmarking with SOBOLEO

Key terMs And definitions Annotation: An annotation is extra information that is associated with some data, usually a document or website. Common uses of annotation are keywords associated to images that aid retrieval or comments about the quality of (parts of) the document in question. An annotation is normally added after the creation of a document and mostly created by other people than the initial author. The verb ‘to annotate’ refers to the process of adding an annotation to some document. An annotation is a special kind of meta-data, distinguished by the property of mostly being added later; however, the difference between annotation and meta-data is not clear cut. Meta-Data: Meta data is data about data; extra information associated with some data, for example a document or website. Common uses of meta-data are creation dates and access right associated with files or information about the tool that was used to create a particular file. Meta-Data is often created at the same time as the data it describes, but can also be created at some later time. Meta-Data can be both embedded in the data it describes or external to it. RDF: Resource Description Framework, a W3C specification for a data exchange format that can support decentralized meta-data exchange on a global scale. RDF is built on the idea of triples that are used to represent everything. A triple consists of a subject, a predicate and an object and represents one statement that is made about the relation between resources. An example for a triple would be ‘Mike has type Human’. Semantic Tagging: Semantic Tagging is the process of associating an element from an ontology with some document, usually a computer file or website. Semantic tagging serves the goal of describing a document in order to facilitate better retrieval later on. Semantic tagging also helps to integrate the tagged document with other resources

240

that are also related to the same ontology. Semantic tagging is a special kind of annotation. Semantic Web: The vision of improving the internet making the content of the web more accessible to machines, this should enable agents to handle more complex task on behalf of the user. The Semantic Web initiative has given rise to standards such as RDF, OWL and SPARQL that aim to make representing information and exchanging information on the web possible. SOBOLEO: SOcial BOokmarking and Lightweight Engineering of Ontologies is a system for the webbased collaborative engineering of SKOS ontologies and annotation of internet resources. SOBOLEO enables the simple creation, extension and maintenance of ontologies. At the same time it supports the annotation of internet resources with elements from this ontology Social Tagging: Social tagging is the process of tagging and use of tagged resources in the context of systems that bring together the tags from a group of people for improved retrieval and in order to foster relationships between the users. Social tagging systems allow to discover other users by finding people that tagged the same resource or use the same tag. These systems also support the discovery of new information using the set of all tags made by all users. SKOS: The Simple Knowledge Organisation System is a RDF vocabulary for the representation of different kinds of structured vocabulary such as thesauri, taxonomies or subject-heading systems. SKOS is build on top of RDF. Tagging: Tagging is the process of associating a keyword or term (a ‘Tag’) with some document, usually a computer file or website. Tagging serves the goal of describing a document in order to facilitate better retrieval later on. The tags used in tagging are usually chosen informally by the person doing the tagging. Tagging is a special kind of annotation.

Social Semantic Bookmarking with SOBOLEO

endnotes 1 2 3 4

http://www.delicious.com http://www.soboleo.com http://www.flickr.com http://mature-ip.eu

241

242

Chapter 14

Social Bookmarking and Web Search Yusuke Yanbe Kyoto University, Japan Adam Jatowt Kyoto University, Japan Satoshi Nakamura Kyoto University, Japan Katsumi Tanaka Kyoto University, Japan

AbstrACt Social bookmarking is an emerging type of a Web service for reusing, sharing, and discovering resources. By bookmarking, users preserve access points to encountered documents for their future access. On the other hand, the social aspect of bookmarking results from the visibility of bookmarks to other users helping them to discover new, potentially interesting resources. In addition, social bookmarking systems allow for better estimation of the popularity and relevance of documents. In this chapter, we provide an overview of major aspects involved with social bookmarking and investigate their potential for enhancing Web search and for building novel applications. We make a comparative analysis of two popularity measures of Web pages, PageRank and SBRank, where SBRank is deined as an aggregate number of bookmarks that a given page accumulates in a selected social bookmarking system. The results of this analysis reveal the advantages of SBRank when compared to PageRank measure and provide the foundations for utilizing social bookmarking information in order to enhance and improve search in the Web. In the second part of the chapter, we describe an application that combines SBRank and PageRank measures in order to rerank results delivered by Web search engines and that offers several complimentary functions for realizing more effective search. DOI: 10.4018/978-1-60566-384-5.ch014

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Social Bookmarking and Web Search

introdUCtion Social bookmarking is one of the main trends of a new generation of the Web called Web 2.0. The idea behind social bookmarking is to let users store URLs to their favorite pages and make them visible to other users. Each social bookmark is annotated with tags that describe the bookmarked resource and that were freely chosen by a bookmarker. Del.icio.us1 is currently the most popular social bookmarking service. It has been operating since 2003 and currently has about 3 million users that bookmarked around 100 million Web documents2. There are also other popular social bookmarking sites such as Furl3 or Simpy4. Non-social bookmarking was proposed first by (Keller, Wolfe, Chen, Labinowitz, & Mathe, 1997) as a way to remember and locally store access points to visited Web documents. In social bookmarking the social aspect of bookmarking allows for discovery of new, potentially relevant resources thanks to the combined effort of many users. This makes it also possible to determine the resources that are both relevant (by the analysis of their tags) and recently popular (by counting their bookmarks) as well as permits to track their popularity and relevance over time. For example, del.icio.us informs users about popular pages that recently obtained many bookmarks and cloudacio.us5 displays historical patterns and trends of bookmarked resources. The incentives of social bookmarkers have been recently categorized by (Marlow, Naaman, Boyd, & Davis, 2006). According to the authors, users decide to bookmark the resources because of the following reasons: future retrieval, contribution and sharing, attract attention, play and competition, self presentation, opinion expression. In most cases, however, bookmarking are useful for individual users who want to externally (hence beyond the limit of a single PC machine) store access points to their selected resources. However, in this way, the users help also to manually arrange the Web in a bottom-up fashion since they categorize

the online resources and enable better estimation of their popularity and, indirectly, quality. An important characteristic of tagging in social bookmarking systems is the lack of any controlled vocabulary. Users are free to annotate bookmarked documents as they wish or they can borrow same tags as others used. However, after some time, certain forms of tag agreements emerge for the resources as demonstrated in (Golder & Huberman, 2006). The process of resource categorization by free tagging is called folksonomy and is inherently different from a rigid classification usually done by domain experts, for example, by librarians. However, the well-known problems with folksonomy result from its advantages, that is, from the uncontrolled, free character of categorizing resources by many users. For example, ambiguity, synonymy or polysemy occur among tags that can undermine the retrieval process. In this chapter we discuss the social bookmarking phenomenon and provide the results of analytical study aimed at analyzing the usefulness of social bookmarks for improving the search in the Web. In particular, we perform a comparative analysis of two popularity measures of Web pages, SBRank and PageRank. PageRank is a popular iterative algorithm that scores Web pages based on the random surfer model (Page, Brin, Motwani, & Winograd, 1998). In short, a page has a high PageRank value if it is linked from a relatively large number of other pages that also have high PageRank scores. By finding popular resources both SBRank and PageRank provide means for selecting high quality Web pages assuming a positive correlation between popularity and quality. In addition, we analyze several other aspects of pages bookmarked in social bookmarking sites. For example, we investigate the dynamics of the both metrics in order to confirm whether mixing them could improve the dynamic characteristics of search results. Next, we also discuss the potentials of social bookmarking systems for providing new, successful kinds of Web services. For example, a

243

Social Bookmarking and Web Search

promising business model could be based on exploiting social bookmarks for improving the precision of direct advertising. We emphasize here the additional aspects of social bookmarks such as the availability of temporal and location-type data that can be leveraged for realizing extended and efficient search models. The historical pattern of social bookmarks accumulation could aid in a better discovery of resources that are in their popularity peaks or that are becoming recently popular among Web users. Additionally, categorizing tags into content-descriptive and sentiment-bearing ones allows for capturing page semantics and estimating user attitudes towards the bookmarked documents. We discuss several potential application examples and explain those characteristics of social bookmarks and, in general, social bookmarking that are likely to be key points in creating successful business models. Here, we also demonstrate the application for extended Web search that we have designed based on some of these ideas.

relAted reseArCh PageRank (Page, Brin, Motwani, & Winograd, 1998) and HITS (Kleinberg, 1999) are the most popular link-based page ranking algorithms for measuring the popularity of pages. Since automatically judging the quality of resources is still not possible to be done by machines, hence the most effective approach is to rely on opinions of large number of Web authors. The calculations of both PageRank and HITS are thus dependent on the number of pages linking to target resources and, basically, favor documents that are commonly refereed to from other pages. Nevertheless, the link-based popularity estimation algorithms suffer from certain problems such as link-spamming, difficulties in link creation and poor temporal characteristics. One reason for this is that only Web authors can cast votes to pages by creating links to them, which is in contrast to socially

244

bookmarking pages, where anyone can “recommend” pages to others. Social bookmarking is thus a more democratic process of resource selection due to its simplicity and accessibility. In addition, since it is necessary for a resource to acquire large number of in-links in order to become more visible on the Web, usually, certain time delay occurs before the page can obtain high score as calculated by the link-based metrics. The observation that PageRank is biased against new pages as it takes some time until pages become noticed and popular among Web authors has been made by (Baeza-Yates, Castillo, & Saint-Jean, 2004) (Yu, Li, & Liu, 2004). Some researchers attempted at eliminating this bias to some degree through incorporating the last-modification dates of pages (Baeza-Yates, Castillo, & Saint-Jean, 2004) or adding exponentially decaying factors to PageRank scores (Yu, Li, & Liu, 2004). Lastly, links may have variety of meanings and purposes as discussed in (Mandl, 2006). Previous studies on social bookmarking focused mostly on the issues related to folksonomy (Golder & Huberman, 2006) (Marlow, Naaman, Boyd, & Davis, 2006) (Wu, Zubair, & Maly, 2006) (Wu, Zhang, & Yu, 2006) (Yu, Li, & Liu, 2004) (Zhang, Wu, & Yu, 2006). For example, (Zhang, Wu, & Yu, 2006) introduced a hierarchical concept model of folksonomies using HACM - a hierarchy-clustering model reporting certain kinds of hierarchical and conceptual relations between tags. In another work, (Golder & Huberman, 2006) investigated the characteristics of tagging and bookmarking and revealed interesting regularities in user activities, tag frequencies, and bursts in popularity of tags in social bookmarks. The authors also analyzed tagging dynamics as well as classified tags into seven categories depending on the functions they perform for bookmarks. None of the previous studies, however, focused on the comparative analysis of link structure and social bookmarking-derived metrics. Perhaps, the closest work to ours is the one done by Heymann et al. (Heymann, Koutrika, &

Social Bookmarking and Web Search

Garcia-Molina, 2008) who has recently made a large scale analysis of data in Del.icio.us to investigate whether social tagging Websites can be of any use for Web search. Among other results, they found that popular query terms and tags overlap significantly, most tags were deemed relevant and objective by users and tags are present in page text of 50% of pages. Our work is, however, unique in its comparison between SBRank and PageRank as well as in its larger focus on temporal characteristics of social bookmarks. We also propose merging link-based ranking metrics with the metric that leverages results of collaborative tagging. In addition, we exploit other characteristics of social bookmarking systems such as agglomerated user behavior, sentiment of users towards bookmarked resources. Recently, (Bao, Wu, Fei, Xue, Su, & Yu, 2007) introduced an authority-centric search model in social bookmarks that is based on the number of social bookmarks that pages have and the characteristics of users bookmarking them. In (Wu, Zubair, & Maly, 2006) HITS algorithm was adapted for identifying high quality resources and users that provide such resources in the Web. In another work, (Damianos, Griffith, & Cuomo, 2006) proposed using social bookmarking for information sharing and management in corporate Intranets. Finally, (Wu, Zhang, & Yu, 2006) described techniques for exploiting social bookmarking for the purpose of fostering the development of Semantic Web. The authors used probabilistic generative model to capture emerging semantics of resource

AnAlysis of soCiAl booKMArKing

using Hatena Bookmark6, which is the most popular bookmarking service in Japan started in February 2005. The datasets were obtained in the following way. We have used popular tags, which are sets of the most popular and recently used tags published by del.icio.us and Hatena Bookmark. 140 tags have been retrieved on December 6th, 2006 from del.icio.us and 742 tags on February 16th, 2007 from Hatena Bookmark. Next, we collected popular URLs from these tags. Usually less than 25 popular pages were listed for each tag in both the social bookmarking systems. At this stage, we obtained 2,673 pages for del.icio. us and 18,377 pages for Hatena Bookmark. In the last step, we removed duplicate URLs (i.e. URLs listed under several popular tags). In result, we obtained 1,290 and 8,029 unique URLs for del.icio. us and for Hatena Bookmark, respectively. Each URL had two attributes: irstDate and SBRank. irstDate indicates the time point when a page was introduced to the social bookmarking system for the irst time by having first bookmark created. SBRank, as mentioned above, is the number of accumulated bookmarks of a given page obtained at the date of the dataset creation. In order to detect PageRank scores of the URLs, we used Google Toolbar7 - a browser toolbar that allows viewing PageRank scores of visited pages. PageRank scores obtained in this way are approximated on the scale from 0 to 10 (0 means the lowest PageRank score). To sum up, the obtained datasets are the snapshots of the collections of popular pages in both social bookmarking systems. Each page has its PageRank and SBRank scores recorded that it had at the time of the dataset creation.

dataset Characteristics

distribution of Pagerank and sbrank

We have created two datasets for our study. We have chosen del.icio.us as a source of the irst dataset, while the second dataset was created

Figures 1 a and 1 b show the distributions of SBRank scores in the both datasets. We can see that only few pages are bookmarked by many us-

245

Social Bookmarking and Web Search

Figure 1. Distribution of SBRank scores, a) del.icio.us dataset, and b) Hatena Bookmark dataset

Figure 2. Distribution of PageRank scores, a) del.icio.us dataset, and b) Hatena Bookmark dataset

ers, while the rest is bookmarked by a relatively low number of users. Figures 2 a and 2 b show, in contrast, the distribution of PageRank scores in both datasets. We found that more than a half of pages (56.1%) have PageRank scores equal to 0 in the del.icio.us dataset, while Hatena Bookmark dataset (81%) contains even more such pages,

246

which is probably due to its more local scope. These pages have relatively low popularity as measured by the link-based metric. Thus it may be difficult for users to find them through conventional search engines. For example, Figure 3 shows the average distribution of PageRank scores in search results delivered from Yahoo!

Social Bookmarking and Web Search

Figure 3. Average PageRank scores of top 500 pages

search engine8 for two sample query sets (see Appendix on how the query sets were created). Nevertheless, many social bookmarkers considered the pages to be of high quality by bookmarking them. We think that it would be advantageous if some of these pages could be added into search results provided they are relevant to issued queries. In general, we think that there may be two possible reasons that caused the occurrence of many pages with low PageRank scores in the datasets, despite their relatively high popularity among social bookmarkers. One is that the pages could have been created rather recently, hence, on average, they did not acquire many in-links and indirectly the high visibility on the Web. Or, in contrast, the pages have been published on the Web since long time, yet their quality cannot be reliably estimated using PageRank measure due to various other reasons. We decided to analyze the temporal characteristics of pages stored in our dataset in order to provide an answer the above question. Figure 4 a and 4 b show the distribu-

tions of the dates of page additions to the social bookmarking systems (i.e., irstDate). Looking at these figures, we can see that more than a half of the pages appeared in the social bookmarking systems within the first three months before the dataset creation. The other half of the pages was bookmarked in the system for the irst time more than three months before the time when the data was crawled. The reason that Hatena Bookmark dataset contains more fresh pages is most likely due to its shorter age than that of del.icio. us. From these figures it can be seen that there are many recently added pages in both datasets. This is especially true for pages with PageRank scores equal to 0 as shown in Figure 5 a and 5 b. Such pages usually do not have enough time to acquire many in-links, hence, they retain low PageRank scores. From these results we observe one of the useful aspects of SBRank when compared to link-based metrics. The latter is not effective in terms of fresh information retrieval since pages

247

Social Bookmarking and Web Search

Figure 4. Histogram of firstDate of pages, a) del.icio.us dataset, and b) Hatena Bookmark dataset

Figure 5. Histogram of firstDate of pages that have PageRank score equal to 0, a) del.icio.us dataset, and b) Hatena Bookmark dataset

require relatively long time in order to acquire large number of in-links. Relatively novel pages may have thus difficulties in reaching top search results in current search engines even if their quality is quite high. We confirm the results of (Baeza-Yates, Castillo, & Saint-Jean, 2004) that PageRank scores of pages are highly correlated with their age in Figure 6 a and 6 b. The correlation

248

coefficients between PageRank and irstDate is r = −0.85 for del.icio.us and r = −0.51 for Hatena Bookmark datasets. To sum up, the results suggest that many popular pages in social bookmarking systems have relatively low PageRank scores. In addition, we confirm that SBRank has better dynamics than the traditional link-based page ranking metric. This

Social Bookmarking and Web Search

Figure 6. Scatter plot of firstDate and PageRank score, a) del.icio.us dataset, and b) Hatena Bookmark dataset

Google Toolbar. We normalize both SBRankj and PageRankj scores by dividing them by the maximum scores found for all N pages. α is a mixing parameter with the value ranging from 0 to 1. Below, we discuss the extensions to a query model that are possible thanks to considering the various types of information attached to social bookmarks.

is because social bookmarks allow for a more rapid, and unbiased, popularity estimation of pages. Complementing PageRank using SBRank has thus potential to improve the search process on the Web.

enhAnCed web seArCh ProPosAl

Metadata search type In this section, we demonstrate a simple method for enhancing Web search by re-ranking top N results returned from conventional search engines using the information about the number of their social bookmarks and their associated tags. To implement a combined rank estimation measure we propose a linear combination of both ranking metrics. In the equation, SBRankj is the number of bookmarks of a page j in del.icio.us, while PageRankj is a PageRank score of the page acquired using

CombinedRank j

SBRank j maxi:1

i N

SBRanki

Tags annotated by users are useful for a so-called search by metadata (“metadata query”). In our search model a user can issue both a traditional query, which we call “content query” as well as “metadata query”. In such a case, pages that contain the content query in their contents will be re-ranked considering the overlap between user-issued metadata query and the tags describing them. For each page we construct a tag vector based on accumulated tags of the page and their

1

PageRank j maxi:1

i N

PageRanki

(1)

249

Social Bookmarking and Web Search

frequencies. We also construct a metadata query vector in the same way. Then the cosine similarity between metadata query and the tag vectors is used for re-ranking search results pages.

temporal search type It is also possible to construct temporally constrained queries thanks to using timestamps of social bookmarks. In consequence, pages could be retrieved according to arbitrary aspects of their popularity patterns. For example, pages that feature quickly rising popularity over time can be returned on top results. Note that this functionality cannot be realized using traditional link-based approaches, as there is no available data on the link evolution of the Web. The temporal extension to the query model that we propose is as follows: First, we propose filtering pages according to their firstDate values, that is, according to the ages of pages inside social bookmarking systems. Users can find those documents that have been recently introduced to the social bookmarking system. Obviously pages may be older than that. Next, we allow users to search for pages according to the popularity variance over time (variance of SBRank over time). For example, search results may contain pages with stable popularity function or pages reflecting large fluctuations in user preferences over time. Lastly, we propose capturing levels of page popularities in certain, specified periods of time through summing the numbers of bookmarks made to pages during those intervals and re-ranking pages accordingly. Thus, user can request relevant pages that were popular within a selected time period.

sentiment search type Pages are often tagged by subjective or “egoistic” tags such as “shocking”, “funny”, “cool”. Although, typically, such tags are viewed negatively

250

as hindering the retrieval process, we consider them as providing useful information about page quality and user impressions. In our enhanced search model, we thus enable users to issue a sentiment-like query. Before implementing this feature, we first measured the number of sentiment tags used by bookmarking users. We have used here the Hatena Bookmark dataset. Tags in this dataset were classified into two groups according to the taxonomy of tags defined by (Golder & Huberman, 2006): • •

tags that identify what or who the resource is about tags that identify qualities or characteristics of resources (e.g., “scary”, “funny”, “stupid”)

We have manually examined top 1,100 tags from Hatena Bookmark dataset in order to distinguish between content and sentiment tags. Table 1 shows the top 10 content and sentiment tags after translation9. In the top 30 tags we have observed the ratio of content tags to sentiment tags to be about 10:1. In Figure 7 we show the distribution of tag frequencies. The results reveal that top 3 sentiment tags are very common, while the other tags are rather less used. After including synonyms we found that the most popular sentiment tags are: useful, amazing and awful. Figure 8 shows the top 54 sentiment tags located on the negative-positive scale with their frequencies. The tags that occur more than 3000 times are placed over the dashed line, while those with frequencies less than 100 times are under the horizontal axis. It is easy to notice that, in general, there are more positive sentiment tags than negative ones. Besides, the positive ones are also more frequently used. Only one negative sentiment tag has been used more than 100 times (“it’s awful”). This means that social bookmarkers rarely bookmark resources for which they have negative feelings. For the purpose of utilizing sentiment tags in our search model, we create page sentiment vec-

Social Bookmarking and Web Search

Table 1. Top 10 content tags (left) and top 10 sentiment tags (right) Tag Name

N

Tag Name

N

Web

16,633

useful (1)

5,381

google

15,674

it’s amazing

5,046

troll

14,453

it’s awful

4,123

javascript

11,840

useful (2)

3,041

youtube

10,858

interesting

638

tips

10,784

funny (1)

616

css

9,411

it’s useful (3)

544

design

8,423

funny (2)

419

2ch

8,381

useful (4)

377

society

7,412

I see

365

tor from the sentiment tags added by users. The similarity of this vector to the vector built using user-issued sentiment queries is then used for computing sentiment-based scores of pages.

1.

final re-ranking

3.

In this section we discuss the way to integrate the above discussed extensions. At query time the system performs the following operations:

4.

2.

Obtain top n pages from search results returned by a search engine P={p1, p2,…, pn} for query q Obtain SBRank scores for each pi where pi∈P Obtain every bookmark and its associated data for each pi that has SBRank > 0 (i.e., the page has at least one social bookmark) Count the number of occurrences of users and tags to be used for providing “related

Figure 7. Frequency distribution of top 20 content and sentiment tags

251

Social Bookmarking and Web Search

Figure 8. Top 54 sentiment tags on the negative-positive sentiment scale

Tags” and “related Bookmarks” structures (described later) In order to incorporate the above re-ranking mechanisms, we have applied the ranking formula shown in Figure 9. The original search results returned by the search engine are re-ranked using Rank(pi) function. Here, B(pi) represents the popularity estimate of pi using the combination of SBRank(pi) and SearchRank(pi), which is the rank of the page in the results returned from a search engine. Not that although, our concept is to combine SBRank with PageRank scores, in the actual implementation, we have used the ranks of pages returned from search engines as an approximation of pages’ popularity on the Web. F(pi) is the freshness level of pi; V(pi) is a variance measure of the function representing added bookmarks to pi. sim(tagi,tagq) is the similarity between page tag vector and query vector,

252

while sim(tagseni, tagq) is the similarity between the page sentiment vector and the query vector. S(pi, tbeg, tend) is the proportion of bookmarks of pi, which have been added in the time period [tbeg, tend] to the total number of bookmarks added to this page in the bookmarking system. α, β, γ and δ are control parameters.

system interface The interface of the prototype application that we have build is shown in Figure 10. It contains 4 slide-bar controls for adjusting α, β, γ and δ parameters. “Time span” control was implemented as the combination of two sliding bars used for selecting desired time periods (the time span can be also specified by entering dates in two textual boxes). Radio buttons were added in order to let users select one of the three most popular senti-

Social Bookmarking and Web Search

Figure 9. Formula used for re-ranking search results

ment expressions: useful, amazing and awful. As it is difficult to list all potential sentiment tags into the interface, we have grouped the similar sentiment expressions with their corresponding weights in order to form the three basic sentiment categories. By default the controls are at the positions in which they do not influence search results so that users can perform the usual content-based search without any additional query features. In addition, the system dynamically generates navigable structures called “related tags” and “related bookmarks” according to issued queries for enabling serendipitous discovery (see the right hand side of Figure 10). “Related tags” is a tag cloud listing 20 most frequently

occurring tags for all the N pages returned. The font sizes of tags are determined by the frequencies of their occurrences in the results. The tag cloud lets users explore other tags related to returned results and indirectly to the issued queries. Clicking on any tag makes pages listing documents assigned to this tag appear from social bookmarking systems. In addition, next to each tag, there are the “+” signs associated with the tags and clicking on them makes the system issue new search queries with the corresponding related tag. On the other hand, “related bookmarks” are tuples of social bookmark users and their tags obtained using the returned search results. Social bookmarkers have scores assigned depend-

253

Social Bookmarking and Web Search

Figure 10. Snapshot of the interface of the enhanced search system

ing how many of the returned results they have bookmarked. Then, for the top-scored users, the system detects the other most frequent tags that these users used provided they are assigned to the pages returned in the results. The links to the corresponding pages of these tags in social bookmarking systems are displayed as “related bookmarks”. In Figure 11 we show a modified interface of the above application. Here, upon request, users are able to observe the temporal pattern of social bookmark creation for each page returned in results.

254

sUMMAry In this section, we summarize positive and negative aspects of social bookmarks from the viewpoint of their usefulness for improving current Web search.

Positive factors The usefulness of social bookmarks for enhancing search in the Web has been recently proved by several research efforts (Bao, Wu, Fei, Xue, Su, & Yu, 2007) (Heymann, Koutrika, & Garcia-Molina, 2008). Similarly to Web links, social bookmarks

Social Bookmarking and Web Search

are kinds of votes cast by Web users to resources such as Web pages. Links, however, are usually created by document authors and, thus, average users are rather constrained in making links. On the other hand, social bookmarks, as being easily generated by users, are, as a consequence, a more democratic means of page quality assessment. As we have previously shown, on average, social bookmarks have better temporal characteristics (i.e., are more dynamic) and allow for early de-

tecting high quality pages that are often still not popular on the Web when judging by conventional approaches (e.g., PageRank). In addition, tags attached to bookmarks provide information about the topical scopes of bookmarked resources or sometimes even convey attitudes of users to bookmarked resources. Next, the timestamps of social bookmarks enable estimation of the fluctuations in their popularity within social bookmarking systems over time. This can be used for various

Figure 11. Snapshot of the system interface with displayed temporal patterns of social bookmarks accumulation

255

Social Bookmarking and Web Search

time-centric improvements of search results. In general the information associated with social bookmarks appears to be useful in improving the precision and extending the query model in information retrieval process.

combating spamming in social bookmarking systems. Lastly, as mentioned before, the lack of controlled vocabulary, misspellings, synonymy, polysemy, etc. can hinder the information retrieval process that utilizes tags in social bookmarks.

negative factors There are, however, several obstacles in directly utilizing social bookmarks for Web search. One is a relatively small number of pages having a considerable amount of social bookmarks on the Web. This problem, however, seems to be solvable in the near future considering the current popularity increase of social bookmarking systems among Web users (Heymann, Koutrika, & Garcia-Molina, 2008). Another issue is related to the characteristics of social bookmarking that makes it popular and useful for Web search. Namely, social bookmarks have high vulnerability to spamming. Since bookmarking pages is quite simple, hence, it is also relatively easy to influence the number of social bookmarks a given page has or to purposefully assign wrong tags to the page. The obvious reason for such manipulation would be the expected increase in the visibility of the page in the system and, indirectly, its visibility on the whole Web. Naturally, certain measures have been undertaken to cope with this problem. For example, filters are set against automatic creation of user accounts using “captcha”, detection of robot behavior or other preventive methods. Social bookmarking users can also report suspicious accounts or spam Web pages in several systems. Nevertheless, it is easy to foresee that the above approaches are superficial and cannot be effective in the future, neither scale well on the Web. If social bookmarks are ever going to be a serious improvement to the Web scale search and, actually, if they are going to be still useful in the future, the problem of spamming has to be appropriately tackled. It is thus apparent that this issue requires much research focus. However, so far, there has been no efficient proposal towards

256

other PotentiAl APPliCAtions We believe there can be many potential business opportunities related to social bookmarks. For example, social bookmarking in enterprises and intranets has been already proposed and implemented (e.g., (Damianos, Griffith, & Cuomo, 2006), (Millen, Feinberg, & Kerr, 2006)). Since there are already several different social bookmarking services on the Web we think that the combination of data from these could be advantageous to users. Similarly to the concept of meta-search engines this should increase the coverage, freshness and trustworthiness of generated resource ranking. An unsolved issue in such meta-search applications is the way to combine popularity estimates drawn from different sources. Such combination should perform normalization score according to various characteristics of different services such as their sizes or scopes. We see also potential in improving contextual advertising on the Web using social bookmarks. Currently, representative keywords are automatically extracted from pages that are later matched to the pool of ads. In an extended advertising model the advertising terms for a given page could be selected among tags associated to its bookmarks or could be even retrieved from among related pages commonly bookmarked by the same users. The integration of social bookmarking with social networks is another potential area. Social circles also help improve Web search. (Mislove, Gummadi, & Druschel, 2006) proposed the Web search system that can search pages using webbrowsing log between members of given social network as well as pages indexed by conventional

Social Bookmarking and Web Search

Web search engines. They reported that the system returned 8% non-indexed but viewed by community previously pages. They also desribed the issues of privacy, membership identification of social networks, ranking of search results, scalability and so on. In 2007, Google together with several social network service providers launched OpenSocial10. This is a unified platform that lets third party developers to utilize relationships between social network users, in their applications. Social networking is closely related to social bookmarking. Social network contains information on human relationships, while social bookmarks are descriptions of user interests. Integrating both systems should enhance the accuracy of recommendation, because social networks correspond to human relationships and the connected users often feature similar interests. Lastly, according to our recent study there is a lot of geographical information included in tags and we expect that some sort of location-aware search adaptation could be feasible. State-of-the art geo-tagging approaches rely on extracting location-related information from page content (e.g., (Amitay, Har’El, Sivan, & Soffer, 2004)) or from among linked resources. Supplementing these methods from the location-based information derived from social bookmarks and their tags may turn out advantageous. The investigation of this potential forms the part of our future work.

ConClUsion In this chapter, we aimed at increasing user’s familiarity with social bookmarking by discussing its major characteristics. We described the key aspects related to social bookmarking and its potential to enhance Web search or to be used for creating novel Web applications. An important part of this description is a comparative analysis of PageRank - a widely-known page popularity measure with a metric derived by the aggregation

of social bookmarks (SBRank). This comparative analysis is important in the view of recent proposals to incorporate social bookmarks into search mechanisms on the Web. In result of our analysis, we are able to indicate the areas where SBRank is superior to PageRank. To remain objective, we also discuss the shortcoming of the popularity estimation based solely on the amount of social bookmarks. In addition, we propose several applications and business models that could utilize social bookmarks for search in the Web or for other purposes.

referenCes Amitay, E. Har’El, N., Sivan, R., & Soffer, A. (2004). Web-a-where: Geotagging Web content. SIGIR 2004, 273-280. Baeza-Yates, R., Castillo, C., & Saint-Jean, F. (2004). Web dynamics, structure, and page quality. In M. Levence & A. Poulovassilis (Eds.), Web dynamics. Springer. Bao, S., Wu, X., Fei, B., Xue, G., Su, Z., & Yu, Y. (2007). Optimizing Web search using social annotations. In Proceedings of the 16th International World Wide Web Conference, Banff, Alberta, Canada. Damianos, L., Griffith, J., & Cuomo, D. (2006). Omni: Social bookmarking on a corporate Internet. The MITRE Corporation. Golder, S. A., & Huberman, B. A. (2006). The structure of collaborative tagging systems. Journal of Information Science, 32(2), 198–208. doi:10.1177/0165551506062337 Grosky, W. I., Sreenath, D. V., & Fotouchi, F. (2002). Emergent semantics and the multimedia Semantic Web. SIGMOD Record, 31(4), 54–58. doi:10.1145/637411.637420

257

Social Bookmarking and Web Search

Heymann, P., Koutrika, J., & Garcia-Molina, H. (2008). Can social bookmarking improve Web search? In Proceedings of the 1st ACM International Conference of Web Search and Data Mining. Stanford. Keller, R. M., Wolfe, R. R., Chen, J. R., Labinowitz, J. L., & Mathe, N. (1997). Bookmarking service for organizing and sharing URLs. In Proceedings of the 6th International World Wide Web Conference (pp. 1103-1114), Santa Clara, CA. Kleinberg, J. M. (1999). Alternative sources in a hyperlinked environment. Journal of the ACM, 604–632. doi:10.1145/324133.324140 Mandl, T. (2006). Implementation and evaluation of a quality-based search engine. In Proceedings of ACM Hypertext 2006 Conference (pp. 73-84). Marlow, C., Naaman, M., Boyd, D., & Davis, M. (2006). HT06, Tagging Paper, Taxonomy, Flickr, Academic Article, To Read. In Proceedings of ACM Hypertext 2006 Conference (pp. 31-40).

Wu, H., Zubair, M., & Maly, K. (2006). Harvesting social knowledge from folksonomies. In Proceedings of ACM Hypertext 2006 Conference (pp. 111-114). Wu, X., Zhang, L., & Yu, Y. (2006). Exploring social annotations for the Semantic Web. In Proceedings of the 15th World Wide Web Conference (pp. 417-426). Yu, P. S., Li, X., & Liu, B. (2004). On temporal dimension of search. In Proceedings of the 13th International World Wide Web Conference on Alternate Track Papers & Posters (pp. 448-449). Zhang, L., Wu, X., & Yu, Y. (2006). Emergent semantics from folksonomies: A quantitative study. Journal on Data Semantics, VI, 168–186. doi:10.1007/11803034_8

endnotes 1

Mathes, A. (2004). Folksonomies-cooperative classification and communication through shared metadata. Computer Mediated Communication. Millen, D. R., Feinberg, J., & Kerr, B. (2006). Doger: Social bookmarking in the enterprise. In Proceedings of the SIGCHI Conference on Human Factors in Computer Systems (pp. 111-120).

2

3 4 5 6 7

Mislove, A., Gummadi, K. P., & Druschel, P. (2006). Exploring social network for Internet search. In Proceedings of the 5th Workshop on Hot Topics in Networks. Page, L., Brin, S., Motwani, R., & Winograd, T. (1998). The pagerank citation ranking: Bringing order to the Web. (Tech. Rep.). Stanford Digital Library Technologies Project. Strutz, D. N. (2004). Communal categorization: The folksonomy. [Content representation.]. Info, 622.

258

8 9

10

Del.icio.us: http://del.icio.us De.icio.us – Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/Del.icio.us) obtained on March 3, 2008 Furl: http://www.furl.net Simpy: http://www.simpy.com Cloudalicio.us: http://cloudalicio.us Hatena Bookmark: http://b.hatena.ne.jp Google Toolbar: http://toolbar.google.com/ T4/index.html Yahoo! Japan: http://www.yahoo.co.jp In some cases same tags are listed several times, since there may be several words used to express the same meaning in Japanese. OpenSocial: http://code.google.com/apis/ opensocial

Social Bookmarking and Web Search

APPendiX We have created two query sets. As a first query set, we collected 50 keywords that gained highest usage on Goo10 on each month from January 2006 to September 2007. Goo1 is one of popular search engines in Japan. After removing duplicates (i.e. same queries that appeared within 2 or more months) we obtained 806 queries. As the second query set we collected frequent and recent tags that were used on Hatena Bookmark at the same time. We obtained 531 tags.

259

260

Chapter 15

Social Tagging: Properties and Applications Yong Yu Shanghai Jiao Tong University, China Rui Li Shanghai Jiao Tong University, China Shenghua Bao Shanghai Jiao Tong University, China Ben Fei IBM China Research Lab, China Zhong Su IBM China Research Lab, China

AbstrACt Recently, collaborative tagging Web sites such as Del.icio.us and Flickr have achieved great success. This chapter is concerned with the problem of social tagging analysis and mining. More speciically, we discuss ive properties of social tagging and their applications: 1) keyword property, which means social annotations serve as human selected keywords for Web resources; 2) semantic property, which indicates semantic relations among tags and Web resources; 3) hierarchical property, which means that hierarchical structure can be derived from the lat social tagging space; 4) quality property, which means that Web resources’ qualities are varied and can be quantiied using social tagging; 5) distribution property, which indicates the distribution of frequencies of social tags usually converges to a power-law distribution. These properties are the most principle characteristics, which have been popularly discussed and explored in many applications. As a case study, we show how to improve the social resource browsing by applying the ive properties of social tags.

introdUCtion With the rapid development of new technologies, both ordinary users and service providers are ex-

periencing the coming wave of the next-generation Web. As a representative, tagging based services have achieved a significant success. Services like Del.icio.us (http://del.icio.us), Flicker (http://www. flickr.com), and Technorati (http://Technorati.com),

DOI: 10.4018/978-1-60566-384-5.ch015

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Social Tagging

enable users to annotate and categorize Web resources, e.g., Web pages, photos and blogs, with freely chosen words. Taking the famous social bookmarking service, Del.icio.us, as an example, the service allows users to collect and annotate Web pages with one-word descriptors, which are also known as social tags or social annotations (in this chapter, we use the terms “annotation” and “tag” interchangeably). The social annotations assigned to bookmarks can help users organize their collected Web pages. Social annotations are a little bit like keywords or categories, but they are chosen by the users, and they do not form a hierarchy. Users can assign as many tags to a bookmark as they like and rename or delete the tags later. So, tagging can be much easier and more flexible than fitting users’ information into predefined categories or folders. In 2004, Thomas Vander Wal named these services “Folksonomy”, which came from the terms “folk” and “taxonomy” (Smith, 2004). Social annotations from tagging based Web sites are increasing at an incredible speed. Millions of Web users with different backgrounds are using these services to annotate their favorite Web resources. For example, Del.icio.us has more than 1 million registered users soon after its third birthday, and the number of Del.icio. us users has increased by more than 200% in the past nine months (Bao et al, 2007). Mathes (2004) attributes the success of these services to the following reasons:

3)

4)

The large amount social annotations are not only simple tags for organizing contents but also are useful in sharing information within a social environment. The power of social annotations is that aggregation of information provided by group of users form a social distributed environment. In this chapter, we will summarize current research on social tagging services and discuss how to aggregate the knowledge from social annotations. Specifically, we present a detailed analysis of five main properties of social annotations: 1)

1)

2)

Low Barriers to Entry: The freely chosen keywords enable users - not just professionals - to participate in the system immediately without any training or prior knowledge. Additionally, annotating Web resources is easy in terms of time, effort and cognitive costs. Feedback and Asymmetric Communication: Feedback is immediate, which leads to a form of asymmetrical communication between

users through metadata. The users of a system are negotiating the meanings of the terms through their individual choices of tags to describe documents for themselves. Individual and Community Aspects: Individuals have an incentive to tag their materials with terms that will help them organize their collections in a way that they can find these items later. The services are designed to share materials. Users can contribute to the system and other users by sharing the tags and associated resources. Unanticipated Uses: While most tags developed at Flickr and Del.icio.us have a concrete focus on subject categorization, there are tags being used in some unexpected, interesting ways that reflect communication and ad-group formation facilitated through metadata.

2)

3)

Keywords property: tags, used to organize personal collected Web resources, often describe and summarize the content and usage of Web resources perfectly. Semantic property: semantics of each tag can be generated from the associations between tags and Web resources, since similar tags are usually used to annotate similar resources. Hierarchical property: although tags are flat keywords associated with corresponding resources for personal use, the hierarchical

261

Social Tagging

4)

5)

structure among tags can be derived by aggregating the tagging information in a social environment. Quality property: the frequencies of tags are different, which suggests that the popularity of tags and the quality of their associated Web resources are varied. Distribution property: the distributed collaborative tagging system seems to be a complex system; however, the distribution of frequency of tags stabilizes and is often converged to a power-law distribution.

These properties are the essence of social annotations, which make them as novel data used in various areas. For example, much work has been done on exploring the social annotations for Web search and browsing (Hotho et al, 2006; Bao et al, 2007; Xu et al, 2007; Heymann et al, 2008; Li et al 2007; Xu et al, 2008), event detection (Rattenbury et al, 2007; Dubinko et al, 2007), enterprise search (Dmitriev et al, 2006;), semantic Web (Wu et al, 2006; Mika 2005; Zhou et al, 2007) etc. As a case study, we discuss the problem of how to improve browsing experience with the help of social annotations. As more and more people are involved in tagging services, the social annotations are not only a method of organizing contents to fascinate the users who build it, but also a navigation mechanism for users to discover interesting resources by sharing annotations. Although social tagging became a new interface for surfing the web, directly browsing in such types of services suffers from several problems since the limitation of current social tagging services. For example, the uncontrolled vocabulary and synonym problem make user hard to locate their desired resources. In this chapter, we will give an effective approach to browse Web pages via utilizing above properties of social annotations.

262

bACKgroUnd Early discussions of the social annotation can be found in (Mathes, 2004; Quintarelli, 2005; Smith 2004). Most of them are Web blogs and mailing list. They initiated the idea that sharing tags can lead to the concept known as “Folksonomy”. The term is a fusion of the words folks and taxonomy and first appeared in an information architecture mailing list (Smith 2004). It means the wide-spreading practice of collaborative categorization using freely chosen keywords by a group of people cooperating spontaneously. Quintarelli (2005) suggested that folksonomy is an information organizing tool which provides an approach to address Web-specific classification issues. Author also argued that “Folksonomies are not simply visitors tagging something for personal use: they are also an aggregation of the information that visitors provide. The power of folksonomy is connected to the act of aggregating, not simply to the creation of tags”. Hammond et al (2005) gave a good review of available social bookmark tools and Lund et al (2005) took the Connotea as an example to illustrate how a social bookmark tool works. Following initial discussions on the Web blogs and mailing list, research papers appeared, focusing on analyzing and examining properties of social annotations. Golader & Huberman (2006) gave a specific analysis of the use patterns of these collaborating tagging systems in both static and dynamic aspects. They found regularities in user activities, tag frequencies, kinds of tags used, bursts of popularity in bookmarking and a remarkable stability in the relative proportions of tags within a given URL. Halpin et al (2007) examined the dynamic aspect of collaborating system empirically, they concluded that distribution of tag frequencies in Del.icio.us converges to a stable Power Law distribution and thus emergent a coherent categorization schema.

Social Tagging

Besides analyzing the property of social annotations and their use patterns, researchers began to focus on how to make use of social annotations, which keep an incredible increasing speed. After Mika (2005) initialized the idea that community based semantics can emerge from social annotations, Wu et al (2006) used a probabilistic generative model to obtain a flat emergent semantics hidden behind the co-occurrence of three types of data (tags, users and resources). Based on their emergent semantic model, they also proposed a framework for semantic search. Zhou et al (2007) continued the work of emergent semantics and constructed a hierarchical semantic structure by using an unsupervised model. Besides the applications in the Semantic Web area, researchers have also adopted social annotations to other areas like Web search (Hotho et al, 2006; Bao et al, 2007; Heymann et al, 2008; Yanbe et al, 2007; Xu et al, 2007; Noll & Meinel, 2007; Xu et al, 2008), blog classification (Brooks & Montanez, 2006) and Web browsing (Li et al, 2007). Hotho et al (2006) proposed Adapted PageRank and FolkRank to find communities within the folksonomy. Bao et al (2007) used the annotation for optimizing Web search. They proposed two algorithms for estimating the similarity between annotations and Web queries, and the quality of Web page using social annotations. Xu et al (2007) smoothed the estimation of language model for information retrieval by fully exploring social annotations. Noll and Meinel (2007) proposed Web search personalization via social bookmarking and tagging. Xu et al (2008) further improved the performance of personalized search by exploring the tagging structures. Brooks and Montanez (2006) analyzed the effectiveness of tags for classifying blog entries and argued that there is a topical hierarchy among tags. Li et al (2007) discussed the problem of how to browse resources of a social tagging system. Researchers from Yahoo also discussed other new research problems and solutions, when they are trying to make annotations on the Flickr more useable, friendly and powerful. Dubinko

et al (2006) proposed and solved the problem of visualizing the evolution of tags within Flickr online image sharing service. They gave an efficient algorithm for processing the large data in real time. Their work focused on discovering the hot images and tags in a pre-defined time interval. Recently, Rattenbury et al (2007) gave a solution to the problem of extracting the locations and events from Flickr tags. Researchers from industrial areas also extended the collaborating system to an enterprise environment. They suggested that social tagging system was not only limited in Web environment, but also benefited enterprise environment. Millen et al (2006) discussed how to design a social annotation tool in an enterprise environment with special issues like security and priority issues. Dmitriev and Eiron (2006) implemented a social annotation tool within an enterprise environment and showed that annotations had improved the search efficiency based on their experiments.

ProPerties of soCiAl tAgging Among many successful collaborative tagging systems, Del.icio.us is the representative of such sites. Following we will take Del.icio.us as an example to illustrate our idea.

data Modeling A typical annotation activity in Del.icio.us consists of four elements: a user, a webpage, a tag, and a tagging time. We define an annotation activity as a quadruple: (User, Page, Tag, Time). While a full quadruple captures the complete information of an annotation activity, we usually discard some roles and simplify the data modeling for different applications. For example, simplification (User, Page, Tag) is explored by Mika (2005) and Wu et al (2006) for emergent semantics; simplification (Page, Tag) is popularly used in page centric applications (Bao et al 2007; Li, et al 2007; Xu

263

Social Tagging

et al 2007); simplification (User, Tag) has also been explored for search personalization (Noll and Meinel 2007; Xu et al 2008); (Time, Tag) is frequently used in the time related applications such as visualizing tags over time (Dubinko et al, 2006) and time-related social browsing (Li et al 2007).

Annotation Properties Based on above data modeling, many properties of social annotations have been proposed and discussed. Here, we will give five key properties of social annotations and their applications.

Keywords Property “Individuals have an incentive to tag their materials with terms that will help them organize their collections in a way that they can find these items later” (Mathes 2004). So people often use annotations which may indicate categories of a Web page, describe abstracts of webpage content, or specify the usage of webpage. Here, we give the first property of social annotations. Property 1: (Keyword Property) “Social annotations are usually good summarizations or precise keywords of the associated resources chosen by their readers or users.” To illustrate the property, we first give some examples in Table 1 to investigate whether annotations are matched with topics of the corresponding Web page. We can find that most of

annotations are good summarization for a given Web page. Comparing with keywords extracted from Web page automatically, annotations, served as the human edited keywords for Web pages, have several advantages. Especially keywords extraction algorithms will fail in some cases. The first case is that a Web page contains a little text. e.g., keywords of Google’s homepage only contain “Google”, on the other hand, annotations in Del.icio.us for Google provides a set of meaningful words like “internet”, “search” and “searchengine”, which are not appeared in the page. The second case is that a Web page is changing dynamically, such as a homepage of a news site, a personal blog. Keywords extracted from Web pages can not reflect the main function of Web pages. E.g., the content of homepage of Del. icio.us is changing every minute and most of the texts are related to the pages tagged recently. So the keywords extracted from Del.icio.us contain words like “iPod” are not related to Del.icio.us. The last case is that for a Web page with lots text information, e.g. http://reference.sitepoint.com/ css. Automatically algorithms extract keywords from each part, but annotations summarize the content well with human knowledge. The keyword property is the most fundamental property of social annotations. This property and its advantages have also been discussed and applied in several applications. Brooks and Montanez (2006) directly used the annotations as a representation of Web pages for clustering. Bao et al (2007) used annotations as additional metadata of pages for

Table 1. Pages with their annotations and keywords Web Pages

Annotations

Keywords

http://www.ebay.com

shopping; ebay; auction; online; Business

Ebay company; ebay inc; ebay new used;

http://www.google.com

google; search; searchengine; search_engine; internet

Google

http://del.icio.us

Del.icio.us; social; web2.0; bookmarks; delicious

tags; del icio.us; posted; ndash, ipod

http://reference.sitepoint.com/css

css; reference; design; web ; webdesign;

sitepoint; css reference; properties; css layout;

Note. It shows the Web pages with top 5 social annotations as well as several automatically extracted key phrases.

264

Social Tagging

Web search. If a query matches any annotation of a page, then the page will be related to the query with high probability. Xu et al (2007) used annotations as a new representation of webpage to smooth the estimation of language model.

Semantic Property Due to asymmetrical communication between users in the social tagging systems, different tags assigned to the same web page are usually semantic related. Below, we give the second property of social annotations: Property 2: (Semantic Property) “Similar tags will annotate similar pages and similar pages will be annotated by similar tags in a social annotation environment. Then, the semantic of a tag can be reflected by resources which it tagged and vice versa.” To derive the semantic properties of social tags, we take the simplified data modeling of (Page, Tag) as discussed in the data modeling section. Then the data in Del.icio.us can be modeled as an m×n adjacency matrix MTP of tags and Web pages, where m and n are the number of tags and Web pages respectively. Each mij denotes the number of users who annotate the jth Web page with the ith tag. Given the matrix M, a tag can be represented as a row vector Ti (p1, p2, … pn) of M, and a webpage can be represented as a column vector Pi (t1, t2, … tm) of M. Based on the co-occurrence between tags and pages, there are several approaches to measure the semantic relationship between tags. One simple symmetric measurement is as follows: Sim(ti , t j ) = cos(Ti ,Tj ) ,

where Ti and Tj are tag vectors corresponding to tags ti and tj, respectively. Linguistic features can also be used for calculating Sim(ti, tj). When tags are freely assigned to

the relative URL, tags are used in various forms, such as the plural form and gerundial form. For example, “Programs”, “Programming” and “Program” all exist in the annotation data. Additional weight is added to Sim(ti, tj) if two terms share the same etyma after stemming. Besides, if the two terms share the etyma after eliminating the external punctuations, a lighter additional weight is added to the Sim(ti, tj) score. We demonstrate a similarity graph of several tags in the Figure 1 by simply using cosine similarity. The length of an edge between two tags indicates strength of their relationship. The similarity calculation is simple but effective. It can find meaningful related tags of a given tag. For example, “Chat” has strongly related tags “im” and “jabber”. “Movie” has its synonymous words “film” as its related tag. There are some other approaches for exploring the link structure of tags to estimate their similarity or semantics, such as KL- divergence used in Zhou et al (2007); Separable Mixture Model introduced by Wu et al (2006) and SocialSimRank provided in Bao et al (2007). The semantics property has been the most important property. Applications of this property come up with initial discussions of social annotations. Researchers in the semantic Web area are expecting that semantics will emerge from social annotations in a social negotiated way. Mika (2005) built a light weight ontology and Wu et al (2006) derived the emergent semantics based on semantic property of social annotations.

Hierarchical Property As a social classification method, social annotations drop the classical hierarchical classification mechanism, since it was introduced in Del.icio. us, and that makes social annotations easy and popular. Quintarelli (2005) and Mathes (2004) both argued that the tagging space is a flat space and a hierarchical representation of topics does not reflect the associative nature of social anno-

265

Social Tagging

Figure 1. Tag similarity graph of several tags

tations. Golder and Huberman (2006) stated that the different expertise and purpose of tagging participants may result in tags at various levels of abstraction to describe a resource. For example, a photo can be tagged at the “basic level” by “cat”, at a super ordinate level by “animal” or at various subordinate levels by “Persian cat”, “Felis silvestris cats” or “longhair Persian”. Given discussions above, we give the third property of social annotations. Property 3: (Hierarchical Property) “There is no neat tree structure like taxonomy or human built ontology with rigid hierarchies and predefined categories with clear boundaries for social annotations, but annotations used in social tagging services do locate in different semantic levels.” So we assume that there is an implicit semihierarchy behind the link structure of social annotations. Several methods have been proposed for constructing such a structure. Brooks and Montanez (2006) used an existing agglomerative clustering algorithm and construct a binary tree structure for annotations. Li et al (2007) used a machine learning method to identify sub-

266

relationship between two tags based on several statistical features. Zhou et al (2007) explored the hierarchical semantics of semantic Web using an unsupervised probabilistic model. Although their models are different, the key idea behind, which is used to measure whether a tag is a sub-tag of another tag or whether a tag is a representative tag of a cluster, are the same. Tags located at different semantic levels have different content coverage. A tag in a high semantic level covers more resources than its sub-tags. For example, there are more Web pages associated with “Web” than those associated with “Google”, and Web pages associated with “Google” are more than those associated with “Googletips”. The hierarchical relation can be derived from their coverage relation between two tags, which is defined as below: Coverageij =

P (ti ) P (t j )

;

where P(ti) denotes the number of pages tagged with tag ti.

Social Tagging

Table 2 and Figure 2 show some results of hierarchical structure derived from Li et al (2007)’s algorithm and Zhou et al (2007)’s algorithm, respectively. Result above implies that the algorithms are able to organize a hierarchical structure among tags as people think in their daily life. For example, when the user clicks “science”, algorithm in (Li et al, 2007) is able to generate a series of sub categories such as “math”, “physics”, “psychology”, etc.; which are meaningful and distinguishable. Although the hierarchical structure of social annotation is well matched with people’s common knowledge, it is hard to specify the relation between the high level tag and low level tags. Apparently, the relationship between “MIT” and “Science” is different from that between “Physics” and “Science”. Some tags share “is a” or “is kind of” relationship. For example, “RPG” is a kind of game. And some share a “concept related” relationship, for example, “hotel” and “transport”. So we can make a conclusion that although the hierarchical structure can be obtained from social annotations, it is still different from the structure in a pure ontology and it is difficult to obtain a hierarchical structure with specific sub relationship.

Quality Property Today, ranking methods serve as the key part in Web search. The existing ranking methods have explored many resources for measuring pages’ quality. The estimation of PageRank (Page et al, 1998) is based on the link graph of Web pages and fRank (Richardson et al, 2006) is calculated based on many features including click through data. These ranking algorithms reflect the Web page creators’, or the search engine users’ point of view. We think annotations reflect the quality of Web pages from annotators’ perspective. Here we give the fourth property of social annotations. Property 4: (Quality Property) “Since high quality Web pages are usually popularly anno-

tated, the quality of annotated Web resources can be derived from the number of assigned social annotations.” Intuitively, if a page is bookmarked by many users, it could be considered as popular page, and if a tag is used frequently, it may be a hot tag. Therefore, three simple principles can be applied to page ranking and browsing. 1.

2. 3.

Pages are ranked by the number of bookmarked frequencies or the number of users who collected it. Tags are ranked by their frequencies. Users are ranked by the number of pages they bookmarked.

Figure 3 shows the average counts of annotations and annotators over Web pages with different PageRank values, and for the “Unique Annotation” line, the value means the count of annotations that are different with each other. From the figure, we can conclude that in most cases, the page with a higher PageRank is likely to be annotated by more users with more annotations. To have a concrete understanding, we list top 10 popular Web pages, hot tags of the domain “Java” in Table 3. The result is promising. The quality property has been discussed in many applications. Bao et al (2007) and Hotho et al (2006) used iterative algorithms on the link graph to obtain a statistic rank for a Web page. Then the ranking combined with other features is used to optimize the performance of web search. Other applications, e.g., visualizing tags over time (Dubinko et al, 2006), also used the quality feature for getting a hot topic during a specific time interval. So quality feature provides a static ranking for Web pages, tags and other objects involved in tagging system from social perspective.

Distribution Property Nowadays, millions of users are using the collaborative tagging system without centrally principles.

267

Social Tagging

Table 2. Results of hierarchical relationship between social annotations Programming

Music

Science

Microsoft

AJAX JavaScripts Ruby rails PHP Python Java Framework C Cpp Dhtml Lisp Perl

Bittorrent Torrents Ipod Radio MP3 Itunes Guitar Chords Sound Soundfx Player Songbird Indie Drm Lyrics song

Health Sleep Math Mathematics Physics Quantum Psychology Brain Space Astronomy Algorithms MIT Biology Lectures sicp Evolution

XP Tweaks Excel Word Writely asp.net dotnet XBOX MSN WindowsLive Outlook Boot Bootdisk Spyware Vista Longhorn

Arts

Basketball

Book

Computer

Graffiti Streetart Museum mus Knots topology Poetry Artistis Painter

ESPN Fox Autism Dallas NBA

Lisp Literature ebook Audiobooks Amzaon Scheme Sicp

Developers IE favorites Algorithms comupeterscience Spyware Adware

Note. Results demonstrated in the paper (Li et al, 2007). For each of the four concepts in the first line, we listed only ten subordinate concepts and for each of the rest of concepts, we listed only five subordinate concepts.

Figure 2. Hierarchical structure derived from social annotations

The tags are assigned freely and openly. Researchers suspect that a collaborative tagging system exhibits features like a complex system. Here we give the fifth property of social annotations.

268

Property 5: (Distribution Property) “In the collaborative tagging system, the distribution of tag frequencies stabilizes and converges to a power low distribution.”

Social Tagging

Figure 3. Average count distribution over PageRank

Table 3. Popular Web pages and hot tags in the category “Java” Popular Web pages

Hot Tags

http://www.eclipse.org http://java.sun.com http://pmd.sourceforge.net http://www.lowagie.com/iText http://www.onjava.com http://ant.apache.org http://argouml.tigris.org http://java.sun.com/docs/books/tutorial http://jakarta.apache.org http://www.caucho.com

java programming development software eclipse tools opensource Java .imported ide

Figure 4 and Figure 5 demonstrate this property. X axis represents tags in the order of their counts and Y axis represents the counts of the tags. Figure 4 illustrates the distribution of the counts of tags which are associated with a certain URL. We discover that people always use the most popular tags to annotate the Web page and unpopular tags are barely applied to annotate. Figure 5 demonstrates the distribution of the counts of tags associated with the whole Del.icio.us data. We find out that the popular tags are frequently and extensively used in the whole Del.icio.us data although there are thousands of tags used. Both figures show that the distribution of tag frequencies stabilizes and converges to a power low distribution. Here stability and convergence mean “tagging eventually settles to a group of tags that describe the resource well and new users mostly reinforce existing tags in the same frequency as they have been given in the stable distribution, but not mean that user stop tagging resources”(Halpin et al, 2007).

This property suggests popular tags and Web pages play an important role in social tagging service. People use popular tags to annotate Web pages and the popular pages are annotated by the majority of tags. This property has also been explored for many applications. For example, Li et al (2007) introduced sampling method and applied to the social browsing system. This property also suggests that although people assign the tags freely, the social annotation used over time would lead to a convergent classification schema that could that be formalized into social knowledge.

A CAse stUdy: browsing with soCiAl AnnotAtions Now, we consider a typical problem—how to browse large scale social data based on above discussion. Currently, there are two main methods of helping users to seek the information through

269

Social Tagging

Figure 4. Tag counts distribution over three specific Web pages

Figure 5. Popular tag counts distribution over Del.icio.us

annotations. The first one is the keyword-based search, which is the most common way for finding information on the Web. Systems of this type will display all contents associated with the given annotation. The second one is a method called tag cloud view (Delicious, Tag). It usually displays the social annotations alphabetically with different font sizes and colors indicating their popularities. Selecting a specific annotation will generally lead to the keyword search with the selected annotation as input. Compared with the direct searching method, tag cloud provides a better user interface for browsing the popular social annotations. However, the drawbacks of the methods are obvious, especially when the scale of social annotations is quite large: 1) Contents and annotations are simply matched by

270

the literal representations instead of the semantic meanings. The synonymy and ambiguity problems exist in these systems. The synonymy problem arises when different annotations have similar meanings. As a result, simple matching cannot find all the relevant resources. The ambiguity problem arises when an annotation has several different meanings. It will introduce noisy pages. Both of these problems influence the performance of existing browsing systems. 2) The returned results are not categorized and ranked. It is difficult for users to browse thousands of returned results to find their interested resources. To overcome the above problems, we propose a new algorithm to help users browse the large scale social annotation data. All five properties of social annotations are exploited for effective browsing:

Social Tagging

1.

2.

3.

4.

5.

Keyword Property: Keyword property is the basis of social browsing. It assures the accurate association between the social annotations and Web resources. Given an information need, users can find the related resources by browsing the meaningful social annotations. Semantic Property: We use the similarity calculation method to measure semantic similarity among annotations. Each annotation is assigned with a semantic concept consisting of the strong related annotations, thus the synonym problem can be controlled by matching the semantic concept of the selected annotation. For example, assuming one selects the annotation “book”, the resources annotated by either “book” or “books” will be returned since “books” also appears in the concept of “book”. Hierarchical Property: We utilize method discussed in the structure property of the social annotations to present concepts in different semantic levels and we build a hierarchical structure for social annotations. When annotations are organized in this way, users can locate their desired resources more easily. Quality Property: The quality property is applied to provide a popular rank of pages, which make user find the popular pages easily. Distribution Property: The time cost for browsing increases with the growth of the size of social annotations. The algorithm samples most frequent tags and Web pages, which are applied for calculating the similarity score and hierarchical structure, based on the distribution property.

Now we give the explanations of the algorithm as shown in Table 4. The corresponding property used by the browsing algorithm at each step is also presented in the right column. More details of the algorithm can be found at (Li et al, 2007).

From step 1-1 to 1-2, the algorithm initializes the first view of annotations. NT, NP, NC, and NCT denote the number of tags, pages, clusters and tags in each cluster. In our experiment, these parameters are set to 2000, 2000, 20, and 5, respectively, which means the top 100 tags distributed in 20 clusters on 2000 most frequent tags and pages are presented to the users as the default browsing interface. These popular tags, which are associated with a large number of resources, are selected as the roots in hierarchical browsing. When users select a tag as the entrance to annotation browsing, the algorithm outputs its related resources and a set of annotations as subtags. The related resources are ranked according to their popularity. Users can iteratively select any annotation from the displayed sub-tags for further exploration. The iterative process consists of four components as follows: Tag selection (from 2-1 to 2-3): to provide a semantic browsing, the algorithm takes the selected tag as a semantic concept which consists of several highly related tags using cosine similarity calculation. The user’s path from the root to the current annotation forms a set of concepts and specifies the user’s interests. The pages and tags related with the concept set will be selected. Pages and tags ranking (step 3): this step ranks the pages according to their frequencies and ranks the tags according to number of users. Page and tag sampling (step 4): this is an optional step. We introduce a sample mechanism to sample tags and pages which match the specified concept set. The application of sampling assures that the algorithm is always running on a data set with controlled size. Sub-Tag Generation (step 5-1): we develop a set of features and rules to find the sub-tags of the current tag. The resources of the current tag can be further classified into concepts of these sub-tags. Similarity based Clustering (step 5-2): a clustering algorithm is used to find a proper number of clusters from the generated sub-tags in the

271

Social Tagging

Table 4. Social annotation browsing algorithm Algorithm 1 Social Browsing Algorithm Applied Properties An empty concept set SC

Input Step 1

Output the initial view of annotations

1-1

Select top NT tags ST and top NP Pages SU

Distribution Property

1-2

Social Clustering (ST, SP, NC),

Semantic Property

Return

NC clusters CT with top NCT tags in each cluster

Loop

User select a tag Ti

Step 2

Concept matching

2-1

Calculate strong related tags to Ti to construct concept Ci

2-2

Add Ci to SC

2-3

Select related pages set SPi and related tag set STi

Keyword Property

Step 3

Ranking related pages set SPi and related tag set STi

Quality Property

Step 4

Get sample pages set SPs and sample tag set STs

Distribution Property

Step 5

Hierarchical Browsing

5-1

Calculate sub-tag set SSTi w.r.t. concept Ci

Hierarchy Property

5-2

Social Clustering (SSTi, SP, NC), obtained CTj

Semantic Property

Return

Top NCT tags in each cluster

IF

Termination condition Satisfied; Return

ELSE

Loop

previous step. Then the sub-tags in each cluster are presented to the user. In step 5, the user can click one of these presented sub-tags to further seek his desired resources. Figure 6 gives a snapshot of the system implemented based on above algorithm. The page behind is the initial interface of the system. It contains popular annotations distributed in different clusters. The size of each annotation indicates its popularity. The page in the front is the result after a user selects the annotation “programming”. On the right side is a set of pages related to the current annotation. Each line on the left side is a sub-category of the current annotation, which consists of several related annotations. Users can click the tag on the left side to further investigate that category.

272

Semantic Property

fUtUre reseArCh direCtions The collaborative tagging system will be more and more popular and important with mass of diverse data type are appearing in the web, such as photos, videos and products. These data, with less text, are hard to search and find. Tags will serve as the keywords description that help search engine to index these data with meaningful keywords, and help users to search and browsing these resources with tags easily. So we assume that besides used in the bookmarking service, photo sharing service and the blog service, collaborative tagging function will be plugged into much more services, such as online shopping Web site, video sharing Web site and online news Web site. Finally, tags will make the Web much easier to surf. However, as we have observed that the social annotations do benefit many Web applications, there are still several problems to further address:

Social Tagging

Figure 6. Interface of social browsing system

Annotation Noises: While useful as it is, social annotations may also contain noises. Brooks and Montanez (2006) categorized the social annotations into three basic strategies for tagging: 1) Annotating information for personal use; 2) Placing information into broadly defined categories; and 3) Annotating particular articles so as to describe their content. Given a page, the list of most popular tags is usually mixed with different categories of annotations described above. The keywords property is mainly derived from the 2nd and 3rd categories of social annotations. To guarantee the keywords property of social annotations, Wu et al (2006) filtered out the social annotations which appear less than 10 times. Bao et al (2007) and Xu et al (2007) manually identified a list of personalized annotations e.g. “toread”, “todo”, to filter out the 1st category and splitted the combined annotations, e.g. “searchengine”, “search.engine”, with the help of WordNet (http:// wordnet.princeton.edu/). Annotation Coverage: While social annotations are useful, many applications of social annotations suffer from the annotation sparseness problem. The Web resources with social annotations are still limited on the World Wide Web. Taking Del.icio.us as an example, more than 1 million Web users collected over 10 million Web pages

with millions of social annotations. However, compared with 1.173 billion Web users (http:// www.internetworldstats.com/stats.htm) and about 30 billion (http://www.boutell.com/newfaq/misc/ sizeofweb.html) Web pages on WWW, the ratio of both social annotators and annotated Web pages still remain less than 0.1%. One possible way to improve the coverage is to propagate the social annotations via Web sitemaps and hyperlinks (Bao et al 2008). Annotation Spamming: Initially, there are few ads or spams in social annotations. However, as social annotation becomes more and more popular, the amount of spam could drastically increase in the near future and spamming will become a real concern for the novel properties discussed above. Many applications discussed in this chapter take the assumption that the social annotations are good summaries of Web pages, so malicious annotations have a good opportunity to harm the quality. There are generally two ways in preventing the spam annotations. 1) Manually or semi-automatically deleting spam annotations and punishing users who abuse the social annotation system. Such work usually relies on service providers; 2) Filtering out spam annotations by using statistical and linguistic analysis before the using the social annotations.

273

Social Tagging

ConClUsion As the boosting increasing of social annotations, researchers have applied social annotations to lots of areas, such as improving the performance of Web search, Web page classification, Semantic Web and ontology extraction. In this chapter, we mainly discussed and analyzed five principle properties of social annotations based on a survey of existing research work. These properties make social annotation as a novel data on the Web and more useful in Web applications. Furthermore, we gave a case study to illustrate the usage of these properties by solving a specific problem, i.e. how to browse large scale Web pages with the help of social annotations.

referenCes Bao, S., Wu, X., Fei, B., Xue, G., Su, Z., & Yu, Y. (2007). Optimizing Web search using social annotations. In Proceedings of the 16th International Conference on World Wide Web (pp. 501-510). New York: ACM. Bao, S., Yang, B., Fei, B., Xu, S., Su, Z., & Yu, Y. (2008). Boosting social annotations using propagation. In Proceedings of the ACM 17th Conference on Information and Knowledge Management. New York: ACM. Brooks, C. H., & Montanez, N. (2006). Improved annotation of the blogosphere via autotagging and hierarchical clustering. In Proceedings of the 15th International Conference on World Wide Web (pp. 625-632). New York: ACM. Delicious, T. (n.d.). Del.icio.us tag cloud view. Retrieved on March 29, 2008, from http://del. icio.us/tag/ Dmitriev, P. A., Eiron, N., Fontoura, M., & Shekita, E. (2006). Using annotations in enterprise search. In Proceedings of the 15th International Conference on World Wide Web (pp. 811-817). New York: ACM.

274

Dubinko, M., Kumar, R., Magnani, J., Novak, J., Raghavan, P., & Tomkins, A. (2006). Visualizing tags over time. In Proceedings of the 15th International Conference on World Wide Web (pp. 193-202). New York: ACM. Delicious, Help. (n.d.). What are tags? Retrieved on March 29, 2008, from http://del.icio.us/help/tags Golder, S. A., & Huberman, B. A. (2006). Usage patterns of collaborative tagging systems. Journal of Information Science, 32(2), 198–208. doi:10.1177/0165551506062337 Halpin, H., Robu, V., & Shepherd, H. (2007). The complex dynamics of collaborative tagging. In Proceedings of the 16th International Conference on World Wide Web (pp. 211-220). New York: ACM. Hammond, T., Hannay, T., Lund, B., & Scott, J. (2005). Social bookmarking tools (I) a general review. D-Lib Magazine, 11(4), 1082–9873. doi:10.1045/april2005-hammond Heymann, P., Koutrika, G., & Garcia-Molina, H. (2008). Can social bookmarking improve Web search? In Proceedings of the International Conference on Web Search and Web Data Mining. New York: ACM. Hotho, A., Jaschke, R., Schmitz, C., & Stumme, G. (2006). Information retrieval in folksonomies: Search and ranking. In Proceedings of 3rd European Semantic Web Conference (pp. 411-426). Springer. Lund, B., Hammond, T., Flack, M., & Hannay, T. (2005). Social bookmarking tools (II) a case study-Connotea. D-Lib Magazine, 11(4), 1082–9873. Li, R., Bao, S., Yu, Y., Fei, B., & Su, Z. (2007). Towards effective browsing of large scale social annotations. In Proceedings of the 16th International Conference on World Wide Web (pp. 943952). New York: ACM.

Social Tagging

Mathes, A. (2004, December). Folksonomies– cooperative classification and communication through shared metadata computer mediated communication. LIS590CMC Graduate School of Library and Information Science University of Illinois Urbana. Mika, P. (2005). Ontologies are us: A unified model of social networks and semantics. In Proceeding of 4th International Semantic Web Conference and the 2nd Asian Semantic Web Conference (pp. 522-536). Springer Press. Millen, D. R., Feinberg, J., & Kerr, B. (2006). Dogear: Social bookmarking in the enterprise. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 111120). New York: ACM. Noll, M. G., & Meinel, C. (2007). Web search personalization via social bookmarking and tagging. In Proceedings of 6th International Semantic Web Conference (ISWC) and 2nd Asian Semantic Web Conference (ASWC) (pp. 367-380). Springer. Page, S., Brin, R. Motwani, & Winograd, T. (1998). The PageRank citation ranking: Bringing order to the Web. (Tech. Rep.). Stanford Digital Library Technologies Project. Quintarelli, E. (2005, June). Folksonomies: Power to the people. Paper presented at the ISKO ItalyUniMIB Meeting, Milan, Italy. Rattenbury, T., Good, N., & Naaman, M. (2007). Towards automatic extraction of event and place semantics from flickr tags. In Proceedings of the 30th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval (pp. 103-110). New York: ACM. Richardson, M., Prakash, A., & Brill, E. (2006). Beyond PageRank: Machine learning for static ranking. In Proceedings of the 15th International Conference on World Wide Web (pp. 707-715). New York: ACM.

Smith, G. (2004, August 3). Atomiq: Folksonomy: Social classification. Retrieved on March 29, 2008, from http://atomiq.org/archives/2004/08/ folksonomy_social_classification.html Wu, X., Zhang, L., & Yu, Y. (2006). Exploring social annotations for the Semantic Web. In Proceedings of the 15th International Conference on World Wide Web (pp. 417-426). New York: ACM. Xu, S., Bao, S., Cao, Y., & Yu, Y. (2007). Using social annotations to improve language model for information retrieval. In Proceedings of the Sixteenth ACM Conference on Information and Knowledge Management. New York: ACM. Xu, S., Bao, S., Fei, B., Su, Z., & Yu, Y. (2008). Exploring folksonomy for personalized search. In Proceedings of 31st Annual International ACM SIGIR Conference on Research & Development in Information Retrieval (pp. 155-162). New York: ACM. Yanbe, Y., Jatowt, A., Nakamura, S., & Tanaka, K. (2007). Can social bookmarking enhance search in the Web? In Proceedings of the 7th ACM/IEEE Joint Conference on Digital Libraries (pp. 107116). New York: ACM. Zhou, M., Bao, S., Wu, X., & Yu, Y. (2007). An unsupervised model for exploring hierarchical semantics from social annotations. In Proceeding of 6th International Semantic Web Conference and the 2nd Asian Semantic Web Conference (pp. 680-693). Springer Press.

AdditionAl reAding Al-Khalifa, H. S. & Davis, H. C.2006. Measuring the semantic value of folksonomies. Innovations in Information Technology, 1-5

275

Social Tagging

Al-Khalifa, H. S., & Davis, H. C. (2007). Exploring the value of folksonomies for creating semantic metadata. [IJSWIS]. International Journal on Semantic Web and Information Systems, 3(1), 13–39. Begelman, G. Keller, P., & Smadja, F. 2006. Automated Tag Clustering: Improving search and exploration in the tag space, In Proceedings of WWW2006, Collaborative Web Tagging Workshop. Benz, D., Tso, K., Schmidt-Thieme, L. 2007. Supporting Collaborative Hierarchical Classification: Bookmarks as an Example, Special Issue on Innovations in Web Communications Infrastructure, Journal of Computer Networks Volume 51:16, 4574-4585 Catutto, C., Schmitz, C., Baldassarri, A., Servedio,V. D. P., Loreto, V., Hotho, A., Grahl, M., & Stumme. G. 2007. Network properties of folksonomies. AI Communications Journal, Special Issue on “Network Analysis in Natural Sciences and Engineering”. Cripe, B. 2007, Folksonomy, Keywords, & Tags: Social & Democratic User Interaction in Enterprise Content Management. An Oracle Business & Technology White Paper Lambiotte, R. & Ausloos., M. 2005. Collaborative tagging as a tripartite network. Technical report, 181-202 Merholz., P. 2004. Metadata for the Masses. http:// www.adaptivepath.com/publications/essays/ archives/000361.php Schachter, J. 2004. Del.icio.us about page. http:// del.icio.us/doc/about Schmitz, C., Hotho, A., Jaschke, R., 2006. Gerd Stumme Mining Association Rules in Folksonomies, Data Science and Classification, 261-270

276

Schmitz, P. 2006. Inducing ontology from Flickr tags In Proceedings of WWW2006, Collaborative Web Tagging Workshop. Specia, L., Motta, E. 2007. Integrating Folksonomies with the Semantic Web, The Semantic Web: Research and Applications, 624-639

Key terMs And definitions Folksonomy: A fusion of the words folks and taxonomy. It is a practice and method of collaboratively creating and managing tags to annotate and categorize all kinds of Web resources. Ontology: A formal representation of a set of concepts within a domain and the relationships between those concepts. It is used to reason about the properties of that domain, and may be used to define the domain Social Browsing: A browsing mechanism which leverages social data contributed by Web users. Social Search: A type of Web search method or algorithm which leverages all kinds of user interaction data, such as social tags, query logs. It is a promising research direction and successful search practice of combining human intelligence with computer algorithms Social Tagging Services: A method for Internet users to store, organize, search, and manage Web resources on the Internet with the help of tags Social Tags: A non-hierarchical keyword or term assigned to different Web resources. They are chosen informally and personally by Web users. The collection of tags becomes a folksonomy. Taxonomy: A kind of classification method which organizes all kinds of things into predefined hierarchical structure.

277

Chapter 16

Improving Cross-Language Information Retrieval by Harnessing the Social Web Diana Irina Tanase University of Westminster, UK Epaminondas Kapetanios University of Westminster, UK

AbstrACt Combining existing advancements in cross-language information retrieval (CLIR) with the new usercentered Web paradigm could allow tapping into Web-based multilingual clusters of language information that are rich, up-to-date in terms of language usage, that increase in size, and have the potential to cater for all languages. In this chapter, we set out to explore existing CLIR systems and their limitations, and we argue that in the current context of a widely adopted social Web, the future of large-scale CLIR and iCLIR systems is linked to the use of the Web as a lexical resource, as a distribution infrastructure, and as a channel of communication between users. Such a synergy will lead to systems that grow organically as more users with different linguistic skills join the network, and that improve in terms of language translations disambiguation and coverage.

introdUCtion In 1926, Bertolt Brecht was making the following suggestion about the utility of the radio: “a one sided” device “when it should be two“, “an apparatus, for mere sharing out” (Kaes et al., 1994, p. 616), that should be used not just for distribution, but also for communication. These visionary statements, apply about eighty years later to the new generation of services that transformed the web into a two-sided DOI: 10.4018/978-1-60566-384-5.ch016

“device” that not only distributes content, but serves as the newfound communication medium on a social, cultural, and economic level. Researchers, software developers, and enterprises found themselves challenged to create these new web services in order to accommodate our communication needs in the dynamic context of digital technologies. Our geographical boundaries disappear when connected to the Internet. Multilingual users meander through the web in search of facts, of answers to questions, in an attempt to discover new information, or just to keep alert on what goes on throughout the

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Improving Cross-Language Information Retrieval by Harnessing the Social Web

world. There are though, the language boundaries. They restrict access to the web in its entirety to users that have a good command of English, the predominant language of documents distributed on the web. Balance however is changing, and is very difficult to quantify accurately. In this context, the focus of the research field of CrossLanguage Information Retrieval (CLIR) is to develop systems that support users in how they locate and present answers to their queries from resources in other languages, regardless of the querying language. One of the pivotal aspects of CLIR is translation. This process entails mapping information encoded in the query-language to information encoded in the document-language. There are two approaches to translation: a) machine translation and b) user-assisted translation. The first is widely used and is supported by a variety of language resources from bilingual lists, to dictionaries, parallel corpora, wordnets, or interlingual indexes. Due to the quality of the language resources and the way they are used, current implementations of machine translation are far from perfect. Let us compare for example any two bilingual dictionaries for English to French and English to Japanese. These two resources will differ in coverage, source style (human or machine readable), number of translations alternatives given, form of entries (root or surface words), etc. These differences in characteristics will then need to be handled by separate parts of the translations component, making scaling to other languages challenging. Years of research, with mixed success, focused on developing methods that would perform well, independently of the pairs of languages the translations are run between (Levow et al., 2005), and one anticipated solution comes from interactive cross-language information retrieval (iCLIR). The iCLIR approach relies on the synergy between human and machine linguistic knowledge for improving the overall performance of a CLIR system. Currently, iCLIR systems have yet to fulfill

278

their mission, since generally human-computer interaction systems take a very long time to tune and test. At the same time, while iCLIR researchers are working on pinning down the best ways for users to help with the cross-language retrieval task, the web landscape has been flooded by a large number of web services that support the creation of online communities and collective knowledge pools (referred to as web 2.0 services). These communities are based on ad-hoc mechanisms for sharing information, communicating on events, stories, or things to-do, and overall facilitating each other to find and identify relevant resources. The latter is in fact the goal of any information retrieval task and the motivation for changing the setting for the users involved in assisting with a CLIR task and immersing them in the highly dynamic web community. In other words, can users collaborate online to share their linguistic knowledge in the context of information retrieval and how can it be achieved? This sets the premise for the explorations in this chapter. We will assess the potential to get users actively involved in interactive cross-language information retrieval, specifically, on how the human users can contribute to a CLIR task by: a) creating multilingual resources, b) annotating web resources with metadata in different languages, c) mapping a query to its appropriate translation, or d) marking relevant results obtained from a cross-language query. This chapter is organized as follows: The Background section describes the challenging aspects of existing iCLIR systems (MIRACLE, CLARITY, PanImages) and of global language projects (OmegaWiki, Wiktionary, Global WordNet Grid); We follow with the CLIR in the current web space section which looks into what are the required architectural components for integrating a CLIR system in the current web collaborative space, with recommendations on how to give more weight to a user’s context when performing the translation or the retrieval steps. The emerging

Improving Cross-Language Information Retrieval by Harnessing the Social Web

solution relies on creating personalized interlingual indexes (myILI) based on feedback received from the searcher and from using information extraction techniques on clusters of information a user associates with on the web (personomies). The myILI structures could be exchanged between users with the same domain interests. We argue that the implementation of such architecture is feasible due to a variety of frameworks and APIs suitable for syndicating and processing web data. Given that a systemic approach carries the danger of overestimating its benefits, we look into ways of evaluating this approach and we present some supporting evidence from the iCLEF 2008 track (Evaluating iCLIR systems). We conclude with Future Research Directions and Conclusions, a series of reflections on the social dimension and the global impact of truly multilingual information access systems can have.

bACKgroUnd This section gives a general overview of the components of a cross-language information retrieval system, existing examples of interactive CLIR systems (MIRACLE, CLARITY, PanImages), and their dependency on language resources, as well as the challenges of balancing user-assisted translation with machine translation. We also describe three ambitious global scale language projects selected due to their support for the emergent “lexical web” and for their potential to become an integral part of future cross-language information retrieval systems (OmegaWiki, Wiktionary, and Global WordNet Grid).

Cross-language information retrieval systems A CLIR system is a complex piece of software built to interpret the user’s information need, translate it, match it against a collection of documents, and finalize the retrieval task by presenting a list of

search results. Such a system needs to accommodate both polyglots (fluent in several languages), and monolingual users. These requirements can be met by carefully orchestrating the collaboration between several components: a set of translation resources, natural language processing tools, and algorithms for handling query processing and retrieval. Out of these components, we will focus predominantly on the challenges of building those parts of the system that handle translation.

Translation Challenges The translation aspect of a cross-language information retrieval system entails determining: a) how to obtain translation knowledge, b) what to translate, and c) how to use translation knowledge or reconcile its absence (He & Wang, 2007). In terms of translation knowledge CLIR systems rely on: dictionaries, parallel and comparable corpora, and human expertise. Out of these types, the dictionary-based CLIR has the most flexibility and scalability. For a dictionary-based CLIR system, the key concerns refer to dictionary coverage and ambiguity in both the query language and the document language. Moreover, there are three main approaches for what to translate: a) translating the query before submitting it to the search engine; b) translating the documents collection prior to searching; or c) using inter-lingual technologies that map both queries and documents to a language independent representation (interlingual index). Out of the three, the query translation has proved more efficient, flexible, and less costly with the major drawback of having to handle translation ambiguities of query terms. The document translation requires running machine translation programs on a batch set of documents before queries can be submitted which suffers from scalability problems when the document collection is dynamic. And last, the interlingual index approach that requires encoding concepts in different languages and defining mappings in-between them. There have

279

Improving Cross-Language Information Retrieval by Harnessing the Social Web

been a number of projects that achieved partial success in creating manual (e.g EuroWordNet) or automatic interlingual indexes by using latent semantic indexing, the generalized vector space model, or the kernel canonical correlation analysis (a classical multivariate method) on parallel corpora (He & Wang, 2007). It is apparent though that the latter techniques do not work well for low or medium density languages. The last challenging aspect concerning translation is how to make use of translation knowledge or reconcile its absence. We will assume the CLIR system supports the query translation approach. In this case there are several possible instances: one translation per query term, more than one, or none. The first instance poses no problems, while in the case when more than one translation is known for a query term, CLIR systems either make an automatic selection or assign weights for each translation. For the latter strategy of accommodating uncertainty, there exist several algorithms that distribute the weights evenly or by exploiting the structure induced by the translation process techniques (Pirkola, 1998). The key idea for Pirkola’s method for example, is to separately estimate the term frequency (tf) and document frequency (df) of each query term based on the tf and df of individual translations. This approach has since been extended to accommodate translations learned from examples (where any term might conceivably be the translation of any other) by leveraging translation probabilities rather than a limited set of known translations (Darwish & Oard, 2003). In the instance when no translation is known for a query term, stemmers are used to extract the root of the word (e.g “smile” rather than “smiling”) and thus improve its chances to find a matching dictionary entry for it. Stemmers such as the Porter Stemmer are linguistic tools trained probabilistically, and hence some of the morphological variants extracted are not always valid words. Also, for certain languages such

280

tools are not available at all. This requires that CLIR systems do allow unknown query terms to remain un-translated or with minor changes (e.g, removal of diacritic marks) in the hope they will match terms in the document languages (proper names or “loan words”) (Oard 2008). From the overview above of translation challenges, we can conclude that a CLIR system needs to bring together a good number of linguistic tools and language resources, from tokenizing tools, stemmers, list of stopwords, phrase identification tools, dictionaries and/or corpora, and algorithms to handle translation disambiguation or weighing of translation alternatives. Thus, there is a strong interdependency between the language resources’ characteristics (coverage, structure, quality, etc.) and the performance of a CLIR system.

interactive Cross-language information retrieval systems An interactive CLIR system extends a classic CLIR application by integrating users in determining how queries can be formulated, expanded, or translated. This poses five key problems in implementing it: (1) interaction design, (2) query formulation and translation, (3) cross-language search, (4) construction of translated summaries, and (5) machine translation of documents. Each of these challenges has been explored individually in the interactive track of the Cross Language Evaluation Forum (iCLEF) started in 2001. Results obtained so far struggled to prove statistically the user’s impact on an iCLIR system’s precision and recall. The difficulty in making such assessments lies in the complexity of the interaction between a user and a CLIR system, and in defining suitable measures for interactive CLIR systems in general. Nevertheless, the results of user studies were encouraging from a user satisfaction perspective, and emphasized a series of system features that did improve the respective CLIR systems. The specifics of some

Improving Cross-Language Information Retrieval by Harnessing the Social Web

of these studies are presented below, contrasting three CLIR systems: MIRACLE, CLARITY, and PanImages.

MIRACLE (Maryland Interactive Retrieval Advanced CrossLanguage Engine) This is a project of the University of Maryland that was used as a testbed for several experiments for iCLEF. These experiments looked at the human-retrieval system interaction from several perspectives: i) query formulation and reformulation, ii) translation selection and reselection, and iii) document selection and reselection (see Figure 1, for an overview of the data flow in MIRACLE). The searchers in these experiments are predominantly monolingual (English speakers). The system’s initial design incorporated four innovations: a) user-assisted query translation, b) progressive refinement (“search results are presented immediately using all known translations and then updated in response to control

actions”), c) weighted structure query methods, and d) configurable translation (establishing a balance between accuracy, fluency, and focus -term highlighting -- when displaying translated documents) (He and al., 2003). Experiments were done using English, French, German, Cebuano, Hindi, and Spanish, and the system was built with a strong statistical machine translation component trained on online bilingual lists, sentence-aligned parallel text, or comparable corpora. Apart from this type of translation, MIRACLE got users directly involved in translating queries by presenting a set of automatically generated explanations (cues): synonym lists, translation probabilities, and examples of usage (keywords in context – KWIC) of the potential translation. The presentation of cues and their quality has been refined in time, but overall their degree of utility proved to be variable. Nonetheless, the cues did reveal an interesting search pattern, namely query reformulation based on words that appeared in either the synonym lists or KWIC. But this is still a partial success, since there is no direct mapping between the query term translations being used

Figure 1. Data Flow for user-assisted query translation for MIRACLE system

281

Improving Cross-Language Information Retrieval by Harnessing the Social Web

in the search, and the document collection index terms that need to match.

CLARITY Clarity is another example of an interactive CLIR system that has been designed incrementally, based on information collected from user studies, to perform multilingual searches for low-density languages. It uses a dictionary-based approach for translation and includes the following languages: English, Finnish, Swedish, Latvian, and Lithuanian. This project was set up to investigate how likely it is to get users actively involved in the process of translations. As opposed to MIRACLE, it is a system that considers polyglots as its main category of users, retrieving documents from different language collections simultaneously. In terms of interaction, this system sequenced the translation step and the search step (Petrelli et al., 2006). Comparative studies have been set up to gauge a user’s willingness to supervise the translation step by selecting or deselecting from a list of potential translations. Results indicated that in supervised mode, when users verify and refine the translated query, the system has performed better than in delegated mode, when users do not intervene in the translation process. Though differences were not statistically significant in terms of precision and recall. It was discovered that the supervised mode helped some of the users to reformulate their initial query based on suggested translations. This search pattern was also observed by the experiments with MIRACLE, as mentioned previously. Out of the Clarity investigations, a set of guidelines for setting up iCLIR systems have been established: a) enable the user to bypass query translation and to fix incorrect or missing translations; b) use rich dictionaries; and c) consider cross-language phrase search (Petrelli et al., 2006). The system architecture, to be introduced in the main section of this chapter, only incorporates the first two.

282

PanImages In contrast with the iCLIR systems described above, PanImages is a cross-language image retrieval system. Its translation mechanism relies on the TransGraph (a graph where a node represents a word in a particular language, and an edge denotes a sense shared between words in a pair of languages). In terms of interaction, the system parses a user query, and for each query term displays a list of distinct word senses along with glosses, if available, and the number of translations for each word sense. The user can than click on a word sense to see the list of translations for that sense. PanImages presents the word sense with the largest number of translations first, and selects this as the default word sense. The user selections combined with the default selections are submitted to Google’s image search retrieval engine (Etzioni, 2007). The distinguishing element of this system is the TransGraph that is automatically constructed from a collection of independently authored, machine-readable bilingual dictionaries, and multi-lingual Wiktionaries. Merging these differently structured resources has a drawback, namely word sense inflation. This problem is triggered by the TransGraph assumption that each translation constitutes a distinct word sense. To help handle this aspect of the graph, the problem of lexical translation has been formalized as probabilistic inference. Determining word-sense-equivalence in the graph equates to computing the probability that two words senses are equivalent (note this is not always defined). This translation graph does produce translation errors as well, since it is built on the premise that entries in the dictionary are distinguished by sense. Further developments of the project are looking into ways of altering the probabilistic inference mechanism in order to alleviate the impact of these errors. Nevertheless, the TransGraph has scaled to 1,267,460 words in more than 100 languages, with 3 of the languages having 100,000 words and

Improving Cross-Language Information Retrieval by Harnessing the Social Web

58 of the languages having at least 1,000 words. This structure gives a greater coverage than any of individual resources and has proved to improve precision for image retrieval (especially for lowdensity languages). It is though a representative example of usage of community-built language resources, i.e Wiktionaries. So far, we have accentuated the importance of multilingual resources for the performance of CLIR and iCLIR systems, and that most problems concerning translations stem from the divergences between language resources. Therefore, let us investigate the present initiatives in developing better resources coming from the natural language processing community.

language resources and the web Currently, many of the world’s language resources are made available through the Linguistic Data Consortium (LDC) in the United States, the European Language Resources Association (ELRA), and a number of web portals (see Ningthoujam, 2007 for examples of language communities web sites, offering courses and tools for learning multiple languages, translation tools and services). The Natural Language Processing (NLP) community, in particular, has acknowledged that the current ways of developing language resources (LR) need more coordination and consideration for the requirements of future human language technologies. The following principles have been actively promoted for consideration (Calzolari, 2008): i) interoperability of LRs; ii) collaborative creation and management of LRs (along the lines of a wiki model); iii) sharing of LRs; iv) dynamic LRs, able to auto-enrich themselves; and v) distributed architectures and infrastructures for LRs, to encompass and exploit the realization of the previous notions. The solution that addresses all these aspects of developing LRs is the creation of “distributed language services, based on open content interoperability standards, and made accessible to

users via web-service technologies” (Calzolari, 2008, p. 9). Global projects that respond to these initiatives are: i) OmegaWiki that aims to provide information on all words of all languages; ii) Wiktionary, a collaborative project for creating a free lexical database in every language; and iii) Global WordNet Grid that is converging from languagespecific WordNets to a global interconnected WordNet (Fellbaum & Vossen, 2007). We will have a closer look to each of these three projects in the following sections.

Wiktionary and OmegaWiki Do web spaces really incite people to contribute to world’s knowledge? The classic story of success is Wikipedia. From there other projects spawned among which are a multilingual dictionary (Wiktionary), and a wiki for all words (OmegaWiki). These projects share the same paradigm of collaboration, giving any user the opportunity to add, edit, even request word entries. But for these projects to actually succeed in terms of validity of submitted information, the wiki entries need to follow strict guidelines. Apart from these specific rules, the OmegaWiki uses the “Babel template” page to keep a record of ones fluency in different languages. For the Wiktionary one needs to follow well-defined templates and carefully create word definitions without replicating copyrighted dictionary entries. It is also a user’s responsibility to create interlanguage links (smart links), which have no semantic labeling (as opposed to relations in a WordNet). Though both wiki instances require very dedicated users, the data entries count for most languages have gone up (e.g 48120 entries on the German side in three years of Wiktionary usage – Witte, 2007), with other languages still having very low represention (Samoa, Fillipino, etc.). None of the projects can argue that they have gained as much momentum as Wikipedia, but research shows that they could serve as large containers for entries created by bots with content collected

283

Improving Cross-Language Information Retrieval by Harnessing the Social Web

from users at moments of user interaction with the web (Witte & Gitzinger, 2007; Zesch et al., 2007). It is also important that wiki-type systems have a clear workflow for validating entries embedded, which is paramount for sanctioning any invalid entries that may be published by bots. The next project to be introduced contrasts with these two projects by the rigorous process involved in creating entries, by the richness of semantic relations between concepts that can be added and that cannot be captured solely by a wiki infrastructure.

Global WordNet Grid This project is one of the most ambitious concerted efforts to build a global scale multilingual resource. It started in the 90s with the Princeton WordNet (a lexical database for English) that spawned other WordNets for over 40 languages. Projects such as EuroWordNet incorporate an Interlingual Index (ILI) that maps synsets (groups of semantically equivalent concepts) from one language to an equivalent synset in the ILI (Miller & Fellbaum, 2007). It is worth mentioning that large parts of WordNets are built manually. After almost 20 years of research on creating WordNets and after getting more perspective on the differences and communalities of semantic networks and ontologies a new initiative was setup: the Global WordNet Grid, an attempt to encode and interrelate concepts from as many languages as possible. Two inclusion criteria for this global lexical database have been laid out for discussion: linguistic and cultural. The first criterion refers to how current and salient a concept is, within a community of speakers. The second criterion determines that specific concepts must be included in an interlingual index, “although there may be no equivalence relations to any languages other than the one that lexicalizes such concept.” (Fellbaum & Vossen, 2007, p. 5). After weighing the limitations of the current concept relations captured by WordNets,

284

and abiding by the same principles as the one mentioned in the first paragraph of this section, a proposed architecture has been outlined. The Global WordNet Grid will “comprise a languageneutral formal ontology as its ILI. This ontology will differ from ILI in EuroWordNet, which is a list of unstructured concepts derived from English WordNet” (Fellbaum & Vossen, 2007). The architectural decision to use an ontology as the core of the Global WordNet Grid brings with it an important advantage for applications: the ability to use the Knowledge Interchage Format (KIF). This allows for inferences to be made and for more expressivity when adding relations or properties for concepts. The drawback to this decision is that KIF is a language that requires a certain level of expertise from the human user adding concepts to the lexical database. The realization of this project depends strongly on engaging in a collaborative framework people from diverse linguistic and cultural backgrounds. A wiki infrastructure was suggested as the environment for collaboration and contribution of new concepts.

Cross-lingUAl inforMAtion retrievAl in the CUrrent web sPACe As outlined in the previous sections of this chapter, one of the main challenges in scaffolding a large-scale CLIR application is the availability of global language resources. Projects such as the Wiktionary, OmegaWiki, or Global WordNet Grid are progressing slowly, and depend on the direct participation of the web community or on programming bots to create stub entries. Moreover, these resources are by design global and generic, and do not reflect the associated conceptualizations of specific groups of users. To balance this aspect, we claim it is possible to extract more specific sets of annotated metadata from online communities to act as personalized

Improving Cross-Language Information Retrieval by Harnessing the Social Web

language resources during the query processing and retrieval stages of a CLIR system. To support this statement, this section puts into perspective the concept of “emergent semantics” from web communities and its direct application to web search personalization, and user profiling. The research in emergent semantics is still in its early stages and so far, there is not a large body of work that targets exploiting the multilingual aspect of the web communities for CLIR. We suggest an extended architecture for iCLIR systems integrated with the current web space parameterized by user language profiles, tags, and documents.

web spaces The web spaces referred to in this chapter are the Semantic Web and the Social Web (web 2.0). They represent two complementary approaches that focus on incorporating the notion of semantics into the web architecture. The “Semantic Web is a web for machines, but the process of creating and maintaining it is a social one. Although machines are helpful in manipulating symbols according to pre-defined rules, only the users of the Semantic Web have the necessary interpretive and associative capability for creating and maintaining annotations of web resources” (Mika, 2007, pp. 13-14). For instance, an application enhanced with a social tagging component allows the user to add free-text annotations (tags/keywords/metadata) of web pages, images, videos, or bibliographic entries. The tags represent personalized associations in the user’s mind, between a concept and a resource, or a memory cue. One user’s tags are used as suggestions for tagging by another user. This creates a powerful informal knowledgesharing channel between users and it leads us to investigate further to what extent web users’ generation of knowledge pools representing both the collective and the individual mindset, can be employed by other applications to improve their overall performance?

In the present context, the question we are interested in is narrower: to what extent a user’s aggregated digital manifestations/annotations on the web (i.e personomy) can be employed for personalizing web search? Furthermore, considering multilingual web search, is it possible to establish channels of informal collaboration between users with different language skills to facilitate the identification of relevant web resources regardless of the document language? There is supporting evidence for the first question, suggesting that a user’s personomy can positively influence the re-ranking of the retrieval list in a web search (Noll & Meinel, 2008) or that a user’s domains of interest can be circumvented from his personomy, and may be employed for web page recommendation or a personal resource manager (Yeung et al., 2008). These experimentations have to compensate the fact that keywords/tags are not fixed sets of words, nor is there a one-to-one mapping between concepts and keywords. This drawback is attenuated by focusing on the keywords that are shared by larger groups of users. A formal model for the study of emergent semantics from these annotations (Mika, 2005), in particular from folksonomies, is based on a tripartite graph with hyperedges. The set of vertices of the graph is partitioned into three disjoint sets, corresponding to the set of actors (users), the set of concepts (tags, keywords), and the set of annotated objects (bookmarks, photos, etc.). Based on this formalism, Halpin et al. (2007) demonstrated that the dynamics of tags associated to a resource could be described by a power law distribution. Furthermore, the “tag co-occurrence networks for a sample domain of tags can be used to analyze the meaning of particular tags given their relationship to other tags” (p. 1). This constitutes the basis of the assumption that it is possible to discover implicit relationships between tags formulated in different languages, considering that social tagging systems are accessed by multilingual communities. As opposed

285

Improving Cross-Language Information Retrieval by Harnessing the Social Web

to implicit relations between tags, the architecture suggested in the following section envisions to stimulate explicit multilingual associations by keeping track of the relevant resources obtained for a given query.

integration Architecture for iClir systems The previous sections built the case for expanding the existing CLIR architecture to allow it to work with global language resources and to tap into the data and knowledge pool collected by web 2.0 services. The step forward taken by several research communities (see Background section) to consolidate the language resources and distribute them through the web space will impact several applications, among which are the deployment of large-scale CLIR systems for extended multilingual communities. Such a synergy will lead to systems that grow organically as more users with different linguistic skills join the network, and that improve in terms of language translations disambiguation and coverage (i.e. more in tune with cultural changes of meaning within a community). The examples of interactive CLIR systems presented in the Background section incorporate the following core system components: (1) an interaction interface that supports user-assisted query translation, (2) a query processing component that reacts to input from the user and links to a diverse set of language specific NLP tools and language resources, and (3) a retrieval engine that executes the query, and delivers and consolidates the query results. Our envisioned iCLIR system’s architecture requires overall adjustments to allow its components to handle multiple lexical and knowledge translation resources of different origins and representations, as well as a user’s personalized set of annotations. The proposed architecture incorporates several extra components that consolidate the linguistic sources: a Multiple Language

286

Resources Compiler, a Personalized Interlingual Index, and a generic Interlingual Index structure to interface with the web in its multiple functions as a social, cultural, and economic space, and as a container of multilingual data and knowledge (see Figure 2).

overview of the extended iClir Architecture The scaffold of the amended architecture relies on: a) an Interaction Interface, b) Query Processing Component, c) Multiple Language Resource Compiler, d) Interlingual Index (ILI), e) Personalized Interlingual Index (myILI), f) Interaction Aware Component, and g) Web Search Engine (see Figure 2). The system based on this architecture will enable an iCLIR task to work as follows: 1.

2.

3.

4.

5. 6.

The user decides on the collection of resources to be searched and the corresponding document language; The user formulates a query and waits for translation suggestions from the interlingual index and the personalized interlingual index; The user checks or unchecks suitable translations for each of the query terms that were identified (the provided cues are filtered based on similarity measures with the user’s personomy, alleviating the problem of overloading the interface with too much information); The user is also allowed to submit his own suggestions for the query term translations (this feature suits polyglot users). This functionality is also meant to populate the multilingual index(es) where no other translation knowledge resources are available. The user submits the query and inspects the documents retrieved; If the results delivered are unsuitable the user will reiterate through query formulation or translation (see Figure 1);

Improving Cross-Language Information Retrieval by Harnessing the Social Web

Figure 2. iCLIR system component interactions

7.

If the user identified a relevant resource he can bookmark it; the system interprets this as a successful mapping between the query in the source language and the translated query; it will store on the user’s computing device the successful match as part of the myILI; it will also parse the initial query into keywords in the source language to annotate the found document; this gives the document’s set of metadata a multilingual dimension, and creates a path for monolingual search engines to identify it;

A. Interaction Interface According to observations from the development of the iCLIR systems described in the Background section, defining an interface entails supporting

three main aspects of the system: i) transparency -- how the system is affected by the query formulation and query translation, ii) control -- what translated query terms are used for retrieval, and iii) iterative query refinement – the ability to revise the initial query or its translation when search results are not relevant. These requirements can be met by an interface that engages the user in selecting or deselecting translations based on cues (such as glosses, synonyms from the ILI, or co-occuring terms from myILI), as well as filters based on myILI which are presented as options to the user. Several iCLEF runs (Wang & Oard, 2001; He et al., 2002; Dorr et al., 2003; Wang & Oard, 2006) pointed out that users learn from translation cues and go back and reenter the query. Re-formulating the query has the greatest impact on retrieval, in

287

Improving Cross-Language Information Retrieval by Harnessing the Social Web

other words choosing the best terms for translation. Hence, it is of great concern that translation cues are meaningful to the user. Extracting cues from large resources and combining them with personalized conceptualizations addresses this problem. In order to understand the role of each of the proposed system’s components and their interaction, let us consider the following scenario: A Romanian native speaker is trying to find information on “rezultatele campionatelor Europene de gimnastica” (i.e., “results of the European championships of gymnastics”). The user is looking for news coverage on both Romanian and English online papers. The query is parsed into query terms, by tools incorporated in the Query Processing Component, and then a set of suggested translations are displayed. Note that a back-off algorithm needs to be applied to ensure suitable matches are found for each of the words that show in articulated form. If the user chooses to provide no input on the query translation the system will process an unsuitable mapping for the query term “rezultatele” and pick “issue” instead of “result” (automatic translation -- InterTran online system). In supervised mode, the system lists several translation alternatives for the previously mentioned query term: “harvest”, “issue”, “outcome”, ”produce”, ”product”, ”purpose”, and “result”. Note that the list is relatively long, and a decision on determining which words to show and which to hide will either rely on probabilities of occurrence in a corpus or on a similarity measure within the user’s personomy (see Noll & Meinel, 2008, for sample measures). Based on the list of translations, alternatives, and available cues (glosses, synonyms, or KWIC examples) the user will finalize the decision on what translations are best at preserving the initial sense of their query. There are though instances when polyglot users know better translations, or the system cannot suggest any translations because of lack of translation knowledge. This is particularly important for

288

minority or low-density natural languages, or for natural languages for which no dictionaries are available or are not cost-effective to be developed. For those cases the interface will also capture a user-inputted translation to be shared by subsequent users, since this will also become part of the system generated suggestions for future query translation or query term disambiguation problems in the same language. It is also worth mentioning that this user generated translation mapping may provide a simple, however, effective way of populating multi-lingual indexes to be reused and shared by web communities.

B. Query Processing This is the core component that determines what to translate when a query is submitted (query term extraction), what translation knowledge is available, and how to perform the translation step. The processed query is then passed on to the search engine. In order for this component to interface well with the Personalized Interlingual Index (for user context, and more specific vocabulary) and with the Interlingual Index (for mapping concepts from one language to another) there needs to be an agreement between the knowledge representations used by each component. A possible solution is to parse the query to a conceptual tree that is language independent, and captures through vertices and edges the concepts and their relationships. Kapetanios et al. (2006) initiated work on deep parsing a query into a conceptual tree using Lexical Conceptual Structures (LCS). These are compositional abstractions with language-independent properties that transcend language structural idiosyncrasies. In effect an LCS is a directed graph with a root. Each node is associated with certain information, including a type, a primitive and a field. LCS resources are hard to build, but simplified structures that express valid associations between concepts (e.g {win, war} or {win, love}) are easier to derive automatically.

Improving Cross-Language Information Retrieval by Harnessing the Social Web

At this stage, we cannot prove that the conceptual tree is the best representational structure, but it is a feasible choice, considering that the formal model for personomies are bipartite graphs (see Personalized Interlingual Index) and that successful merges between lexical resources have been obtained by building translation graphs such as TransGraph. Furthermore, the user generated and inputted translations can be directly interfaced with the API for the construction of the query tree graph, since each user inputted translation becomes a node to enhance the query graph.

C. Multiple Language Resources Compiler The Multiple Language Resources Compiler plays a central role in ensuring the effectiveness of translations. The PanImages’ TransGraph structure that was detailed in the Background section has great potential to handle the difficult task of successfully merging several dictionaries, distinguishing word senses, and providing good translations that preserve the initial word sense. This component will update itself whenever its source resources are changing. The compiler will combine solely lexical translation resources and consult for word senses against knowledge translation resources. What results is the creation of a unified structure (the Interlingual Index) that is easier to use by the other components of the system. This component is computationally intensive and ideally will be set up as a shared web service that can serve other applications.

D. Interlingual Index (ILI) The output of the Multiple Language Compiler is an interlingual index, a translation knowledge representation structure similar to the TransGraph. If available, the ILI ontology (see the Global WordNet Grid) should be linked to this component. It is important to mention that automatically

compiled structures such as TransGraph rely on statistical techniques to learn the empirical probabilities for the translation mappings. This typically yields, instead of a direct mapping between words, a few highly probable translations and a large number of possibilities that occur with very low probability. Applying the weighting translations methods (Pirkola’s method) across the full set of possible translations in such circumstances can give quite poor results (Oard et al., 2008) since very large sets of possible translations are likely to include at least one term that is very common. Hence, how to strike the right balance between large word/ concept coverage and precision when building an ILI structure is still under scrutiny (see Witte, 2007; Sammer & Soderland, 2007) This component is also receiving input from users in the community through the Interaction Aware Component.

E. Personalized Interlingual Index (myILI) This component has the task to derive the conceptual universe a user has created online, his personomy. It is a plausible assumption that a high number of users have tagged resources online through web services like Flickr, del.icio. us, stumbleupon, Technorati, or digg. Also, most news website provide a feature that encourages the users to make an annotation. All the web 2.0 type services just mentioned offer programmatic access to the data they store. Hence, it is straightforward to collapse them into an unstructured pool of metadata kept on the user side. A size-bounded subset of this will constitute the initial seed data for myILI. To this set, as the user performs more successful cross-language searches, the Interaction Aware Component will capture the relevance feedback by adding the initial query search and its successful translation to myILI. In the case where a user does not have their own trail of tags on the web, it should be possible to download such a set from users that share the

289

Improving Cross-Language Information Retrieval by Harnessing the Social Web

same interests or cultural background (sharing bundles of knowledge). This is a slightly speculative assumption, but it relies on web 3.0’s overall trend to make sense of small granularity data.

F. Interaction Aware Component This component is paramount for monitoring the user-system interaction. Its function is to capture the relationship between queries, their translations, and relevant resources (updating myILI with a query and its translated mappings when identifying relevant resources). It also handles the automatic submission of the initial query terms as tags for the found document. This enables indexing the document collection with the same keywords as the ones used for translation.

G. Web Search Engine The components described so far, take over the responsibility of processing the query before it gets submitted to a search engine. This gives the user the freedom to connect to different search engines and use their myILI data structure with the most suitable one. For example, Russian native speakers will favor the Yandex search engine that is unknown for English speakers. Hence, a system based on this architecture will enable web searchers to disseminate metadata regardless of geographical or language boundaries, and also to preserve their individuality in the search process. The system architecture presented in this section is a first attempt at zooming into the challenges of building a collective knowledge system that exploits collective intelligence. By collective intelligence we refer to the large amount of human-generated knowledge that can enable emergent knowledge through computation and inference over the collected information, and can lead to better indexing of collections, discoveries, or other results that are not explicitly specified by human contributions.

290

evAlUAting interACtive Cross-lingUAl systeMs The ideas expressed so far cannot yet be evaluated in their entirety, but we were able to draw some partial conclusions after participating in the interactive Cross-Language Evaluation Forum (iCLEF) track at CLEF 2008. This is a specific evaluation track for iCLIR systems and the task setup allowed us to look at the correlation between user language skills and behavior. The organizers provided a test iCLIR system called Flickling, a CLIR front-end to the Flickr database that uses the Flickr API for the retrieval of images. Flickling was disguised as a game and made available to everybody on the web. Players signed up on a voluntary basis and they had to look for images from a specially selected set. These images had generic content and annotations were evenly distributed between English, Spanish, German, French, Dutch, and Italian. Basic dictionaries were incorporated into the system for each of these languages, and users were able to interact with the system for selecting or deselecting translations, adding new words, or updating entries. Hence, each user was able to create a personal dictionary (note that one players changes were not shared with other gamers). The total number of players was approximately 300, with more than 200 that actively played the game. While signing up each player filled in a short pre-game questionnaire specifying their mother language, the active languages (fluent writing and reading), passive languages (some level of fluency) and unknown languages. During the game, players were asked to fill in short questionnaires after finding or giving up on each image search, as well as after looking for more than 15 images. The results of these questionnaires and of the step-by-step actions of the players were distributed to the interested research groups. In our analysis we focused on extracting the entries that related to the users’ language profiles, the interactions

Improving Cross-Language Information Retrieval by Harnessing the Social Web

with the translation mechanism, the addition of new entries in the personal dictionary and on the overall user’s results for the game. Below we summarize the main questions we have tried to answer based on the log’s content. We have grouped the users in five categories based on the number of languages they were familiar with. We took the decision not to distinguish between active or passive skills: firstly, they are a subjective measure, secondly, three of the languages of the system are Latin-based with lots of words in common from an etymological point of view, and thirdly, given a random image the user that knows more languages is more likely to do better. We used a coarse index from 1 to 5 to measure the degree of confidence with languages. Below are the main research questions we have answered: a. Does the degree of confidence with languages affect usage and creation of personal dictionary entries, i.e., do those users with little knowledge of a language make use of the personal dictionary and to which extent? The interesting aspect of the log data was that users with a smaller degree of confidence (” character (e.g. Business > hotel > Hilton or France > Paris > Hilton). Basic issues, like polysemy, homonymy and base level variations, are solved in this way, using contextualization and user-added semantic value. Unfortunately some open issues emerge: •









there is no distinction between the different types of hierarchies. This also means that at each level of the hierarchy, the relationships between concepts are not semantically expressed; multiple hierarchies may exist to identify the same terminal values: a concept may belong to different classes (poly-hierarchy); there is no identiication of pseudo-hierarchies (e.g., showbiz > California > Paris Hilton) and of bogus hierarchies (weapons of mass destruction > blonde > Paris Hilton); there is no reference vocabulary for instances (such as proper names) which count for more than 25% of all the tags of documents. ontologies do not solve words ambiguity and are not updated on natural language evolution, neither on metaphorical uses of lexical units.

Next section discusses in more detail current open issues and defines some possible research lines for ontology emerging from folksonomies.

360

fUtUre reseArCh direCtions Word Sense Disambiguation is a problem well recognized and addressed in computational linguistic (Yngve, 1995). But while in computational linguistic the disambiguation can be performed on the neighbouring sentences and words, in folksonomic tags we have almost no context around. For this reason, the above-mentioned limits impose new and innovative approaches. We are currently experimenting with a few of them.

Clustering In order to study the tags behaviour, it is important to do a statistical analysis of tags in order to identify groups, or more appropriately clusters, of related tags. Clustering is the classification of objects into different groups, sometimes even overlapping, so that the data in each group (ideally) share some common trait, often expressed as proximity according to some defined measure of distance. In particular, semantic clustering is the clustering of objects based on proximity in their meaning. Through clustering it is possible to determine similarity in meaning based on the contexts according to which the documents are tagged, i.e., by examining not only the individual tag, but also all the tags that are associated to the document, and all the tags that are associated to all documents that include the individual tag. The distance between tags is then computed by considering the relationships that compose the context of use of these tags. This technique allows us to differentiate each context of use of an ambiguous tag. For instance, “apple” is clustered differently to refer either to a fruit or a company, and is disambiguated by considering whether it is appearing near tags such as “computer” rather than “pie”. Consider for instance a theoretical tagging of a document by different users (Table 1) The fact that these terms all refer to the same document does allow us to infer that their semantic distance is limited, and that in some way at least

Towards Disambiguating Social Tagging Systems

Table 1. An example of document tagging Tag

User

Kids

Joe Green

Document A

Cartoon

Joe Green

Document A

Aladdin

Joe Green

Document A

Disney

Mary Violet

Document A

Cartoon

Mary Violet

Document A

Movie

Hugh Orange

Document A

Kids

Hugh Orange

Document A

one meaning of both “Aladdin” and “Disney” belongs to the same neighbourhood of at least one meaning of the word “cartoon”, given the fact that this term appears in both tag sets where they appear, i.e. we can infer that they are clustered together because of some (unspecified) semantic justification involving “cartoon”. A reasonable expectation is also that the other meanings of these words are clustered differently, and therefore have different distances between them. There exists different approaches to tags clustering. Motta and Specia (2007) in their paper show a specific analysis based on tags co-occurrence, in order to find “similarity” of tags. They use two smoothing heuristics to avoid having a high number of these very similar clusters. For every two clusters: •



Document

if one cluster contains the other, that is, if the larger cluster contains all the tags of the smaller one, remove the smaller cluster; if clusters differ within a small margin, that is, the number of different tags in the smaller cluster represents less than a percentage of the number of tags in the smaller and larger clusters, add the distinct words from the smaller to the larger cluster and remove the smaller.

That would be an important classification of tags, in a specific prospective, and generates a set of clusters resulting from distinct seeds that are

similar to each other. Another possible algorithm that we have considered would be based-on a fuzzy approach. Clustering is hard if it produces an exact partition of the data set, as in the case of the Motta and Specia approach, and it is termed fuzzy if it produces a fuzzy set covering the data set, whereby objects belong to clusters with a certain degree of truthness expressed as a number between 0 and 1. In order to talk about disambiguation of polysemic terms we prefer to rely on fuzzy clustering, since hard clustering does not allow any ambiguity, and forces to resolve it automatically by selecting only the best cluster for each term and excluding all the others. Fuzzy clustering, on the other hand, allows terms to belong to multiple clusters with different degrees of certainty, and can take semantic ambiguity in consideration.

identifying Proper names Another approach to disambiguation is to provide a way by which (at least a few) users add structure and depth to social tags. This can be obtained by providing a syntactically simple mechanism to qualify the terms used. As mentioned, a similar mechanism has been proposed (Quintarelli et al., 2005), but limited to expressing is_a relations (i.e., the BT/NT generic hierarchies between terms) as pairs of generic/specific tags such as feline > cat (see section “Thesauri” and “Ontologies”).

361

Towards Disambiguating Social Tagging Systems

We intend to concentrate on a different hierarchy, the instance_of relationship (Fisher, 1998), which connects an instance to a category, i.e., a proper name to a common name or an individual to its category. Rather than requiring the author of the tag to identify the immediately broader term of each relevant term, we only expect a categorical term (and, in fact, just about any reasonable categorical term) for each proper name (be it of individuals, organizations, places, etc.), such as person:Paris Hilton as opposed to hotel:Paris Hilton, or fruit:apple as opposed to company:apple, at the same time expecting any degree of variability in the categorical term, i.e., allowing for variations such as socialite:Paris Hilton, heiress:Paris Hilton, inn:Paris Hilton, destination:Paris Hilton, or really any other category, generic or specific, that the mind of the reader comes up with in the spur of the moment. Such social tags would be exactly composed of exactly two parts, the category and the proper name. In fact, the relationship instance_of only matters for proper names, and the tag author needs only answer the simple questions “Is this a proper name? And if so, what is its category?” Among the advantages of this approach: •





362

The instance_of social tag has always exactly two levels, and never more. Therefore the categorical term can be chosen from any level of a multi-level is_a hierarchy of terms (such as WordNet). The instance_of social tag easily deals with the fact that no vocabulary of proper names exists, but only of categories. Proper names constitute a hearty percentage of tags in real life folksonomies. A method for devising a meaningful measure of such percentage is under way within our research team, but our initial considerations for sites such as del.icio.us suggests that over 20% of tags are proper names. All inferences and experiments in ontology building are always performed on the

categories only, and never on the proper names, which are by deinition open and are simply rewritten as non-controlled vocabulary. Also note that a tag separator that is explicitly different from space would allow for spaces to be available in tags, and thus for first name/family name pairs (as well as for city names such as San Francisco and New York) to be recognizable as such and to be considered as single tags rather than as two separate ones. One of the most interesting key points of proposing a richer syntax for disambiguation in folksonomies is that it is not necessary for all users to adopt it: in fact, it suffices for a few, and actually even just one author to use the syntax, to disambiguate all other associations of the same tag to the same document, even if they keep on relying on the unsophisticated syntax.

Adding Qualification Qualification can be used to conceptualize the tags of a folkonomy, and to let a real fully-fledged ontology to emerge from the concepts described therewith. The simple addition of a tag in a list is not sufficiently eloquent to determine if it describes facts about the document or about the content of the document. Tags in folksonomies, in fact, are used to describe the subject of the content of the document (i.e., what the document talks about), as well as incidentals about the characteristics of the document, its intended or perceived uses, and the relevance to the author of the tagging. Consider for instance the list of tags “DVD release date”, “kids”, “cartoon”, “Disney” “Aladdin” and “Christmas presents”. A human could immediately and reasonably infer that the document associated to this list of tags talks about the “DVD release date” for the movie titled “Aladdin”, which is of type “cartoon”, produced (or authored) by “Disney” and that the author of

Towards Disambiguating Social Tagging Systems

the tags is interested in it in relation to making “Christmas presents” aimed at some “kids” (their own, possibly). In order to qualify correctly the justification and meaning of these tags, a possible solution may be to populate some faceted classification properties such as Dublin Core. For instance, the mentioned tags could populate properties such as, respectively, dc:subject, dc:audience, dc:content-type, dc:creator, dc:title, dc:relation, and so on. In fact, it is not even necessary to use the Dublin Core properties correctly (in our case, “kids” for dc:audience is a bit of a stretch, Disney and Aladdin may be the dc:author and dc:title of the movie, but most surely not of the document talking about the DVD release date, and “cartoon” for dc:content-type is technically wrong) as long as reasonably distinguished qualifiers are used. Enticing users to qualify their tags can be done in at least two different ways: •



By using a positional organization of the tags, in a similar way to a Colon Classiication (see section “Faceted classiication systems”) on which are based Dublin Core facets. By providing them with a speciic list-like selector with terms from a controlled vocabulary for at least a few of the facets of the Dublin Core schema.

Faceted qualifications not only allow the association of tags to their category, but they also provide relationships that enable the correct generation of metadata property statements. Metadata plays a role very important in both cases, and also the use of the RDF standard (Resource Description Framework, http://www.w3.org/RDF), based upon the idea of making “statements” about Web resources in the form of subject-predicate-object expressions, makes it is possible to associate a computable form of the correct role of each tag of every document.

disambiguating slang words Dealing with folksonomies a big problem is to contextualize the tags according to the document they are associated to. This implies (as explained) describing the semantic distance of the tags in relation with other tags used by the different users for the same resource or by the same users for different resources. Contextualization also means defining the role of the tag as regard to its specific scope of use in terms of categories, and facets. To correctly assign the terms to their category it is possible to use linguistic resources to associate at least approximately the terms to their context. Some existing linguistic resources include WordNet (see also section “Integrate ontologies and facets with folksonomies”), a large lexical database of the English language. Nouns, verbs, adjectives and adverbs are grouped into sets of cognitive synonyms (synsets), each expressing a distinct concept. Synsets are interlinked by means of conceptual-semantic and lexical relations. The resulting network of meaningfully related words and concepts can be navigated through an API or directly within the browser. WordNet, though, provides definitions only for terms belonging to the “official” language, which is often a limited, bowdlerized, averaged view of the multifaceted, multi-localized and ever-evolving language that is really used by people for folksonomic tags. A large number of tags - again, as per proper names, we cannot provide reliable figures yet, but we notice a visible incidence - does not in fact belong to the view of the English language proposed by WordNet, either because the word simply does not exist in the official language (e.g., fanfic) or because the official definition provided does not really match the meaning intended in current or local usage of the word (e.g., douche bag). These terms, which we cumulatively call slang (even if there are subtle distinctions that should be made in using such term) cannot be satisfyingly

363

Towards Disambiguating Social Tagging Systems

catered for by traditional linguistic resources, both because of their often irreverent tone, and because of their frenetic creation and evolution. There are therefore two additional resources we are considering, although way less sophisticated technically than WordNet, that may give hints so as to disambiguate and provide some meaning to terms unreliably described by WordNet: •



364

Urban Dictionary (http://www.urbandictionary.com) is a dictionary of slang with deinitions provided by users. For each term it is possible to have different deinitions ordered according to credibility or just simply coolness. All slang terms we have encountered so far in folksonomies (except for foreign words) are present in Urban Dictionary with more or less credible deinitions. One disadvantage of Urban Dictionary is the level of noise that is present: a large number of terms are really extremely limited in scope (even down to usage within a single US High School) and many deinitions are clearly nothing but jokes, exercises in low-level humour, or personal offences, with limited usefulness except possibly for the self-esteem of their compilers. Wikipedia (http://en.wikipedia.org) is the well-known largest multilingual online encyclopaedia, built collaboratively using Wiki software. Wikipedia articles have been written by volunteers around the world, and nearly all of its content can be edited by anyone with access to the Internet. While much better guarded against humorous exploitation of its deinitions, the encyclopaedic rather than linguistic purpose of Wikipedia makes concrete disambiguation of tags quite dificult manually, and impossible automatically: almost every categorical word in English has multiple pages related to it (including people, places, books, records and movies

with that term as name or title), and often is associated to a disambiguation page to (manually) guide the reader to the actual meaning sought. On the other hand, Wikipedia does provide adequate light to all public personas, all large corporations, all main brands, or all major places whose proper name is used in folksonomic tags, as most of them have a page on Wikipedia, so it is a relevant source of information for disambiguation of such tags.

ConClUsion Metadata represents one of the most popular ways for retrieving relevant information in search engines. The conceptual basis of social tagging is that users’ information associated to documents via folksonomies provides a good and reliable set of descriptors of the documents themselves, i.e., social tags are really representative of the aims and content of the documents. The analysis of this “data on data” is fundamental in the new frontiers of the Web, as it aims at establishing a collective knowledge and allowing a global collaboration environment for the production and the description of electronic resources. However, the polysemy of natural language requires us to not get rid of controlled vocabularies already, especially whenever it is necessary to convey meaning through concepts rather than potentially ambiguous natural language words. In this paper we have presented a collection of works and efforts to bring together formal classification methods and social classification efforts. The path towards joining in a single allencompassing environment these radically different approaches is still long. We have listed a few of the still unanswered issues (proper names, slang, facets) and proposed a few possible ways to approach them (cluster analysis, syntactical extensions to tags, and socially generated linguistic resources). Of course, the realization and concrete usefulness of these approaches are, as of

Towards Disambiguating Social Tagging Systems

now, fully undemonstrated, but we are confident that they will at least be considered interesting initial steps. We also need to discuss some intrinsic limitations in what we are proposing, that makes solutions harder to implement and exploit. In particular: •



As already mentioned, both Urban Dictionary and Wikipedia are not designed to be used as linguistic resources in automatic engines, but rather as interactive reference tools for humans. Thus, besides the obvious problems of reliability, noise and information overload that their use imply, accessing deinition features of the terms (even the simple distinction between common names and proper names) is dificult, error-prone and heavily dependent on NLP algorithms to work on their deinitions. Clustering algorithms, and in fact any algorithm that attributes relevance to items by considering information available outside of the items themselves, is open to malicious attacks by determined individuals and organizations planning to take advantage of the algorithm. The practice of edit wars, spamdexing, or Googlebombing are clear examples of these kinds of exploitations, and are impossible to deal with in an automatic way (i.e., by the algorithm itself), since any kind of prevention becomes automatically part of the algorithm and as such is open to (possibly different kinds of) further exploitation. Only manual operations on clearly identiied attacks can be considered adequate responses to these practices, and they require massive manpower for even a starting and limitedly successful Web service.

It is hard to see a simple solution to these problems, but on the other hand they are shared with a large number of other (and fairly successful)

services, which we would never think of giving up to. As such, these problems will make all these services float together or sink together, and solutions found for one will work for all the others.

ACKnowledgMent The authors would like to thank the colleagues and students that have contributed and are contributing to this research. In particular, a big thank you goes to Giovanni Rossi, as well as to the folksonomy folks (Nicola Di Matteo, Ferdinando Tracuzzi, Barbara Angius and Natalino Mondella) of the department of Computer Science, for their ongoing work and early contributions to these activities. The author would also like to acknowledge the European Thematic Network Project Acume 2 (http://acume2.web.cs.unibo.it/), within which a part of the activities here described are being delivered.

referenCes Agirre, E., & Edmonds, P. (2006). Word sense disambiguation: algorithms and applications. Dordrecht: Springer. Au Yeung, C. M., Gibbins, N., & Shadbolt, N. (2007). Mutual Contextualization in Tripartite Graphs of Folksonomies. In The 6th International Semantic Web Conference (ISWC 2007), LNCS (4825/2008) (pp. 966-970). Berlin-Heidelberg: Springer-Verlag. Baruzzo, A., Dattolo, A., Pudota, N., & Tasso, C. (2009). Recommending New Tags Using DomainOntologies. Proceedings of the 2009 IEEE/WIC/ ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology, September 15-18, Milan, Italy, vol. 3, pp. 409-412. ISBN 978-0-7695-3801-3, doi http://doi.ieeecomputersociety.org/10.1109/WI-IAT.2009.313.

365

Towards Disambiguating Social Tagging Systems

Bateman, S., Brooks, C., & McCalla, G. (2006). Collaborative tagging approaches for ontological metadata in adaptive e-learning systems. In Proceedings of the 4th International Workshop on Applications of Semantic Web Technologies for E-Learning (SWEL’06). (Lecture Notes in Learning and Teaching, (pp. 3-12). Dublin: National College of Ireland. Casoto, P., Dattolo, A., Omero, P., Pudota, N., & Tasso, C. (2008). Accessing, Analyzing, and Extracting Information from User Generated Contents. Chapter XXVII of this handbook. Christiaens, S. (2006). Metadata Mmechanisms: From Oontology to Ffolksonomy… and Bback. In R. Meersman, Z. Tari, & P. Herrero (Eds.), On the Move to Meaningful Internet Systems 2006: OMT 2006 Workshops, (Vol. 4278, (pp. 199-207). Berlin-Heidelberg: Springer-Verlag. Dattolo, A., Tasso, C., Farzan, R., Kleanthous, S., Bueno Vallejo, D., & Vassileva, J. (Eds.). (2009). Proceedings of International Workshop on Adaptation and Personalization for Web 2.0 (AP- WEB 2.0 2009), Trento, Italy, June 22, 2009, CEUR Workshop Proceedings, ISSN 1613-0073, online http://ceur-ws.org/Vol-485. Echarte, F., Astrain, J., Cordoba, A., & Villadangos, J. (2007). Ontology of Ffolksonomy: A new Mmodeling Mmethod. Semantic Aauthoring, Aannotation, and Kknowledge Mmarkup (SAAKM), K-CAP 2007. Retrieved on September 14, 2008, from http://ceur-ws.org/Vol-289/p08.pdfhttp:// ceur-ws.org/Vol-289/p08.pdf Farrell, S., Lau, T., & Nusser, S. (2007). Building Communities with People-Tags. In C. Baranauskas, P. Palanque, J. Abascal, & S.D.J. Barbosa (Eds), Proceedings of Human-Computer Interaction - INTERACT 2007, 11th IFIP TC 13 International Conference (pp. 357-360). BerlinHeidelberg: Springer-Verlag.

366

Fisher, D. H. (1998). From thesauri towards ontologies? In W. Mustafa el Hadi, J. Maniez &, S. Pollitt (Eds.), Structures and relations in knowledge organization: Proceedings of the 5th International ISKO Conference (pp. 18-30). Würzburg: Ergon. Golder, A. S., & Huberman, B. A. (2005). The structure of collaborative tagging. Information Dynamics Lab. Retrieved on June 10, 2008, from http://arxiv.org/ftp/cs/papers//0508/0508082. pdfhttp://arxiv.org/ftp/cs/papers//0508/0508082. pdf. Golder, S. A., & Huberman, B. A. (2006). Usage patterns of collaborative tagging systems. Journal of Information Science, 32(2), 198–208. doi:10.1177/0165551506062337 Gordon-Murnane, L. (2006). Social bookmarking, folksonomies, and Web 2.0 tools. Searcher Mag Database Prof, 14(6), 26–38. Heymann, P., & Garcia-Molina, H. (2006). Collaborative Ccreation of Ccommunal Hhierarchical Ttaxonomies in Ssocial Ttagging Ssystems. (Technical. Report. InfoLab). Retrieved on June 10, 2008, from http://dbpubs.stanford.edu:8090/ pub/showDoc.Fulltext?lang=en&doc=2006-10& format=pdf& compression=&name=2006-10.pdfhttp://dbpubs.stanford.edu:8090/pub/showDoc. Fulltext?lang=en&doc=2006-10& format=pdf& compression=&name=2006-10.pdf. Hotho, A., Jaschke, R., Schmitz, C., & Summe, G. (2006). BibSonomy: A Social Bookmark and Publication Sharing System. In Proceedings of the Conceptual Structures Tool Interoperability. Workshop at the 14th International Conference on Conceptual Structures, July. Retrieved June 30, 2008, from http://www.kde.cs.uni-kassel.de/ jaeschke/paper/hotho06bibsonomy.pdf.

Towards Disambiguating Social Tagging Systems

ISO 2788. (1986). Guidelines for the establishment and development of monolingual thesauri (2nd ed.). Genevea: International Organization for Standardization. ISO 5963. (1985). Documentation: mMethods for examining documents, determining their subjects, and selecting indexing terms. Genevea: International Organization for Standardization. Jacob, E. K. (2004, Winter). Classification and categorization: Aa difference that makes a difference. [from http://sils.unc edu/~fu/IR/fulltext/jacob_classification_and_categorization.pdfhttp://sils.unc. edu/~fu/IR/fulltext/ jacob_classification_and_categorization.pdf.]. Library Trends, 52(3), 515–540. Retrieved on June 10, 2008. Jurafsky, D., & Martin, J. H. (2000). Speech and Language Processing: an introduction to natural language processing, computational linguistics, and speech recognition. Upper Saddle River, New Jersey: Prentice Hall. Kome, S. H. (2005). Hierarchical Ssubject Rrelationships in Ffolksonomies, Master’s thesis, University of North Carolina at Chapel Hill, Chapel Hill, NC. Maedche, A. (2002). Emergent semantics for ontologies – support by an explicit lexical layer and ontology learning. IEEE Intelligent Systems - Trends & Controversies, 17(1), 78–86. Manning, C., & Schütze, H. (1999), Foundations of Statistical Natural Language Processing, Cambridge, MA: MIT Press. Marlow, C., Naaman, M., Boyd, D., & Davis, M. (2006). HT06, Ttagging Ppaper, Ttaxonomy, Fflickr, Aacademic Aarticle, TtorRead. In [New York: ACM Press.]. Proceedings of Hypertext, 2006, 31–39.

Mathes, A. (2004, December). Folksonomies Ccooperative Cclassification and Ccommunication Tthrough Sshared Mmetadata. December 2004. Retrieved on June 10, 2008, from http:// www.adammathes.com/academic/computer-mediated-communication/folksonomies.pdfhttp:// www.adammathes.com/academic/computermediated-communication/folksonomies.pdf. Mika, P. (2005). Ontologies Aare Uus: A Uunified Mmodel of Ssocial Nnetworks and Ssemantics. In Y. Gil, E. Motta, V. R. Benjamins, & M. Musen (Eds.), Proceedings of the 4th International Semantic Web Conference (ISWC2005) (pp. 522-536). Berlin-Heidelberg: Springer-Verlag. Mitkov, R. (2003). The Oxford Handbook of Computational Linguistics. Oxford/New York: Oxford University Press. Morrison, P. J. (2008). Tagging and searching: Search retrieval effectiveness of folksonomies on the World Wide Web. Information Processing & Management, 44(4), 1562–1579. doi:10.1016/j. ipm.2007.12.010 Ohmukai, I., Hamasaki, M., & Takeda, H. (2005). A proposal of Community-based folksonomy with RDF metadata. In Proceedings of the 4th International Semantic Web Conference (ISWC2005). Parameswaran, M., & Whinston, A. B. (2007). Research issues in social computing. Journal of the Association for Information Systems, 8(6), 336–350. Peckham, A. (2005). Urban Dictionary: Fularious Street Slang Defined. Kansas City: Andrews McMeel. Quintarelli, E. (2005). Folksonomies: pPower to the people. Proceedings of ISKO Italy-UniMIB mMeeting. Retrieved on June 10, 2008, from http:// www.iskoi.org/doc/folksonomies.htmhttp://www. iskoi.org/doc/folksonomies.htm.

367

Towards Disambiguating Social Tagging Systems

Quintarelli, E., Resmini, A., & Rosati, L. (2006). FaceTag: Integrating Bbottom-up and Ttop-down Cclassification in a Ssocial Ttagging Ssystem. Paper presented at the EuroIA Conference, BerlinHeidelberg, DEGermany. Schmitz, P. (2006), Inducing ontology from flickr tags. In Collaborative Web Tagging workshop. Proceeding of the 15th International World Wide Web Conference. Retrieved June 30, 2008, from http://www.ibiblio.org/www_tagging/2006/22. pdf. Sinclair, J., & Cardew-Hall, M. (2008). The folksonomy tag cloud: when is it useful? Journal of Information Science, 34(1), 15–29. doi:10.1177/0165551506078083 Smith, G. (2004). Folksonomy: sSocial classification. Atomiq (August 3, 2004). Retrieved on June 10, 2008, from http://atomiq.org/archives/2004/08/folksonomy_social_classification.htmlhttp://atomiq.org/archives/2004/08/ folksonomy_social_classification.html. Specia, L., & Motta, E. (2007). Integrating folksonomies with the Semantic Web. Proceedings of the ESWC 2007, Workshop “Bridging the Gap between Semantic Web and Web 2.0”, (pp. 624639). Retrieved on June 10, 2008, from http:// www.eswc2007.org/pdf/eswc07-specia.pdfhttp:// www.eswc2007.org/pdf/eswc07-specia.pdf. Spiteri, L. F. (2007). The structure and form of folksonomy tags: The road to the public library catalog. Information Technology and Libraries, 26(3), 13–25. Spyns, P., De Moor, A., Vandenbussche, J., & Meersman, R. (2006). From folksologies to ontologies: how the twain meet. In R. Meersman, Z. Tari et al. (Eds.), OTM 2006, LNCS 4275 (pp. 738-755). Berlin-Heidelberg: Springer-Verlag. Taylor, A. G. (2004). The organization of information. Westport/London: Libraries Unlimited.

368

Van Damme, C., Hepp, M., & Siorpaes, K. (2007). FolksOntology: An Iintegrated Aapproach for tTurning fFolksonomies into oOntologies. Proceedings of the ESWC 2007 Workshop “Bridging the Gap between Semantic Web and Web 2.0”, (pp. 71-84). Retrieved on June 10, 2008, from http:// www.kde.cs.uni-kassel.de/ws/eswc2007/proc/ FolksOntology.pdfhttp://www.kde.cs.uni-kassel. de/ws/eswc2007/proc/FolksOntology.pdf. Vander Wal, T. (2005). Folksonomy Ddefinition and Wwikipedia. Off the Top (November 2, 2005). Retrieved in June 2008, from http://vanderwal.net/ random/category.php?cat=153http://vanderwal. net/random/category.php?cat=153. Veres, C. (2006). The Language of Folksonomies: What Tags Reveal About User Classification. LNCS (3999/2006). Natural Language Processing and Information Systems (pp. 58-69). BerlinHeidelberg: Springer-Verlag. Weinberger, D. (2007). Everything is miscellaneous: the power of the new digital disorder. New York: Times Books. Wright, A. (2008). Glut: Mastering Information Through the Ages. New York: Cornell University Press. Wu, X., Zhang, L., & Yu, Y. (2006). Exploring social annotations for the semantic web. In Proceedings of the 15th international conference on World Wide Web (pp. 417-426). Yee, K. P., Swearingen, K., Li, K., & Hearst, M. (2003). Faceted metadata for image searching and browsing. Proceeding of ACM CHI 2003, (pp. 401-408). Retrieved on June 10, 2008, from http://flamenco.berkeley.edu/papers/flamencochi03.pdfhttp://flamenco.berkeley.edu/papers/ flamenco-chi03.pdf.

Towards Disambiguating Social Tagging Systems

AdditionAl reAdings Yngve, V. H. (1995). Syntax and the problem of multiple meaning. In W. N. Locke and D. A. Booth (Eds.), Machine Translation of Languages (pp. 208-26). New York: John Wiley and Sons. Zhang, L., Wu, X., & Yu, Y. (2006). Emergent Semantics from Folksonomies: A Quantitative Study. Journal on Data Semantics VI: Special Issue on Emergent Semantics. LNCS(4090/2006) (pp. 168-186). Berlin-Heidelberg: Springer-Verlag.

Key terMs And definitions Categorization: The basic cognitive process of arranging into classes or categories. The word classification identifies especially the system used in libraries for describe, with a specific notation, the content of a book. Categorization is a more theoretical theory Folksonomies: Folksonomy is the result of personal free tagging of information and objects (anything with a URL) for one’s own retrieval. The tagging is done in a social environment (usually shared and open to others). Folksonomy is created from the act of tagging by the person consuming the information. Metadata: Data that describes other data. The term may refer to detailed compilations such as data dictionaries and repositories that provide a substantial amount of information about each data element. It may also refer to any descriptive item about data, such as a title field in a media file, a field of key words in a written article or the content in a meta tag in an HTML page Ontologies: Definition (computer_science): An ontology is a collection of concepts and relations among them, based on the principles of classes, identified by categories, properties that

are different aspects of the class and instances that are the things Tags: A tag is a generic term for a language element descriptor. The set of tags for a document or other unit of information is sometimes referred to as markup, a term that dates to pre-computer days when writers and copy editors marked up document elements with copy editing symbols or shorthand Taxonomies: Taxonomy is the science of classification according to a pre-determined system, with the resulting catalogue used to provide a conceptual framework for discussion, analysis, or information retrieval. In theory, the development of a good taxonomy takes into account the importance of separating elements of a group (taxon) into subgroups (taxa) that are mutually exclusive, unambiguous, and taken together, include all possibilities Thesaurus: A thesaurus is the vocabulary of an indexing language, that is a controlled list of accepted terms. The role of a thesaurus is to specify a preferred term (descriptor) to be use in indexing and to establish relationships between concepts at different levels: define synonyms, specify hierarchies, individuate related terms Web 2.0: Web 2.0 is the popular term for advanced Internet technology and applications including blogs, wikis, RSS and social bookmarking. The expression was originally coined by O’Reilly Media and MediaLive International in 2004, following a conference dealing with nextgeneration Web concepts and issues Web 3.0: Web 3.0 is defined as the creation of high-quality content and services produced by gifted individuals using Web 2.0 technology as an enabling platform. Web 3.0 refers to specific technologies that should be able to create the Semantic Web.

369

Section 6

Web Quality, Trust, Security, and Effort Estimation

371

Chapter 21

Modeling Content Quality for the Web 2.0 and Follow-on Applications Roberto Sassano University of Trento, Italy Luis Olsina National University of La Pampa, Argentina Luisa Mich University of Trento, Italy

AbstrACt The consistent modeling of quality requirements for Web sites and applications at different stages of the life cycle is still a challenge to most Web engineering researchers and practitioners. In the present chapter, we propose an integrated approach to specify quality requirements to Web sites and applications. By extending the ISO 9126-1 quality views characteristics, we discuss how to model internal, external quality, and quality in use views taking into account not only the software features, but also the own characteristics of Web applications. Particularly, we thoroughly analyze the modeling of the content characteristic for evaluating the quality of information–so critical for the whole Web application eras. The resulting model represents a irst step towards a multi-dimensional integrated approach to evaluate Web sites at different lifecycle stages.

introdUCtion The Web platform –as other Internet service- is approaching to the 20 years of existence. Web sites and applications (WebApps) built on top of this platform have become one of the most influential

developments in the recent computing history with real impact in individuals and communities. For instance, as cited in Murugesan (2007: p.8), websites such as amazon.com, google.com, yahoo.com, myspace.org, wikipedia.org, ebay.com, youtube. com, napster.com, blogger.com and saloon.com are considered as the top ten websites that changed the world. Some of these sites have over 100 million

DOI: 10.4018/978-1-60566-384-5.ch021

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Modeling Content Quality for the Web 2.0 and Follow-on Applications

users (yahoo.com, ebay.com, myspace.com); about 1 billion visits a day (wikipedia.org); over 1 billion searches per day (google.com). As the Web usage continues to grow, users expect WebApps to be more mature and useful with regard to the functions, services and contents delivered for their intended goals and tasks at hand. In addition, they expect these functions and contents be meaningful, accurate, suitable, usable, reliable, secure, personalised, and ultimately with perceived quality. Despite the major breakthrough in Web developments and technologies there are some Web Engineering branches still in their infancy, with lack of a wide consensus: Modeling of quality requirements for WebApps at different lifecycle stages is one of them. In this chapter we propose – based on related research literature and some ideas published in Olsina et al (2005) – an integrated approach to specify quality requirements for contents, functionalities and services to WebApps. Websites, from the very beginning, were conceived as document-, content-oriented artifacts. Few years later, websites started to provide not only contents but also software-like functionalities and services. Since then, WebApps have at a fast pace emerged for many different sectors as e-commerce, e-learning, e-entertainment, and so forth. After that epoch named Web 1.0, a new recent era that considers a set of strategies and technologies focusing on social networking, collaboration, integration, personalization, etc. is emerging – currently called Web 2.0 and follow-ons. We argue that WebApps will continue being centered on functionalities, services and contents independently of the new technologies and collaboration strategies. However, as aforementioned a challenging issue still is how to specify and assess the quality and the quality in use of WebApps at different lifecycle stages since their intrinsic nature will continue being both content and function oriented. In the present work, by reusing and extending some ISO 9126-1 (2001) quality views’ characteristics, we discuss how to model internal, external

372

quality, and quality in use views taking into account the previous concerns. Particularly, we thoroughly discuss the modeling of the content characteristic for evaluating the quality of information, which is not included in this standard and is so critical for the whole WebApp eras. Moreover, we underline how some specific attributes could be considered to Web 2.0 and follow-on applications. So far no integrated approach for WebApps that considers all these dimensions at the same time to different lifecycle stages has been issued as we know – not even the in-progress SQuaRE project intended to replace and to make more consistent many ISO standards related to quality models, measurement and evaluation processes (ISO 15939, 2002; ISO 14598-1, 1999). The rest of this chapter proceeds as follows. First, we give an overview of the Web eras as well as the unique intrinsic features of WebApps compared to traditional software applications. A review of the ISO quality models to specify nonfunctional requirements at different lifecycle stages of software products follows; besides, we also discuss what is missing in these quality models for the Web features highlighted in the previous section. Then we illustrate the proposed extension of the ISO quality models’ characteristics in order to specify the quality of information. In the sequel, we analyse our proposal in the light of related work, and, finally, we draw the main conclusions and future trends.

feAtUres of web erAs And APPliCAtions: An overview The first goal of this section is to introduce the main features of the Web evolution, but without deepening in the different dimensions of eras, while they are detailed commented in Murugesan (2007); the second goal is to outline the distinctive intrinsic features of WebApps compared to traditional software applications. The first WebApps can be grouped in the Web 1.0 era, and they can be categorized into static and

Modeling Content Quality for the Web 2.0 and Follow-on Applications

dynamic; most recent WebApps can be grouped in the so-called Web 2.0 era as per O’Reilly (2005). These allow people collaborate, share and edit information online in seemingly new ways of interaction. Others applications could be grouped in the mobile Web era, where applications could offer some additional features such as personalization and context-aware capabilities and services; and the semantic Web era (Berners-Lee et al, 2001), where applications offer the automatic processing of information meaningfully. According to Murugesan (2007: p.11), at the very beginning, most websites were a collection of static HTML pages intended to deliver just information. After a while, WebApps became dynamic delivering pages created on the fly. The ability to create pages from the data stored on databases enabled Web developers to provide customized information to the visitors, in addition to more complex functionalities. However, “… these sites provided primarily one way interaction and limited user interactivity. The final users had no role in content generation and no means to access content without visiting the sites concerned”. In the last few years, new classes of Web 2.0 WebApps have emerged. Examples of these applications include social networking sites such as ‘myspace.com’, media sharing sites such as ‘youtube.com’ and collaborative authoring sites such as ‘wikipedia.org’. We can feature Web 2.0 WebApps as follows: •

User generated content: if we check the rating of the most popular sites, e.g. at http:// www.alexa.com/site/ds/top_sites (accessed by August, 2008), we can igure out that currently, after ‘google.com’ and ‘yahoo. com’, one of the most visited is ‘youtube. com’. Maybe the latter is the best example to explain how big has become the Web 2.0 phenomenon and what user generated content means.







User active involvement: the active participation of users is one of the most important features, which has changed the way users have to interact with WebApps. Now users’ role can be deined as ‘prosumer’ since s/ he is content producer and consumer at the same time. WebApps like blogs are signiicant examples. Sharing information: in social network people share interests and activities. Examples of these applications are ‘myspace.com’, ‘facebook.com’ and ‘orkut.com’. Endless beta condition: considering the above three features it is easy to understand that Web 2.0 apps are mostly dynamic and under ongoing changes. Wikipedia is for instance continually subject to editing by users so there is no a ‘inal version’ of it.

According to Murugesan (2007: p.11), these new sites “… offer smart user interfaces and builtin facilities for users to generate and edit content presented on the Web and thereby enrich the content base. Besides leveraging the users’ potential in generating content, these applications provide facilities to keep the content under the user’s own categories (tagging feature) and access it easily (Web feed tool). These new breed of Web sites are also able to integrate multiple services under a rich user interface” Also he remarks “Web 2.0 is gradually becoming recognized as an important collection of technologies, business strategies, and social trends. As a result of these developments the Web is changing fast from a static, one-way publishing medium to a highly interactive, dynamic application platform for fielding new kinds of applications”. On the other hand, as commented in Olsina et al. (2005: pp. 120-121), WebApps taken as product or product in use entities (without talking about distinctive features of Web development processes) have their own features distinct from traditional software, namely:

373

Modeling Content Quality for the Web 2.0 and Follow-on Applications















374

WebApps will be even more informationdriven, content-oriented. Most WebApps, besides the increasing support to functionalities and services –seen since the dynamic Web 1.0 era- will continue aiming at showing and delivering multimedia information. This info orientation is a basic feature stemming from the early, static Web 1.0 era, which is currently empowered by the Web 2.0 and follow-on applications; WebApps are interactive, user-centered, hypermedia-based applications, where the user interface plays a central role; thus, they will continue to be highly focused on the look and feel. Web interfaces ought to be easy to use, understand, operate, and navigate because thousands of users with different proiles and capabilities interact with them daily; in addition, WebApps currently have to cope with a variety of display devices and screen sizes. The Web embodies a greater bond between art and science than that encountered in software applications. Aesthetic and visual features of Web development are not just a technical skill but also a creative, artistic skill. Internationalization and accessibility of contents for users with various disabilities are real and challenging issues in WebApps, independently of eras. Searching and browsing are two basic functionalities used to ind and explore information and services. These capabilities are inherited from hypermedia-based applications and will continue to be there. Security is a central issue in data- transaction-oriented WebApps. Likewise, performance is also critical for many WebApps, although both are also critical features for traditional applications. The entire Web applications, and its parts, are often evolutionary pieces of information – even more for the current Web 2.0 sites.





The medium where WebApps are hosted and delivered is generally more unpredictable than the medium where traditional software applications run. For instance, unpredictability in bandwidth maintenance, or in server availability, can affect the perceived quality that users could have. Contents privacy and intellectual property rights of materials are current issues too. They involve ethic, cultural, and legal aspects as well. Most of the time, it is very dificult to establish legal boundaries due to the heterogeneity of legislations in different countries, or even worse, to their absence.

Most of the above features make a WebApp a particular artifact. However, like any software application, it also involves source and executable code, persistent structured data, and functional requirements, architectural design and testing specifications as well. Ultimately, many of the above characteristics will influence the way non-functional requirements are modeled and instantiated. We need to deal not only with usability, functionality –and its sub-characteristics like accuracy, suitability, function and data security, and interoperability-, efficiency, reliability and maintainability, as in traditional software products but also with information quality, i.e. with content accuracy, suitability, accessibility, legal compliance and so forth.

iso/ieC 9126-1 QUAlity views: A disCUssion iso software Quality Perspectives: A review According to Olsina et al (2005) the quality of an entity – e.g. a product as a software program or a WebApp- is easy to recognize but hard to define, measure, and evaluate. The concept of quality is

Modeling Content Quality for the Web 2.0 and Follow-on Applications

not simple and atomic, but a multidimensional and relative one –as also indicated in Mich et al (2003a). Common practice assesses quality by means of the quantification of lower abstraction concepts, such as attributes of entities. The attribute can be briefly defined as a measurable property of an entity category. Therefore, quality –and its sub-dimensions, called characteristics and sub-characteristics in the ISO 9126-1 standard-, is an abstract relationship between attributes of an entity and a specific information need, with regard to its purpose, context, and user’s viewpoint (Olsina et al, 2007a). On account of such multidimensionality, a quality model, which specifies the relationships between characteristics, subcharacteristics and associated attributes, is usually necessary. Further, an instantiated quality model can in the end be calculated and evaluated in order to determine the level of satisfaction achieved. The ISO 9126-1 (2001) standard –and the ongoing SQuaRE project- distinguishes among three different approaches to software product quality, viz. internal quality, external quality, and quality in use. These three views of quality in ISO 9126-1 can be summarized as follows (Olsina et al, 2005: p. 114): 1.

2.

Internal Quality, which is specified by a quality model (i.e. a set of six characteristics –functionality, usability, reliability, efficiency, maintainability and portability- and a set of sub-characteristics per each characteristic are prescribed), and can be measured and evaluated by static attributes of documents such as specification of requirements, architecture, or design; pieces of source code; and so forth. In early phases of a software or Web lifecycle, we can evaluate and control the internal quality of these by-products, but assuring internal quality is not usually sufficient to assure external quality. External Quality, which is specified by a quality model (equally to the previous model), and can be measured and evaluated

3.

by dynamic properties of the running code in a computer system, i.e. when the module or full application is executed in a computer or network simulating as closely as possible the actual environment. In late phases of a software lifecycle (mainly in different kinds of testing, or even in the acceptance testing, or furthermore in the operational state of a software or WebApp), we can measure, evaluate and control the external quality of these late products, but assuring external quality is not usually sufficient to assure quality in use. Quality in Use, which is specified by a quality model (i.e. a set of four characteristics –effectiveness, productivity, safety and satisfaction- is prescribed), and can be measured and evaluated by the extent to which a software or WebApp meets specific user needs in an actual, specific context of use.

The internal quality definition in ISO 9126-1 is “the totality of attributes of a product that determines its ability to satisfy stated and implied needs when used under specified conditions”; the external quality definition is “the extent to which a product satisfies stated and implied needs when used under specified conditions”; and the quality in use definition is “the capability of the software product to enable specified users to achieve specified goals with effectiveness, productivity, safety and satisfaction in specified context of use”. These three slightly different definitions of quality refer particularly to the software product when it is used under specified conditions and context of use, so making it clear that quality is not an absolute concept, but depends on specific conditions and context of use by specific users. The same quality model have been maintained both to internal and external views. For instance, functionality characteristic is defined as “the capability of the software product to provide functions which meet stated and implied needs when the software is used under specified conditions”.

375

Modeling Content Quality for the Web 2.0 and Follow-on Applications

In turn, its five sub-characteristics, namely: accuracy, suitability, security, interoperability and compliance are defined in Table 1. Functionality –from the non-functional requirement point of view- is concerned with what the software does to fulfill the user needs (software is defined as a set of programs with the associated data and documentation). Considering for example the accuracy and security definitions both function and data attributes can be associated in order to assess them. This is also valid for WebApps where programs and persistent, structured data (and its effects) are there as well. Note that in the information quality literature data and information quality are treated very often as synonymous terms but we make a clear difference as we discuss later on. Besides, usability characteristic is defined

as “the capability of the software product to be understood, learned, used and attractive to the user, when used under specified conditions”. Usability is subdivided in turn into five sub-characteristics –understandability, learnability, operability, attractiveness, and compliance – which are defined in Table 2. Usability and its sub-characteristics apply also to specifying internal and external quality requirements for WebApps. Lastly, the core aim in designing an interactive (software or Web) application is to meet the user needs; that is, to provide degrees of excellence or quality in use by interacting with the application and by performing its tasks comfortably. Regarding the spirit of the ISO 9126-1 standard, quality in use is the end user’s view of the quality of a running system containing software, and is

Table 1. Definition of functionality sub-characteristics prescribed in ISO 9126-1 for specifying internal and external quality requirements Sub-characteristic

Definition

Accuracy

The capability of the software product to provide the right or agreed results or effects with the needed degree of precision.ISO Note: This includes the needed degree of precision of calculated values.

Suitability

The capability of the software product to provide an appropriate set of functions for specified tasks and user objectives.

Security

The capability of the software product to protect information and data so that unauthorised persons or systems cannot read or modify them and authorised persons or systems are not denied access to them.ISO Note 1: This also applies to data in transmission. ISO Note 2: Safety is defined as a characteristic of quality in use, as it does not relate to software alone, but to a whole system.

Interoperability

The capability of the software product to interact with one or more specified systems.

Functionality Compliance

The capability of the software product to adhere to standards, conventions or regulations in laws and similar prescriptions.

Table 2. Definition of usability sub-characteristics prescribed in ISO 9126-1 for specifying internal and external quality requirements Sub-characteristic

Definition

Understandability

The capability of the software product to enable the user to understand whether the software is suitable, and how it can be used for particular tasks and conditions of use.

Learnability

The capability of the software product to enable the user to learn its application.

Operability

The capability of the software product to enable the user to operate and control it.

Attractiveness

The capability of the software product to be attractive to the user.

Usability compliance

The capability of the software product to adhere to standards, conventions, style guides or regulations relating to usability.

376

Modeling Content Quality for the Web 2.0 and Follow-on Applications

measured and evaluated in terms of the result of using the software, rather than by properties of the software itself. A software (or Web) product’s internal and external quality attributes are the cause, and quality in use attributes are the effect. Table 3 shows the definition of the four quality in use dimensions. Ultimately, taking into account meaningful software or WebApp attributes for internal quality are a prerequisite to achieve the required external behavior, and considering meaningful software attributes to external behavior are a prerequisite to achieve quality in use.

Usefulness of iso Quality Models for webApps: A discussion Consequently, we argue that the ISO software quality models introduced above are also applicable to a great extent to intermediate and final lifecycle Web products. A discussion of this statement follows, as well as how we could adapt specific features of WebApps –those outlined in the previous section- into quality models. Note this discussion is an extension of that made in Olsina et al (2005, pp. 121-123) reinforcing the same line of argumentation. Like any software production line, the Web lifecycle involves different stages of its products, whether in early phases as inception and development, or in late phases as deployment, operation, and evolution. To assure the quality of products, we can plan to do it by evaluating

and controlling the quality from intermediate products to final products. Thus, to the general question, if we can apply to WebApps the same ISO internal and external quality, and quality in use models, the natural answer is yes – and we believe this almost does not need more explanation. Nevertheless, to the more specific question whether we can use the same six prescribed quality characteristics (and their sub-characteristics) for internal and external quality requirements, and the four characteristics for quality in use requirements, our answer is yes for the latter, but some other considerations might be taken into account for the former. In particular, as highlighted in the previous section, the very nature of WebApps is a mixture of information (media) content, functionalities and services. We argue that the set of six characteristics, i.e. functionality, usability, reliability, efficiency, maintainability and portability, and their sub-characteristics respectively, are not well suited (or they were not intended) to specify requirements for information quality. At this point, we would like to introduce the slight difference in meaning between data and information terms. A piece of data is raw material; even though it has a degree of information. Data come from attribute measurements, facts, formula calculations, etc. and basically they have categorical or numerical values, a scale type, and may also have an explicit procedure to produce or collect them. Structured data sets are often represented in databases. On the other hand, information has

Table 3. Definition of the four quality in use characteristics prescribed in ISO 9126-1 Characteristic

Definition

Effectiveness

The capability of the software product to enable users to achieve specified goals with accuracy and completeness in a specified context of use.

Productivity

The capability of the software product to enable users to expend appropriate amounts of resources in relation to the effectiveness achieved in a specified context of use.

Safety

The capability of the software product to achieve acceptable levels of risk of harm to people, business, software, property or the environment in a specified context of use.

Satisfaction

The capability of the software product to satisfy users in a specified context of use.ISO Note: Satisfaction is the user’s response to interaction with the product, and includes attitudes towards use of the product.

377

Modeling Content Quality for the Web 2.0 and Follow-on Applications

an added value over data. This is, information is the meaningful interpretation of data for a given context, purpose, and user viewpoint. Usually, a traditional software program is a mixture of functions and data. On the other side, a webpage is very often content oriented, i.e. is intended to deliver information (usually unstructured semantically, from the Semantic Web representation point of view). For example, this article could be hyperlinked and posted as content (textual information) Web pages. Also a webpage component, e.g. a shopping cart, can edit an item quantity and recalculate prices (a function over data). Therefore, to follow the thread of our argument, the central issue is how we can specify and model the content quality of WebApps from the internal and external quality viewpoints. We need to deal not only with usability, functionality, efficiency, reliability, and maintainability –as modeled by ISO to software products- but also with the content quality characteristic, which in turn can be subdivided into content accuracy, content suitability, content accessibility, and legal compliance sub-characteristics. As a consequence, we propose to include the content characteristic and its subcharacteristics in the internal and external quality model of the ISO standard, as shown in Figure 1. A point worth mentioning is that in the spirit of the ISO 9126-1 standard is stated that “evaluating product quality in practice requires characteristics beyond the set at hand”.

On the other hand, the quality in use definition may be rephrased as “the capability of the software or WebApp product to enable specified users to achieve specified goals with effectiveness, productivity, safety and satisfaction in specified context of use”. Note that these four characteristics are influenced not only by the usability, functionality, reliability, efficiency, and content of a WebApp, but also by two resource components of the context of use. The context of use depends on both the infrastructure (i.e. the computer, network, or even the physical working medium) and the user-oriented goals (i.e. the supported WebApp tasks and the properties of the user type such as level of IT training, expertise, age, and cultural issues as well). Tasks are the steps or sub-goals undertaken to reach an intended application goal. Care should be taken when generalizing the results of any assessment of quality in use to another context of use with different types of users, tasks, or environments. See for example the quality in use case study for an e-learning WebApp in Covella & Olsina (2006), where user tasks were designed not only to deal with services and functions but with contents as well. Next, we thoroughly discuss the proposed ISO internal and external requirement extension in order to include the content characteristic for WebApps independently of Web eras. However, specific attributes associated to content sub-

Figure 1. ISO model for internal and external quality along with our extension to the content characteristic

378

Modeling Content Quality for the Web 2.0 and Follow-on Applications

characteristics may be considered for Web 2.0 and follow-on applications, as we comment in the next section.

eXtending the iso QUAlity Models to inforMAtion QUAlity As aforementioned, information has added value over data, and hereafter we consider Web information as Web content, which can be textual or other media. Hence, we define content as “the capability of a Web product to deliver information which meets stated and implied needs when used under specified conditions”. Taking into account previous contributions made in the area of information quality –as we will discuss in the related work section-, we have primarily identified for the content characteristic four major sub-characteristics, which can help to evaluate information quality requirements for WebApps. This initial proposal was made in Olsina et al (2005, pp. 122-123). In the present work we contribute by redefining them (see Table 4) and also by extending and defining the sub-sub-characteristics (see Figure 2). The content sub-characteristics are: first, content accuracy, which addresses the very intrinsic nature of the information quality; second, content suitability, which addresses the contextual nature of the information quality; it emphasizes the importance of conveying the appropriate

information for user-oriented tasks and goals; in other words, it highlights the quality requirement that content must be considered within the context of use and the intended audience; third, content accessibility, which emphasizes the importance of technical and representational aspects in order to make Web contents more accessible for users with various disabilities as regarded in the WAIWCAG initiative (W3C, 1999); and lastly, legal compliance, as defined in Table 4. In Figure 2, we define the sub-sub-characteristics for both the content accuracy and content suitability dimensions. Some of them could be just treated either as measurable attributes or as sub-dimensions to which attributes should be further associated accordingly. On the other hand, Figure 2 does not show the content accessibility sub-characteristic decomposition for instance. For space reasons we will model its sub-sub-characteristics and attributes in other related paper; however, the reader may surmise that the WCAG’s 14 content accessibility guidelines and the 65 checkpoints are very useful for this purpose, and could be reused completely, even the available tools. Accessibility guidelines deal with both representational and technical aspects such as navigability and orientation, device independence transformations, equivalent alternatives to auditory and visual content, internationalization, among others.

Table 4. Definition of the proposed Content sub-characteristics for specifying information quality requirements in order to extend ISO’s internal and external quality models Sub-characteristic

Definition

Content Accuracy

The capability of a Web product to deliver information that is correct, credible and current.

Content Suitability

The capability of a Web product to deliver information with the right coverage, added value, and consistency, considering the specified user tasks and goals.

Content Accessibility

The capability of a Web product to deliver information that is accessible for all users (with or without disabilities) taking into account both technical and representational aspects.

Content Legal Compliance

The capability of the Web product to adhere to standards, conventions, and legal norms related to content as well as to intellectual property rights.

379

Modeling Content Quality for the Web 2.0 and Follow-on Applications

Figure 2. Definition of the proposed Content Accuracy and Suitability sub-dimensions for specifying internal and external information quality requirements

1. Content Accuracy 1.1. Correctness, the extent to which information is reliable in the sense of being free of errors. Note: information errors can be both syntactic and semantic; for example, semantic errors can be checked by known experts, or by peer contributors like in Wikipedia relying on the concept of ‘wisdom of crowds’. 1.2. Believability (synonym: Credibility), the extent to which the information is reputable, objective, and verifiable. 1.2.1. Authority (synonym: Reputability), the extent to which the source of the information is trustworthy. Note: it is well known that almost anyone can become a Web publisher and collaborate with content edition, even more in Web 2.0 applications. Although it is one of the Web 2.0’s great strengths also poses new evaluation challenges. 1.2.2. Objectivity, the extent to which the content (i.e., information or facts) is unbiased and impartial. 1.2.3. Ve r i f i a b i l i t y ( s y n o n y m : Traceability), the extent to which the

380

owner and/or author of the content can be verified. Note: This also poses new evaluation challenges for Web 2.0 applications, because it relies often on social control mechanisms. On the other hand, we can trace versions, as in Wikipedia for instance. 1.3. Currency (synonym: Up-to-dateness), the extent to which the information can be identified as up to date. Note: Attributes of information currency like creation, posted, and revised dates can be used.

2. Content suitability 2.1. Value-added, the extent to which the content can be novel, beneficial, and contribute to react to a given user for the task at hand. 2.1.1. Novelty (synonym: Freshness), the extent to which the information is fresh and contributes to make new decisions for an intended user goal. 2.1.2. Beneficialness, the extent to which the information is advantageous and contributes to make new decisions for an intended user goal. Note: e.g., in marketing this attribute is related to the

Modeling Content Quality for the Web 2.0 and Follow-on Applications

image of the company or organization as it is projected by the website and also to the target identified as relevant by the marketing people (Mich et al., 2003b). 2.1.3. Reactiveness, the extent to which the information is compelling and contributes to react for an intended user goal. 2.2. Coverage, the extent to which the content is appropriate, complete but also concise for the task at hand to a given user. 2.2.1. Appropriateness, the extent to which the information coverage fits to an intended user goal. 2.2.2. Completeness, the extent to which the information coverage is the sufficient amount of information to an intended user goal. Note: e.g. see the shopping cart case study performed in Olsina et al (2007b, p. 417), where five attributes for completeness and appropriateness have been measured and evaluated. 2.2.3. Conciseness, the extent to which the information coverage is compactly represented without being overwhelming. Note: e.g. to the writing for the Web heuristic, usually, shorter is better. 2.3. Consistency, the extent to which the content is consistent to the site’s piece of information or page with respect to the intended user goal.

As regard the last sub-characteristic –Content Legal Compliance- the pervasive nature of WebApps demands an increasingly attention to laws, regulations and policies. To align Web content and the uneven international regulations is a challenging issue: both cooperation with lawyers and supporting tools are needed. Also, sub-characteristics for Legal Compliance have to be defined accordingly to a given WebApp specific rules. Even if we have

identified some attributes, we are not addressing this aspect in this work. In addition to the above content sub-characteristics, others to information architecture and organization could be addressed. Many of these sub-characteristics, such as global understandability –implemented by mechanisms that help to understand quickly the structure and content of the information space of a Web site like a table of contents, indexes, or a site map-, learnability, and also operability and attractiveness, can be related to the usability characteristic. Besides, other particular features of WebApps such as search and navigation functionalities can be specified in the functionality sub-characteristics; for example, are the basic and advanced search suitable for the end user? Or, are they tolerant of misspelled words and accurate in retrieving documents? In the same way, we can represent link and page maturity attributes, or attributes to deficiencies due to browsers’ compatibility into the reliability sub-characteristics. On the other hand, from the quality in use perspective, we have proposed to use the ISO model. However, for the satisfaction characteristic, specific (questionnaire) items for evaluating quality of content should be included. Also, for other quality in use characteristics such as effectiveness and productivity, specific user-oriented evaluation tasks that include performing actions with content and functions can be designed and tested. Finally, we have performed some preliminary studies. One of them was a quantitative evaluation for a shopping cart of a Web 1.0 app, where the content accuracy and suitability sub-characteristics have intervened –see Olsina et al (2007b). Recently, a qualitative evaluation was made with the aim of comparing two WebApps that belong to the tourism domain –see Sassano (2008). In particular, we have evaluated the content accuracy and suitability in addition to the accessibility and legal compliance sub-characteristics on opodo. co.uk, a Web 1.0 app, and on tripadvisor.com, which belong to the Web 2.0. The external quality

381

Modeling Content Quality for the Web 2.0 and Follow-on Applications

evaluation was based on checklist considering a question for each sub-dimension of Figure 2 as well as to the content accessibility and content legal compliance sub-dimensions. Two experts have intervened in the inspection. Though we have non-conclusive evidence from this study, some initial comments and observations can be drawn about content features that distinguish Web 1.0 from Web 2.0 apps. First of all, it should be highlighted that the process of content production in Web 1.0 apps pursues rather a top-down approach, i.e., only content providers supply information to users. Conversely, in Web 2.0 apps this process becomes rather bottom-up; that is, mainly final users upload and update information. Moreover, content is submitted to a social control mechanism since users can share, edit, or comment content of other users, as happens in blogs, wikis, and social networks. In fact, initial observations have shown that some kind of information may be considered more accurate and suitable in ‘tripadvisor.com’ than in ‘opodo.co.uk’; particularly, information referring for instance to hotel review, location comment. In contrast, information as flight timetables, holiday price lists, etc. can be considered more accurate and suitable in ‘opodo.co.uk’. Lastly, in general terms we argue that the WebApp’s content quality does not depend on the kind of applications –whether Web 1.0 or Web 2.0; however, some kind of contents and services are more appropriate for Web 1.0 apps, while others for Web 2.0. Ultimately, we can state the content subcharacteristics we have specified for evaluation purposes can be applied to all WebApps, independently from which era they belong.

relAted worK The model presented in this paper, as an extension of the ISO 9126-1 quality models, has been elaborated taken also into account related researches about dimensions of data and information qual-

382

ity. We next remark the main information quality models, often called frameworks in the literature, developed over the last years. The reader can, however, find broader reviews for WebApps, for instance, in Knight & Burn (2005) and Parker et al (2006). It is worth mentioning that the difference in meaning between data and information –as remarked in the previous section- has often been neglected in these quality frameworks. Moreover, very often –and in some cases explicitly- these terms are used interchangeably. One of the first studies intended to categorize data quality dimensions was made by Strong et al (1997). The focus in their work was on considering the dimensions of data quality for three user roles, i.e. data consumer, data custodian, and data producer. According to the authors, high quality data is data that is fit for use by the intended users. They developed a framework made up by four categories –intrinsic, contextual, representational, and accessibility- including 16 dimensions of data quality. Specifically, the intrinsic category indicates that information has its own quality per se. It contains four dimensions: accuracy, objectivity, believability, and reputation. The accessibility category states that information must be easily accessible but secure. It includes: accessibility, and security dimensions. The third category is contextual data quality, which indicates that information should be provided in time and in appropriate amounts. It includes: relevancy, valueadded, timeliness, completeness, and amount of data. The last category is representational data quality, which focuses on format of data/information and its meaning. It includes: interpretability, ease of understanding, concise representation, and consistent representation. As a matter of fact, the Strong et al data quality framework was initially developed for traditional information systems. Nevertheless, this model has been used for WebApps too. For instance, Katerattanakul & Siau (1999) reuse the four categories and the characteristics including free-of-error webpage content, workable and relevant hyperlinks, and

Modeling Content Quality for the Web 2.0 and Follow-on Applications

the navigational tools provided for accessibility. In a recent study, Caro et al (2007) have reused the Strong et al framework for modeling data quality of Web portals from the data consumer viewpoint. All these data quality frameworks neither consider different lifecycle stages of a WebApp and therefore different quality models as we propose, nor make any distinction between data and information quality. A different slightly way to model and evaluate the quality of information for a WebApp –both at page and site level- is proposed by Alexander & Tate (1999). They take into account six dimensions (criteria) such as authority, accuracy, objectivity, currency, orientation, and navigation. Authors include a checklist for a step-by-step quality evaluation to some kind of websites, namely, for the advocacy, business, informational, personal, news, and entertainment sectors. They evaluate information rather than data without considering different information quality models at different WebApp lifecycle stages. The first published study about extending the ISO 9126 model has been made by Zeist & Hendricks (1996). In a nutshell, their extended model consists of adding some sub-characteristics for each characteristic, with the aim of specifying data/information quality. Unfortunately, this study is quite limited because at that moment the ISO standard did not consider the internal, external, and quality in use views –as these were included only in the 2001 revised standard. Finally, as mentioned in the introduction, there exists an ongoing SQuaRE project that proposes harmonizing many ISO standards related to quality models, measurement and evaluation processes (ISO 15939, 2002; ISO 14598-1, 1999). According to Vaníček (2005: p.139) “these standards have not a unified terminology and do not fully reflect the current state of art in software engineering”. In his contribution he proposes a data quality model regarding the three ISO views, but these models are just for data (data as a new entity) separated of the quality models for software functions. As

the author is aware “the main problem concerning the development of new SQuaRE series of standard and also concerning the data quality standard is the enormous volume of standardisation documents … If we extend the number and span of standards, nobody will use them” Vaníček (2005: p.145). Conversely to the SQuaRE approach, our aim is modeling nonfunctional requirements for WebApps’ functions, services and content, taking into account the three integrated quality models and Web lifecycle views.

fUtUre trends And ConClUding reMArKs While users are becoming more and more mature in the use of WebApps and tools, there is greater demand for the quality of these applications that should match real user needs in actual working environments. Hence a natural trend is the greater demand of quality planning and assessing in early Web development stages to assure the quality of delivered products. On the other hand, most WebApps, besides the increasing support to functionalities and services will continue aiming at showing and delivering multimedia content. This basic feature stemming from the early Web 1.0 applications is currently empowered by the Web 2.0 and follow-on applications. Web 2.0 applications rely strongly on actual users sharing, collaborating and performing content tasks in real contexts of use. So other probably trend will be the greater demand for Web 2.0 application evaluations from the quality in use perspective. This could imply more pervasive user-centered testing and evaluations that is nowadays happening. As concluding remarks, in this chapter we have proposed how to specify quality requirements for functionalities, services and content for WebApps employing a minimalist and integrated approach. By reusing and extending the ISO 9126-1 quality models’ characteristics, we have

383

Modeling Content Quality for the Web 2.0 and Follow-on Applications

discussed the need of modeling and adding the content characteristic for evaluating the quality of information. Specifically, we have argued that the internal and external quality models with the set of six characteristics, i.e. functionality, usability, reliability, efficiency, maintainability and portability, and their sub-characteristics respectively, are not sufficient to specify WebApps’ information quality requirements. As a consequence, we have proposed to include in both models the content characteristic and its sub-characteristics, i.e. content accuracy, content suitability, content accessibility, and content legal compliance. Besides, from the quality in use perspective, we have proposed to use the same ISO model. At the most, by redesigning accordingly the tasks in order to include content goals, and, by adding to the satisfaction characteristic specific (questionnaire) items for evaluating quality of content it would be enough. Ultimately, we have tried to give a minimalist and integral solution to the current concern which is how to identify and model WebApps’ quality and quality in use requirements at different lifecycle stages. Lastly, we did not discuss in this chapter –for space reasons- how to split the content sub-characteristics into measurable attributes or into some questionnaire items. Nor we gave clues about how different methods and techniques could be fit in for measurement and evaluation purposes. This will be discussed in an upcoming paper running a real case study as well.

ACKnowledgMent This work and line of research are supported by the following projects: UNLPam 09/F037 and PICTO 11-30300, from the Science and Technology Agency, Argentina. The Roberto Sassano’s scholarship to GIDIS_Web group was also funded by the University of Trento grant, Italy, in the framework of the University of Trento and National University of La Pampa agreement.

384

referenCes W3C. (1999). WWW consortium, Web content accessibility guidelines 1.0. Retrieved on August 27, 2008, from http://www.w3.org/TR/WAIWEBCONTENT/ Alexander, J. E., & Tate, M. A. (1999). Web wisdom: How to evaluate and create information quality on the Web. Mahwah, NJ: Erlbaum. Berners-Lee, T., Hendler, J., & Lassila, O. (2001). The Semantic Web. Scientific American. Caro, A., Calero, C., Mendes, E., & Piattini, M. (2007). A probabilistic approach to Web portal’s data quality evaluation. In Proceedings of the IEEE 6th International Conference on the Quality of Information and Communications Technology (pp. 143-153), Lisbon, Portugal. Covella, G., & Olsina, L. (2006). Assessing quality in use in a consistent way. In ACM Proceedings, Int’l Congress on Web Engineering, (ICWE06) (pp. 1-8), San Francisco. ISO 14598-1. (1999). International standard, information technology-software product evaluationpart 1: General overview. ISO 15939. (2002). Software engineering-software measurement process. ISO/IEC 9126-1. (2001). Software engineering— software product quality—part 1: Quality model. Geneva: Int’l Org. for Standardization. Katerattanakul, P., & Siau, K. (1999). Measuring information quality of Web sites: Development of an instrument. Proceedings of the 20th International Conference on Information Systems (pp. 279-285), Charlotte, NC. Knight, S., & Burn, J. (2005). Developing a framework for assessing information quality on the World Wide Web. Information Science Journal, 8.

Modeling Content Quality for the Web 2.0 and Follow-on Applications

Mich, L., Franch, M., & Cilione, G. (2003a). The 2QCV3Q quality model for the analysis of Web site requirements. Journal of Web Engineering, 2(1-2), 105–127. Mich, L., Franch, M., & Gaio, L. (2003b). Evaluating and designing the quality of Web sites. IEEE MultiMedia, 10(1), 34–43. doi:10.1109/ MMUL.2003.1167920 Murugesan, S. (2007). Web application development: Challenges and the role of Web engineering. In G. Rossi, O. Pastor, D. Schwabe & L. Olsina (Eds.), Web engineering: Modeling and implementing Web applications (pp. 7-32). Springer. O’Reilly, T. (2005). What is Web 2.0? Design patterns and business models for the next generation of software. Retrieved on August 27, 2008, from www.oreillynet.com/pub/a/oreilly/tim/ news/2005/09/30/what-is-web-20.html Olsina, L., Covella, G., & Rossi, G. (2005). Web quality. In E. Mendes & N. Mosley (Eds.), Web engineering (pp. 109–142). Springer. Olsina, L., Papa, F., & Molina, H. (2007a). How to measure and evaluate Web applications in a consistent way. In G. Rossi, O. Pastor, D. Schwabe & L. Olsina (Eds.), Web engineering: Modeling and implementing Web applications (pp. 385–420). Springer. Olsina, L., Rossi, G., Garrido, A., Distante, D., & Canfora, G. (2007b). Incremental quality improvement in Web applications using Web model refactoring. (LNCS 4832, pp. 411-422). In M. Weske, M.-S. Hacid & C. Godart (Eds.), 1st Int’l Workshop on Web Usability and Accessibility (IWWUA’07), WISE 2007 Workshops. Springer. Parker, M., Moleshe, V., De la Harpe, R., & Wills, G. (2006). An evaluation of information quality frameworks for the World Wide Web. In 8th Annual Conference on WWW Applications, Bloemfontein, Free State Province, South Africa.

Sassano, R. (2008). Content quality model for Web 2.0 Web sites. Unpublished doctoral dissertation (in Italian), University of Trento, Italy. Strong, D., Lee, Y., & Wang, R. (1997). Data quality in context. Communications of the ACM, 40(5), 103–110. doi:10.1145/253769.253804 Vaníček, J. (2005, September 21-21). Software and data quality. In Proceedings Conference Agricultural Perspectives XIV, Czech University of Agriculture in Prague. Zeist, R. H. J., & Hendriks, P. R. H. (1996). Specifying software quality with the extended ISO model. Software Quality Management IV– Improving Quality, BCS (pp. 145-160).

Key terMs And definitions Content Suitability: which represents the capability of a Web product to deliver information with the right coverage, added value, and consistency, considering the specified user tasks and goals. Content Accessibility: which represents the capability of a Web product to deliver information that is accessible for all users (with or without disabilities) taking into account both technical and representational aspects. Content Accuracy: which represents the capability of a Web product to deliver information that is correct, credible and current. Content Characteristic: it represents the capability of a Web product to deliver information which meets stated and implied needs when used under specified conditions. It is composed of sub-characteristics such as content accuracy, suitability, accessibility and legal compliance. Content Legal Compliance: which represents the capability of the Web product to adhere to standards, conventions, and legal norms related to content as well as to intellectual property rights.

385

Modeling Content Quality for the Web 2.0 and Follow-on Applications

Information Quality: which is a quality dimension for a Web product and it is represented by the content characteristic in the quality model. Quality Model: it specifies the quality perspective and the relationships between quality characteristics, sub-characteristics and associated attributes of an entity, which allow the further

386

evaluation or estimation for given information needs. Web 2.0: it is gradually becoming recognized as an important collection of Web technologies, business strategies, and social trends which allow people collaborate, share and edit information online in seemingly new ways of interaction.

387

Chapter 22

A New Web Site Quality Assessment Model for the Web 2.0 Era Minseok Pang University of Michigan at Ann Arbor, USA Woojong Suh Inha University, South Korea Jinwon Hong Inha University, South Korea Jongho Kim Hyundai Research Institute, South Korea Heeseok Lee Korea Advanced Institute of Science and Technology, South Korea

AbstrACt To ind a strategy for improving the competitiveness of Web sites, it is necessary to use comprehensive, integrated Web site quality dimensions that effectively discover which improvements are needed. Previous studies on Web site quality, however, seem to have inconsistent and confusing scopes, creating a need of reconciliation among the quality dimensions. Therefore, this chapter attempts to provide a Web site quality model that can comprise all the quality scopes provided by previous studies. The relationship between the speciic dimensions of the quality model and the characteristics or merits of Web 2.0 was discussed in this chapter with actual Web site examples. It is expected that this study can help Web sites improve their competitiveness in the Web 2.0 environment.

introdUCtion To date, the World Wide Web (WWW) has become rapidly prevalent in our society, tremendously influ-

encing over the length and breadth of human being’s life and business environment. Now it is being utilized as essential media or social infrastructure for personal living and organizations’ business. In spite of these rapid changes and increase in utilization,

DOI: 10.4018/978-1-60566-384-5.ch022

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

A New Web Site Quality Assessment Model for the Web 2.0 Era

the dot-com bubble in 2001 formed a negative perspective on further development in the WWW (Byrne, 2000; Howcroft, 2001). However, a new concept of Web 2.0 brings us new and innovative changes, breaking these negative views. Such changes include ones in business practices using Web and users’ behavioral patterns as well as service development on the Web. Bughin & Manyika (2007) and O’Reilly (2005), all of which are easily found in the contemporary websites. Changes related to the Web take place in user behaviors and rich user experience. In terms of changes in user behaviors, users are getting actively involved in producing and sharing contents and discussing about the contents, as collective intelligence which collects various knowledge and experiences together through user interaction widespread, and UCC (User Created Content) or UGC (User Generated Content) becomes popular (Murugesan, 2007; O’Reilly, 2005; Ogawa & Goto, 2006). Technologies such as AJAX (Asynchronous JavaScript and XML), Mashup, Flux, etc. have been applied in website development, providing richer user experiences through robust functions and elegant user interfaces (Murugesan, 2007; O’Reilly, 2005; Ogawa & Goto, 2006). As a result, websites have been greatly improved in quality, focusing more on interactions between websites and users or among users. With these improvements, websites have become an important means for firms in managing the relationship with customers and partners and with internal employees as well (Bughin & Manyika, 2007). It is important for website administrators to make endeavors to improve website quality by actively utilizing the characteristics and merits of Web 2.0 in order to improve competitiveness of websites. At least two requirements have been emerged to support these kinds of endeavors as follows. First, it is necessary to have comprehensive, integrated dimensions which cover the entire website lifecycle (Murugesan, Deshpande, Hansen, & Ginige, 2001) from conception, development and deployment, to continual refinement,

388

update, and upgrade. Previous studies on website quality, however, have different scopes for considering qualities and so are complementary with each other but incomplete. Second, it is necessary to understand what the characteristics or merits of Web 2.0 are in those terms. One cannot find any explanation on website quality in previous studies, though. Therefore, the purpose of this study is to suggest a website quality assessment model that consists of quality dimensions which are highly relevant to websites’ user experience. We also would like to offer a guideline to help website developers provide services with higher quality and distinctive features of Web 2.0. The outline of this paper is as follows. The next section reviews the related studies, and the following section describes a comprehensive set of website quality dimensions, discussing the related features and advantages of Web 2.0 with their illustrations. The final section concludes this chapter with discussions of future research directions.

bACKgroUnd Many studies have provided various website quality dimensions. Among them, we have examined and summarized 28 studies which comprise quality dimensions that are crucial in the usage of website, rather ones relevant to website developers including maintainability and recoverability. Only those with significant characteristics especially from the perspective of customers are explained in this paper. Barnes and Vidgen have suggested and revised eQual method (previously called WebQual) to assess website quality in various domains of websites including university websites (Barnes & Vidgen, 2000), auction websites (Barnes & Vidgen, 2001), Internet bookstores (Barnes & Vidgen, 2005), and information-extensive websites (Barnes & Vidgen, 2003). Their latest method, eQual 4.0, consists of usability, information, and service interaction instruments (Barnes & Vidgen, 2005).

A New Web Site Quality Assessment Model for the Web 2.0 Era

Szymanski & Hise (2000) constructed a conceptual model of e-satisfaction (i.e. customer satisfaction in e-retailing) which is the outcome of consumer perceptions of online convenience, merchandising (product offerings and product information), site design, and financial security. In this study, they empirically proved such four quality dimensions to be significantly correlated with users’ satisfaction. While most of the research on website quality has conducted empirical research, interestingly, Zhang, Zhu, Greenwood, & Huo (2001) proposed different methods adopted from software engineering techniques - information flow analysis and software hazard analysis. They modelled activities and structures of e-commerce systems using UML diagram, and then, from software hazard analysis, they identified risk factors that are linked to quality dimensions (content, time, presentation). From an analogy between businesses and buildings, Kim, Lee, Han, & Lee (2002) identified three major quality dimensions of websites - the structural firmness (firmitas) that refers to the solidity of the system structure in overcoming all expected and unexpected threats, the functional convenience (utilitas) which means the provision of convenient functions for customers’ processing of transaction activities, and the representational delight (venustas) that indicates interface aspects of the websites with which the user comes into contact. They also suggested six sub-dimensions (refer to Table 1). Agarwal & Venkatesh (2002) presented the dimensions and sub-dimensions of websites quality based on Microsoft Usability Guidelines (MUG). Their major dimensions consist of content, ease of use, promotion, made-for-themedium, and emotion. By adopting a heuristic evaluation procedure, they calculated relative importance weights of their dimensions, and, in order to show the usefulness of their dimensions, they assessed actual websites from five kinds of industries - airline, bookstore, auto manufacturer, and car rental.

Palmer (2002) tried to explore the relationships among website usability, design, and performance metrics. He employed download delay (speed of access and display rate), navigability (organization, arrangement, layout, and sequencing), site content (amount and variety of information), interactivity (customization and interactivity) and responsiveness (feedback and FAQs) as independent variables, and user satisfaction, the likelihood of return, and the frequency of use as dependent variables. From three empirical studies using juries, third-party (Alexa) ratings, and a software agent, he demonstrated significant associations between independent and dependent variables. Santos (2003) formulated two major quality dimensions of websites – the incubative dimensions (easy of use, appearance, linkage, structure and layout, and content), and the active dimensions (reliability, efficiency, support, communication, security, and incentives). These dimensions were identified by focus group interviews with Internet users, and based on service quality theory. The DeLone and McLean’s Information Systems (D&M IS) Success Model is one of the most frequently cited success model in the literature. This study extended their original D&M IS Success Model by adding new constructs, ‘service quality’ and ‘net benefits’ for covering e-commerce system. As a result, this model included system quality, information quality, and service quality as quality factors, considering use, user satisfaction, and net benefits as success factors (DeLone & McLean, 2004). Webb & Webb (2004) discussed SITEQUAL, a set of quality dimensions for e-commerce websites. Using constructs from service quality dimensions and information quality dimensions, they identified desired B2C website quality factors (reliability, assured empathy, tangibility, navigability, relevant representation, accuracy, and security) and minimum B2C website quality factors (reliability, assured empathy, perceived usability, and trustworthiness).

389

A New Web Site Quality Assessment Model for the Web 2.0 Era

Table 1. Summary of previous studies on website quality dimensions Researcher

Target

Research Methodology

Proposed Quality Dimensions Access to website, content, graphics, structure, user friendliness, navigation, usefulness, unique features, online transactions, site usage fee

Bell & Tang (1988)

General websites

Online survey with users

Abels, White, & Hahn (1999)

Academic business community websites

User-based design process

Olsina, Godoy, Lauente, & Rossi (1999)

Academic websites

Web-site QEM methodology

Szymanski & Hise (2000)

E-commerce websites

Online survey with users

Liu & Arnett (2000)

E-commerce websites

Mail survey with webmasters

Information quality, learning capability, playfulness, system quality, system use, service quality

Bauer & Scharl (2000)

Non-profit organizations websites

Snapshot analysis, longitudinal analysis, comparative analysis

Content, interactivity, navigation

Mateos, Mera, Gonzalez, & Lopez (2001)

University websites

Judgmental evaluation

Barnes & Vidgen (2001)

Auction websites

Online survey with users

Site quality, information quality, interaction quality, auction quality

Zhang et al. (2001)

E-commerce websites

Information flow analysis, hazard analysis

Content(correctness, relevance, accuracy, completeness, security, temporal media), time(timeliness, response time), presentation(navigability, layout, consistency)

Kim et al. (2002)

E-commerce websites, online stock brokerages, search portals, online games

Online survey with users

Dustin, Rashka, & McDiarmid (2002)

General websites

N/A

Janda, Trocchia, & Gwinner (2002)

E-commerce websites

Offline interviews with users

Agarwal & Venkatesh (2002)

Airline websites, bookstore websites, Auto manufacturer websites, Car rental websites

Heuristic evaluation procedure, offline survey with users

Palmer (2002)

Fortune 500 corporations’ websites

Offline survey with student users, software agents, Alexa ratings

Madu & Madu (2002)

E-commerce websites

N/A

Ranganathan & Ganapathy (2002)

E-commerce websites

Online survey with users

390

Use, content, structure, linkage, search, appearance

Functionality, usability, efficiency, site reliability

Convenience, merchandising, site design, financial security

Accessibility, speed, navigability, contents quality

Structural firmness(internal stability, external stability), functional convenience(information gathering, order processing), representational delight(system interface, communication interface) Functional, security, performance and security, compatibility, usability(language, layout and graphics, information architecture, user interface) Performance, access, security, sensation, information Content(relevance, media use, depth and breadth, current and timely information), easy of use(goal, structure, feedback), promotion, made-for-the-medium(community, personalization, refinement), emotion(challenge, pilot, character strengths, pace) Download delay, navigation/organization, interactivity, responsiveness, information/content Performance, features, structure, aesthetics, reliability, storage capability, system integrity and security, trust, responsiveness, differentiation and customization, web store policies, reputation, assurance, empathy Information content, design, security, privacy

A New Web Site Quality Assessment Model for the Web 2.0 Era

Table 1. continued Researcher Santos (2003)

Target E-commerce websites

Research Methodology

Proposed Quality Dimensions

Focus group interviews

Incubative dimensions (easy of use, appearance, linkage, structure and layout, content) active dimensions (reliability, efficiency, support, communication, security, incentives)

Online survey with users

Customer care (easy to communicate, security of transaction and personal information, prompt response), information benefit (reliability, completeness, covering personal interest, up-to-date information), interaction benefit

Gounaris & Dimitriadis (2003)

Portals

DeLone & McLean (2003)

E-commerce websites

N/A

Information (completeness, ease of understanding, personalization, relevance, security), system (adaptability, availability, reliability, response time, usability), service (assurance, empathy, responsiveness)

Kim & Stoel (2004)

E-commerce website of apparel products

Mail survey with female users

Informational task-to-fit, tailored communication, online completeness, relative advantage, visual appeal, innovativeness, emotional appeal, consistent image, easy of understanding, intuitive operations, response time, trust

Webb & Webb (2004)

E-commerce websites

Mail survey with users

Service quality (reliability, responsiveness, assurance, empathy, tangibility), information quality (accessibility, contextual, representational, intrinsic)

Gonzalez & Palacios (2004)

200 Spanish commercial websites

Qualitative evaluations

Parasuraman et al. (2005)

E-commerce websites (Amazon, Walmart)

Online survey with users

Lee & Kozar (2006)

E-commerce websites(travel, electronics)

Survey with users and managers/designers

Information quality (relevance, currency, understandability), service quality (empathy, reliability, responsiveness), systems quality (navigability, response time, personalization, telepresence, security), vender-specific quality (awareness, reputation, price saving)

Accessibility, speed, navigability, site content E-s-qual (efficiency, fulfillment, system availability, privacy) e-recs-qual (responsiveness, compensation, contact)

Moustakis et al. (2006)

Commercial websites

Survey with users

Content (utility of content, completeness of information, subject specialization, reliability of content, syntax of content), navigation (convenience of navigation tools, identify of site, means of navigation, links to another site, ease of use of navigation tools, search engines), structure and design (order of elements, loading speed, site map, information structure, software requirements, browser compatibility, real time information), appearance and multimedia (graphics representation, readability of content, multimedia/ images/voice/video), uniqueness (uniqueness of content, aesthetics in content presentation, design characteristics)

Wulf, Schillewaert, Muylle, & Rangarajan (2006)

General websites

Online survey with users

Content (credibility, currentness, relevance, sufficiency), organization (design, interactivity, navigation, readability, speed), technology (progressiveness, reliability)

Lin (2007)

E-commerce websites

Offline survey with users

System quality (website design, interactivity), information quality (informativeness, security), service quality (responsiveness, trust, empathy)

Grigoroudis et al. (2008)

Commercial websites

Survey with users

Relevance, usefulness, reliability, specialization, architecture, navigability, efficiency, layout, animation

From an online survey with users of Amazon and Wal-Mart website, Parasuraman, Zeithaml, & Malhotra (2005) conceptualized and tested empiri-

cally two sets of multiple-item scales for measuring the service quality delivered by websites on which customers go shopping online, namely

391

A New Web Site Quality Assessment Model for the Web 2.0 Era

E-S-QUAL and E-RecS-QUAL. E-S-QUAL consists of efficiency, fulfillment, system availability, and privacy, whereas E-RecS-QUAL comprises responsiveness, compensation, and contact. Lee & Kozar (2006) and Moustakis, Tsironis, & Litos (2006) estimated the importance of website quality dimensions and detail properties in each dimension using AHP (Analytical Hierarchy Process). Lee & Kozar (2006) derived qualities such as information quality, service quality, systems quality, vender-specific quality, etc. from a survey from general users and website administrators/ designers, and estimated the importance measure. Moustakis et al. (2006) derived qualities such as content, navigation, structure and design, appearance and multimedia, uniqueness, etc. and estimated the importance of the quality dimensions and their detailed properties with examining three websites. Lin (2007) derived quality dimensions of B2C E-commerce website affecting customer satisfaction with considering different aspects such as system quality, information quality and service quality as suggested in the DeLone & McLean (2003) model. Grigoroudis, Litos, Moustakis, Politis, & Tsironis (2008) analyzed user satisfaction and performance for 3 Greek websites and performed a satisfaction benchmarking analysis using website estimation metrics such as relevance, usefulness, reliability, specialization, architecture, navigability, efficiency, layout, animation, etc. Considering the fact that Web 2.0 is a new, emerging paradigm that is intertwined with the change in the Internet culture and the progress in the Internet technologies, we argue that the prior studies listed in Table 1 are hard to be applied in Web 2.0 environment for the following reasons. First, the previous works only offered a limited set of criteria to evaluate a variety of emotional factors that websites provide users. It should be noted that such representative Web 2.0 services as social network services or blogs successfully fulfill various emotional needs that users have

392

including empathy, intimacy, playfulness, and emotion. Second, all the studies but three in Table 1 focused on a specific type of websites. Therefore, suggested quality dimensions only explain quality characteristics of the targeted websites, failing to offer a more comprehensive set of website quality dimensions. A growing number of websites that are recently developed are augmenting customer experiences to a greater extent with a more variety of services. For example, Amazon.com is not only selling products as an e-commerce website but supporting its customers to build various communities for their interests. A similar example is Kyobo Book Store (www.kyobobooks.co.kr/), the largest book seller in Korea, offers a blog service in which users can post articles with their reading experiences. Administrators in these websites have to consider both quality factors of e-commerce websites both ones of blogging sites. A broader, more general collection of criteria for evaluating website quality is needed in Web 2.0 environment. The limitations in previous studies and the changes that Web 2.0 paradigm has brought call for identifying new website quality dimensions appropriate for Web 2.0 environment.

A new set of website QUAlity diMensions for web 2.0 In this section, we present a new set of website quality dimensions for use in Web 2.0 applications. This set consists of five first-order dimensions and 25 second-order dimensions (shown in Figure 1) which are discussed in detail later in this section. These dimensions were developed by a synthesis approach through following steps. First, we examined thoroughly the conceptual and operational definitions of each dimension and measurement that are exhibited in all papers shown in Table 1. Second, we eliminated trivial dimensions (e.g., ‘number of documents,’ or ‘number of file types’ in Bauer & Scharl (2000)) and ones that are specific to a certain

A New Web Site Quality Assessment Model for the Web 2.0 Era

Figure 1. New website quality dimensions and its comparison with other models proposed earlier

sort of websites (e.g., ‘auction quality’ in Barnes & Vidgen (2001)). Third, we grouped dimensions that have similar meanings and synthesized them into our final dimensions. When producing final dimensions, we considered an appropriate granularity of our dimensions so that all dimensions could contain similar amount and depth of meanings. As a result of these synthesis steps, our metrics was produced as shown in Figure 1. Dimensions shown in Figure 1 can be used as a checklist that service providers need to keep in mind during the entire website lifecycle from inception, development, and deployment to continual improvement, update, and upgrade.

For example, at the inception stage, considering how quality dimensions should be promoted, the providers can devise a development plan to successfully achieve the goal of a website. In addition, at the development and deployment stage, the provider can instruct website designers, programmers, content developers which quality dimensions are critical to respective parties. For example, website designers have to develop website interfaces that support website requirements specified in the previous stage in terms of such quality dimensions as navigation, appearance, layout, emotional empathy, playfulness, and emotion. And at the stages of continual refinement,

393

A New Web Site Quality Assessment Model for the Web 2.0 Era

update, and upgrade, data on customers’ feedbacks can be collected with respect to the quality dimensions above and be analysed for setting up a plan for maintenance and a large scale of renewal. As explained so far, the integrative website quality model suggested by this study enables the service providers to consistently evaluate their websites and to make a sound decision to effectively fulfill website users’ needs.

interface Quality dimension In websites, interface typically focuses on the front-end aspects - what users mostly look at and interact with. In Web 2.0 environment, the interface quality becomes more essential with an increase in users’ demand for more effective interactions (Huang, Li, & Zeng, 2007). Interface quality can be classified into proximity, compatibility, navigation, appearance, and layout dimensions. ‘Proximity’ means the degree to which users can find and reach a website shortly through search engines or URL which users can easily remember. For visitors who do not know the exact URL of a website, the URL should be found instantly using search engines. In addition, the domain name of a website should be easily recognizable and easy to memorize so that visitors can revisit it. ‘Compatibility’ is the degree to which a website can be accessible and usable in various sorts of user environment such as Web browsers or operating systems. The contents and layout of a website should be rendered properly in Internet Explorer, Mozilla Firefox, or other kinds of browsers. Moreover, a website should also be accessible and visible to visitors who use any kinds of operating systems such as Microsoft Window XP or Linux. At the environment of Web 1.0, it has been criticized that compatibility has not been fully supported by many websites and users have experienced a great deal of difficulty in accessing a website that is not compatible with their Web browser or operating systems. A certain website may be rendered differently across Web browsers.

394

Or, some users may not use a transaction functionality development with ActiveX technology if their browser does not support ActiveX. In addition, there exist a variety of problems in accessing a website with a different kind of devices. However, in the era of Web 2.0 which emphasizes the use of standardized technologies in website development, whichever operating systems (OS) or Web browsers users use, they have to be able to use website services without any trouble. Also, users need to easily transfer data between different devices such as mobile phones through Web regardless of OS platform (Ogawa & Goto, 2006). Thus, taking advantage of this compatibility allows to secure more contents users, that is, customers in Web 2.0 environment. ‘Navigation’ refers to the degree to which the sequence of pages is properly arranged, and navigation-related components such as links, labels, or site maps are consistently and efficiently used. Navigation is considered relatively in much more studies as shown in Figure 1. The links between pages in a website should be properly organized and text labels or image icons of the links should correctly indicate where the links connect so that users can navigate wherever they want to go (Brinck, Gergle, & Wood, 2002). At the Web 1.0 environment, navigation is structured with menu bars, indexes, or links with texts or images. However, this sort of navigation structure is a static one which is predefined by website developers. It is very hard to develop a navigation structure which dynamically supports continuously changing users’ interests or issues. The Tag Cloud service, wherein clicking on tags leads to the related contents, is noteworthy among new changes in navigation methods in Web 2.0 environment. Tags are shown with different degrees of highlight to different frequencies of tag register in this service in real time. This Tag Cloud service automatically generates a navigation structure which reflects dynamic changes in users’ interests in a real-time basis. Also, with “Web 2.0” input as a keyword on Fliker Related

A New Web Site Quality Assessment Model for the Web 2.0 Era

Tag Bowers as shown in Figure 2, you get highly relevant tags, as well as search results limited to Web 2.0 (the image in the center). ‘Appearance’ means the degree to which color, graphics, images, font, style, and animations are properly and consistently used. Some other studies mention this dimension as aesthetics or ‘look and feel’. A website should display visually appealing design (Kim et al., 2002). Selecting right colors with consideration of brightness or contrast makes users visually comfortable, while using inconsistent styles throughout a website makes users confused and lose interest in. In the Web 1.0 environment, when a website user requests certain information, the server transmits the entire information to the client, so that it is a very complex task to render dynamic graphics as shown in applications installed in PC. In the Web

2.0, however, that limitation has been overcome by technological progress. Specifically, RIA (Rich Internet Application), providing richer user interface, has been realized in Web 2.0 environment with the use of Ajax, Adobe Flex, Microsoft Silverlight, etc. (Moroney, 2007; Ogawa & Goto, 2006). As shown in Figure 3, the website using Flex can realize rich user interface more dynamically and elegantly. ‘Layout’ implies the degree to which visual elements such as texts, forms, frames, or tables are well placed and organized in a page to be easily recognizable and usable to user. For example, a table too wide to be showed in a screen without a scrollbar is inconvenient for users’ to browse. Brinck et al. (2002) point out that the goals of proper layout are simplicity, consistency, and focus. Nonetheless, layout needs to be designed

Figure 2. Fliker Related Tag Browers provides relevant subjects with links including search results (from http://www.airtightinteractive.com/projects/related_tag_browser/app/) © 2009 Airtight Interactive. Used with permission.

395

A New Web Site Quality Assessment Model for the Web 2.0 Era

to effectively reflect the purpose and strategy of a website (Brown, 2007, pp. 320-321). For example, while Google, which provides search service, uses a very simple type of layout so that users can search for information in a speedy manner (as shown in Figure 4), Yahoo! which provide a broad range of services uses a complicated type of layout to show various services simultaneously (as shown in Figure 5).

system Quality dimension System quality also has been considered one of the most important dimensions in the previous studies. Particularly in Web 2.0 environment, the amount of contents has rapidly increased through UCC (or

UGC) websites such as YouTube (Kanda, 2006), a trend that also raises potential security issues due to the characteristics of Web 2.0 based technology (LaMonica, 2006). Thus, website system quality affects customers’ satisfaction to a great extent. The system quality dimension may consist of availability, efficiency, reliability, and security. ‘Availability’ means whether a website is available and accessible in 24 hours a day and 365 days a year. Frequent interruptions of website services can damage the reputation of a company in relation to its customers. When it is inevitable for a website to shut down for regular maintenance or expansion, the website should make considerable efforts to minimize the closing time and notify users in advance.

Figure 3. Website of rich user interface established with Flex (from http://examples.adobe.com/flex2/ inproduct/lcds/flexstore/flexstore.html)

396

A New Web Site Quality Assessment Model for the Web 2.0 Era

Figure 4. Google website with simple layout (from http://www.google.com/)

Figure 5. Yahoo! website with complicated layout (from http://www.yahoo.com/) © 2009 Yahoo! Inc. Used with permission

397

A New Web Site Quality Assessment Model for the Web 2.0 Era

‘Efficiency’ is the degree to which a website is accessible efficiently with minimum or no delay. Some literature referred to efficiency as ‘performance’ (Dustin et al., 2002), or ‘download delay’ (Palmer, 2002). To achieve high efficiency of a website, not only should sufficient hardware capacity or network bandwidth be secured, but also visual components with suitable size need to be used. For example, forcing users to download too big images or files makes users wait too much, resulting in their discontent. Moreover, a website should be scalable so that it can support unexpected heavy traffic. ‘Reliability’ implies the degree to which a website can perform as intended correctly and consistently without any error or breakdown. A website must be free of any errors which can expose its vulnerability to external attackers as well as make users dissatisfied. When an error occurs, the website should be immediately recovered, and it must be guaranteed that such an error will not take place anymore in the future. Since the use of multimedia data which requires high computing capacity has increased explosively particularly under Web 2.0 environment, a greater level of reliability is very important for the websites to allow users to access multimedia data without interruption. ‘Security’ means the degree to which a website can be robust against all possible attacks or threats from outside and keep private and confidential information securely. Madu & Madu (2002) pointed out that users were worried about providing personal information online since it could potentially fall into the wrong hands or be abused. Therefore, the quality of a website is intertwined with the website’s ability to safeguard and protect information provided to it. In Web 2.0 environment, it is necessary for website administrators to be more careful against potential information leakage due to XSS (Cross-Site Scripting) attacks against Ajax-established websites (Ritchie, 2007), social networking service, wikis, RSS (Really Simple Syndication), etc. (Espiner, 2007).

398

information Quality dimension Information quality is substantially required for not only the success of information-extensive websites such as news and investor relations websites but also that of transaction-extensive websites such as Internet shopping or online stock exchange websites, since users generally seek information about products and services before making purchasing decisions. Businesses based on Web 2.0 business (often called as Web 2.0 company) in particular need to develop their own information for business success (O’Reilly, 2005). For example, Amazon. com has been attracting visitors with book reviews as one source of information, and YouTube has retained users by allowing them to share various kinds of videos. Google Maps has been able to establish a new service model by providing maps and combining with other geographical information (e.g., searching for a pizza shop on Google Map). One can expand those chances to create values through such combination and utilization in Web 2.0 environment, which can be easily found in information quality perspectives. The information quality dimension may consist of completeness, timeliness, comprehensibility, trustworthiness, presentation variability, architecture, and search capability. ‘Completeness’ means the degree to which a website offers a broad range of information which is relevant to users’ needs. In order to satisfy the completeness, a website should offer relevant information enough for users’ goals and decision-making in terms of breadth and depth of information needs. In the Web 1.0 environment, ‘the Collective Intelligence’ was not materialized; instead, information generated by website service providers or their partners is offered unilaterally. In the Web 2.0 environment, on the other hand, numerous users can actively produce information and knowledge through the Collective intelligence, so that the completeness information is enhanced. The process of creating new, complete information has been constantly improved with collective

A New Web Site Quality Assessment Model for the Web 2.0 Era

intelligence achieved through user interaction in Web 2.0. For example, Wikipedia has been making an encyclopedia through participations of many and unspecified users as shown in Figure 6. Users can participate in writing, editing, and improving contents by adding new words or content, correcting etc. ‘Timeliness’ is the degree to which a website provides current and up-to-date information. It is very crucial for website administrators to update continuously and frequently the most current information. For example, e-commerce websites should hide products unavailable in the inventory. There are three reasons why users have difficulty in getting right information at the right time. First, information service providers are not agile in finding out what information users need, so that information creation has been delayed

significantly. Second, no matter how fast they find out users’ needs, they or their third-party partners cannot fulfill all of the various users’ needs by themselves. Lastly, even though they have enough information, they do not have sufficient communication tools for marketing the information to customers. Web 2.0, however, brings new technologies to bridge between information providers and consumers in a real-time basis, so that the timeliness of information can be significantly improved. For example, NAVER (www. naver.com), a representative portal site in Korea, runs blog service and SNS (Social Networking Service), and also RSS service for those two services. Furthermore, it develops a widget which interacts with blogs and communities so that users can have the real-time information on blogs and communities (As shown in Figure 7). When there

Figure 6. Wikipedia achieving the completness of information through collective intelligence (from http://en.wikipedia.org/)

399

A New Web Site Quality Assessment Model for the Web 2.0 Era

Figure 7. NAVER has secured timeliness by providing Widget interacted with blogs and SNS (from http:// www.naver.com) © 2009 NHN Corp. Used with permission

Figure 8. Google website enables users to download documents with four kinds of file formats (Microsoft Media Player, Real Player, HTML, PDF) (from BCE http://www.google.com/)

400

A New Web Site Quality Assessment Model for the Web 2.0 Era

Figure 9. IBM website providing huge amount of information in efficient structure (fromhttp://www.ibm. com/).© 2009 IBM. Used with permission.

are new comments or posts on their own blog or SNS, the widget informs users of the update in real time. Compared to an RSS reader that simply informs updates on new information, the widget can provide users with even more various update information (e.g., the number of blog visitors, updates on neighbor blogs, new neighbors, new comments etc.). Users can provide others with necessary information on blogs or SNS almost in real time using the widget. This system can be very useful in enhancing information timeliness in case of professional blogs or SNS.

‘Comprehensibility’implies the degree to which information a website exhibits is sufficiently understandable even to users who have little background knowledge. For instance, a website in which information contains too many acronyms or jargons which are not familiar with users will make users reluctant to visit the websites again. It is also desirable for the success of a website whose format and representation of information as well as context are common and understandable to general users. ‘Trustworthy’ means the degree to which information in a website is accurate, credible, and verified. No matter how complete, timely,

401

A New Web Site Quality Assessment Model for the Web 2.0 Era

Figure 10. YouTube recommends new key words (e.g., “Also try:” Key words suggested afterward) for the input key word (e.g., adobe) (http://www.youtube.com/)

and comprehensible the information offered by a website is, if it is incorrect or unbelievable, it damages to users’ loyalty to the website. For example, in Wikipedia, there was an incident that incorrect contents were posted and a writer missed out one major person (Terdiman, 2005). Incidents like this can affect the reliability of Wikipedia itself as well as the related information cited in Wikipedia. In an investor relations website, if there were the financial statements posted in the website were not authorized by auditors, administrators should delete the statements or notify that these are not verified (Xiao, Jones, & Lymer, 2002). ‘Presentation variability’ is the degree to which a website presents information in various sorts of format. In an e-commerce website, for example, it is more advisable to present a variety of audio or video clips which emphasize the usefulness of products than to show the information of products simply with text format. It is recommended to make users acquire documents with various file formats such as Webcast, HTML, or Portable Document Format (PDF) format, as shown in Figure 8. In this case, for users’ convenience, it is desirable to provide readers with the software with which they will be able to read the documents. ‘Architecture’ implies the degree to which information in a website is suitably structured so that users can easily access information they seek. The hierarchy of information should be balanced appropriately and not be too deep so that users can

402

browse the website without difficulties (Rosenfeld & Morville, 1998). IBM website is an interesting example (as Shown in Figure 9). A huge amount of contents on one subject is shown in an efficient structure including Getting Started, Learn, Teach, Connect, SSME Highlight, Related Blogs etc. ‘Search capability’ means whether a website facilitates search function or engine and the degree to which search results are accurate and relevant to users’ intention. Using an appropriate set of words in Web search is crucial to retrieve quality search results. At the Web 1.0 environment, it was not straightforward for users to use effective search rules (Barsky & Bar-Ilan, 2005; Holscher & Strube, 2000). However, in the Web 2.0 environment, new technologies are emerging to help users search more easily, such as recommendations for better key words as shown in Figure 10 or suggestions of the most popular keywords Figure 11.

service Quality dimension The service quality dimensions focus on how satisfactorily a website fulfills services. In Web 2.0 environment, users, which are main producers and consumers of contents, share their opinions actively about website services, which greatly affect the success of websites. That is, the service quality of a website has a great impact on the loyalty of users for the website (Lee & Im,

A New Web Site Quality Assessment Model for the Web 2.0 Era

Figure 11. Google recommends other key words automatically in Korean version (unlikely other versions) (http://www.google.co.kr/)

2006). This service quality dimensions may include customization, support, channel diversity, responsiveness, incentive, and compensation. ‘Customization’ means the degree to which a websites provides a user with contents and interface customized according to the user’s particular characteristics or needs. Many studies have mentioned that customization is an important driver of the success of websites (Agarwal & Venkatesh, 2002; Kim & Stoel, 2004; Madu & Madu, 2002). Examples of customization include product recommendation for a certain customer in an e-commerce website or a news clipping service which arranges news a subscriber is interested in. For another instance, as shown in Figure 12, MSN allows users to choose various options such as contents, color, etc., and to design their own layout by delocalizing or deleting squared areas using the click-and-drag method. ‘Support’ is the degree to which a website facilitates supportive information or learning tools which can contribute to enhancing users’

understanding to the website. Websites should support users by providing FAQs, help pages, or simulations. Especially, websites which are likely to require complex activities should provide sufficient help functions including simulation or multilingual services. Figure 13 shows a coordination service using a simulation technology. ’Channel diversity’ implies the degree to which a website offers a variety of channels which enable users to contact staffs conveniently. Just showing email address or telephone number in the bottom of a website is insufficient. To make users able to contact administrators of a website whenever they want, the website should offer various channels such as an email sending form, a built-in board system, or an online chatting function. ‘Responsiveness’ means the degree to which a website fulfills users’ requests or questions promptly. Remaining question made by users to be unanswered for a long time is inexcusable. An e-commerce website should be able to promise as fast a delivery as possible and also keep that

403

A New Web Site Quality Assessment Model for the Web 2.0 Era

Figure 12. MSN allows users to customize in their own style (from http://www.msn.com/)

promise. Complaints from customers must be taken care of immediately through the cooperation between website administrators and staffs in back-end office. ‘Incentive’ is a benefit given by a website that encourages users to visit it continuously and enhances users’ satisfaction and loyalty to the website (Kim et al., 2002; Parasuraman et al., 2005). Examples can be free coupons, discount, prize draw, or gifts. Santos (2003) pointed out that such incentives is useful in encouraging users to try to use the website, engaging in online purchasing, and increasing retention and word-of-month communication. A Korean portal, NAVER (www. naver.com) provides users with incentives such as points for each access of the email service and allows them to use the points for other services (internet phone, Music Streaming, SMS) for free or in a discounted price.

404

‘Compensation’ implies the degree to which a website compensates users for some problems which arose while they were using it. This dimension is suggested through E-RecS-QUAL proposed by Parasuraman et al. (2005). According to their research, compensation is important for the customers’ satisfaction in handling service problems of customers who have encountered.

emotional Quality dimension Emotional quality should be also considered significantly for the success of a website. In this context, Jarvenpaa & Todd (1997) contended that customers are satisfied by not only an extrinsic reward in purchasing products or services but also personal and emotional reward from a purchasing-derived pleasure. Heijden (2004) stressed emotional aspects as well, stating that websites

A New Web Site Quality Assessment Model for the Web 2.0 Era

Figure 13. Simulator for fashion coordination (from CCODDIhttp://ccoddi.com/). © 2009 CCODDI. com, G&G Commerce. Used with permission.

serve hedonic purposes as well as utilitarian ones. The emotional quality dimension includes of assurance, empathy, interaction, playfulness, and emotion. ‘Assurance’ means the extent to which staffs of a website are knowledgeable about their operation and courteous in their responses and can convey trust and confidence to users. Madu & Madu (2002) pointed out that many virtual operations on websites rarely encouraged any direct communication except through e-mail services and stressed that they should make considerable efforts to provide impeccable responses to users. ‘Empathy’ is the extent to which a website can provide caring and individualized attention to customers’ concerns and requests. Madu & Madu (2002) emphasized that such individualized attention is more effective than typical automatic responses in conveying empathy to the customers; such attention of a website should be cognizant of

users’ needs and express concern and understanding of their needs. ‘Intimacy’ implies the degree to which make users feel a close relationship with or affection to a website through interactive processes. Intimacy can be developed while users are interacting with others through active communications. Such an interaction gives users a chance to understand products or services better, increasing their intimacy to a website. In particular, it is regarded as the most important factor for a website to promote user interactions particularly in case of SNS websites such as Cyworld (Figure 14). This kind of interactions are not only one of the most significant characteristics of Web 2.0, but also the fundamental reason for Web 2.0 to be called “Social Web” (Boulos & Wheeler, 2007). ‘Playfulness’ means the degree to which a website can amuse or entertain users. Liu & Arnett (2000) stressed that website designers 405

A New Web Site Quality Assessment Model for the Web 2.0 Era

Figure 14. Cyworld website focusing on users interaction (from http://us.cyworld.com/)

needed to consider hedonic pleasure seriously in designing a website by motivating customers to participate, by promoting customer excitement and concentration, and by including charming features to attract customers and to make them enjoy the visit. For example, playing games in a websites can give users excitement and enjoyment, enhancing the likelihood of revisit to the website (Liu & Arnett, 2000; Rice, 1997). In the Web 1.0 environment, playfulness is fulfilled simply by interesting contents or services, while in the Web 2.0 environment, even simple behaviors in website interfaces provide users a great pleasure. For example, using RIA (Rich Internet Application), users can enjoy dynamic and sophisticated interfaces and have a fun with them. ‘Emotion’ denotes the extent to which a website evokes emotional reactions from users. Agarwal

406

& Venkatesh (2002) included this dimension in their research, and described that its components include challenge, plot, character strength, and pace. Challenge captures the idea of accomplishment rather than functional complexity or obscurity. Plot relates to how a website piques the user’s interest, especially with a storyline. Character strength relates to credibility conveyed by a website. Pace means the extent to which a website provides users with an opportunity to control the flow of information.

ConClUsion This chapter provided a new website quality model for assessing websites comprehensively by closely reviewing 28 other studies on website quality. The

A New Web Site Quality Assessment Model for the Web 2.0 Era

model comprises five first-order dimensions and 25 second-order dimensions, covering almost all the facets of website quality from the previous studies investigated in this paper. Furthermore, this chapter explained each dimension of the model in conjunction with the merits and characteristics of Web 2.0. The website quality model presented in this chapter provides researchers or practitioners with a more comprehensive and balanced perspective on website quality. Specifically, the model may help them more effectively evaluate websites for their development or improvement. Furthermore, this chapter gives the explanations for many dimensions and the relationships to the beneficial features of Web 2.0, which would lead the professionals to generate ideas for improving the websites more competitive in the area of Web 2.0. We plan to extend this study in the following ways: First, the website quality model will be continuously updated along with new study results. Second, we will develop specific measurement items for empirical validation of the model.

referenCes Abels, E. G., White, M. D., & Hahn, K. (1999). A user-based design process for Web sites. OCLC Systems & Services, 15(1), 35–44. doi:10.1108/10650759910257850 Agarwal, R., & Venkatesh, V. (2002). Assessing a firm’s Web presence: A heuristic evaluation procedure for the measurement of usability. Information Systems Research, 13(2), 168–186. doi:10.1287/isre.13.2.168.84 Barnes, S. J., & Vidgen, R. T. (2000). WebQual: An exploration of Web site quality. Paper presented at the the 8th European Conference on Information Systems.

Barnes, S. J., & Vidgen, R. T. (2001). Assessing the quality of auction Web sites. Paper presented at the the 34th Annual Hawaii International Conference on System Sciences. Barnes, S. J., & Vidgen, R. T. (2003). Measuring Web site quality improvements: A case study of the forum on strategic management knowledge exchange. Industrial Management & Data Systems, 103(5), 297–309. doi:10.1108/02635570310477352 Barnes, S. J., & Vidgen, R. T. (2005). The eQual approach to the assessment of e-commerce quality: A longitudinal study of Internet bookstores. In W. Suh (Ed.), Web-engineering: Principles and techniques (vol. 1). Hershey, PA: Idea Group Publishing. Barsky, E., & Bar-Ilan, J. (2005). From the search problem through query formulation to results on the Web. Online Information Review, 29(1), 75–89. doi:10.1108/14684520510583954 Bauer, C., & Scharl, A. (2000). Quantitative evaluation of Web site content and structure. Internet Research: Electronic Networking Applications and Policy, 10(1), 31–43. doi:10.1108/10662240010312138 Bell, H., & Tang, N. K. H. (1998). The effectiveness of commercial Internet Web sites: A user’s perspective. Internet Research: Electronic Networking Applications and Policy, 8(3), 219–228. doi:10.1108/10662249810217768 Boulos, M. N. K., & Wheeler, S. (2007). The emerging Web 2.0 social software: An enabling suite of sociable technologies in health and healthcare education. Health Information and Libraries Journal, 24(1), 2–23. doi:10.1111/j.14711842.2007.00701.x Brinck, T., Gergle, D., & Wood, S. D. (2002). Usability for the Web: Designing Web sites that work. London: Academic Press.

407

A New Web Site Quality Assessment Model for the Web 2.0 Era

Brown, D. M. (2007). Communicating design: Developing Web site documentation for design and planning. Berkely, CA: New Riders.

Heijden, H. d. (2004). User acceptance of hedonic information systems. MIS Quarterly, 28(4), 695–704.

Bughin, J., & Manyika, J. (2007). How businesses are using Web 2.0: A McKinsey global survey. The McKinsey Quarterly.

Holscher, C., & Strube, G. (2000). Web search behavior of Internet experts and newbies. Computer Networks, 33(1-6), 337-346.

Byrne, J. A. (2000). The fall of a dot-com. Business Week, 1, 150–160.

Howcroft, D. (2001). After the goldrush: Deconstructing the myths of the dot.com market. Journal of Information Technology, 16(4), 195–204. doi:10.1080/02683960110100418

DeLone, W. H., & McLean, E. R. (2003). The DeLone and McLean model of information systems success: A ten-year update. Journal of Management Information Systems, 19(4), 9–30. DeLone, W. H., & McLean, E. R. (2004). Measuring e-commerce success: Applying the DeLone & McLean information systems success model. International Journal of Electronic Commerce, 9(1), 31–47. Dustin, E., Rashka, J., & McDiarmid, D. (2002). Quality Web systems: Performance, security, and usability. Upper Saddle River, NJ: AddisonWesley. Espiner, T. (2007). Security threats toolkit: Businesss warned over Web 2.0 security. Retrieved on February 9, 2008, from http://news.zdnet.co.uk/ security/0,1000000189,39286468,00.htm Gonzalez, F. J. M., & Palacios, T. M. B. (2004). Quantitative evaluation of commercial Web sites: An empirical study of Spanish firms. International Journal of Information Management, 24(4), 313– 328. doi:10.1016/j.ijinfomgt.2004.04.009 Gounaris, S., & Dimitriadis, S. (2003). Assessing service quality on the Web: Evidence from business-to-consumer portals. Journal of Services Marketing, 17(5), 529–548. doi:10.1108/08876040310486302 Grigoroudis, E., Litos, C., Moustakis, V. A., Politis, Y., & Tsironis, L. (2008). The assessment of userperceived Web quality: Application of a satisfaction benchmarking approach. European Journal of Operational Research, 187(3), 1346–1357. doi:10.1016/j.ejor.2006.09.017

408

Huang, M., Li, X., & Zeng, X. (2007). Service quality in Web 2.0 electronic commerce: An integrative perspective from typical service and technique-adoption. Paper presented at the International Conference on Wireless Communications, Networking and Mobile Computing, 2007. WiCom 2007. Janda, S., Trocchia, P. J., & Gwinner, K. P. (2002). Consumer perceptions of Internet retail service quality. International Journal of Service Industry Management, 13(5), 412–431. doi:10.1108/09564230210447913 Jarvenpaa, S. L., & Todd, P. A. (1997). Consumer reactions to electornic shopping on the World Wide Web. International Journal of Electronic Commerce, 2(1), 59–88. Kanda, T. (2006). YouTube KAKUMEI. Tokyo: Softbank Creative Corp. Kim, J., Lee, J., Han, K., & Lee, M. (2002). Businesses as buildings: Metrics for the architectural quality of Internet business. Information Systems Research, 13(3), 239–254. doi:10.1287/ isre.13.3.239.79 Kim, S., & Stoel, L. (2004). Dimensional hierarchy of retail Web site quality. Information & Management, 41, 619–633. doi:10.1016/j. im.2003.07.002

A New Web Site Quality Assessment Model for the Web 2.0 Era

LaMonica, M. (2006). Web 2.0 entering corporate world slowly. Retrieved on January 12, 2008, from http://www.news.com/2100-7345_3-6117854. html Lee, Y., & Kozar, K. A. (2006). Investigating the effect of Web site quality on e-business success: An analytic hierarchy process (AHP) approach. Decision Support Systems, 42(3), 1383–1401. doi:10.1016/j.dss.2005.11.005 Lee, Z., & Im, I. (2006). Web 2.0 business strategy. Seoul, Korea: SIGMA INSIGHT. Lin, H.-F. (2007). The impact of Web site quality dimensions on customer satisfaction in the B2C E-commerce context. Total Quality Management & Business Excellence, 18(3), 363–378. doi:10.1080/14783360701231302 Liu, C., & Arnett, K. P. (2000). Exploring the factors associated with Web site success in the context of electronic commerce. Information & Management, 38, 23–33. doi:10.1016/S03787206(00)00049-5 Madu, C. N., & Madu, A. A. (2002). Dimensions of e-quality. International Journal of Quality & Reliability Management, 19(3), 246–258. doi:10.1108/02656710210415668 Mateos, M. B., Mera, A. C., Gonzalez, F. J. M., & Lopez, O. R. G. (2001). A new Web assessment index: Spanish universities analysis. Internet Research: Electronic Networking Applications and Policy, 11(3), 226–234. doi:10.1108/10662240110396469 Moroney, L. (2007). Silverlight: Get started building a deeper experience across the Web. MSDN MAGAZINE. Retrieved on February 5, 2008, from http://msdn2.microsoft.com/en-us/ magazine/cc163404.aspx Moustakis, V., Tsironis, L., & Litos, C. (2006). A model of Web site quality assessment. The Quality Management Journal, 13(2), 22–37.

Murugesan, S. (2007). Understanding Web 2.0. IT Professional, 9(4), 34–41. doi:10.1109/ MITP.2007.78 Murugesan, S., Deshpande, Y., Hansen, S., & Ginige, A. (2001). Web engineering: A new discipline for development of Web-based systems. In S. Murugesan & Y. Deshpande (Eds.), Web engineering: Managing diversity and complexity of Web application development (Vol. 2016/2001, pp. 3-13). Beriln, Heidelberg: Springer. O’Reilly, T. (2005). What is Web 2.0 design patterns and business models for the next generation of software. Retrieved on February 4, 2008, from http://www.oreillynet.com/pub/a/oreilly/tim/ news/2005/09/30/what-is-web-20.html Ogawa, H., & Goto, Y. (2006). Web 2.0 book. Tokyo, Japan: Impress Japan Corporation. Olsina, L., Godoy, D., Lauente, G., & Rossi, G. (1999). Assessing the quality of academic Web sites: A case study. The New Review of Hypermedia and Multimedia, 5(1), 81–104. doi:10.1080/13614569908914709 Palmer, J. W. (2002). Web site usability, design, and performance metrics. Information Systems Research, 13(2), 151–167. doi:10.1287/ isre.13.2.151.88 Parasuraman, A., Zeithaml, V. A., & Malhotra, A. (2005). E-S-QUAL: A multiple-items scale for assessing electronic service quality. Journal of Service Research, 7(3), 213–233. doi:10.1177/1094670504271156 Ranganathan, C., & Ganapathy, S. (2002). Key dimensions of business-to-consumer Web sites. Information & Management, 39(6), 457–465. doi:10.1016/S0378-7206(01)00112-4 Rice, M. (1997). What makes users revisit a Web site? Marketing News, 31(6), 12.

409

A New Web Site Quality Assessment Model for the Web 2.0 Era

Ritchie, P. (2007). The security risks of AJAX/ Web 2.0 applications. Network Security, (3): 4–8. doi:10.1016/S1353-4858(07)70025-9 Rosenfeld, L., & Morville, P. (1998). Information architecture for the World Wide Web. Seastopol, CA: O’Reilly & Associates, Inc. Santos, J. (2003). E-service quality: A model of virtual service quality dimensions. Managing Service Quality, 13(3), 233–246. doi:10.1108/09604520310476490 Szymanski, D. M., & Hise, R. T. (2000). Esatisfaction: An initial examination. Journal of Retailing, 76(3), 309–322. doi:10.1016/S00224359(00)00035-X Terdiman, D. (2005). Wikipedia’s open-source label conundrum. Retrieved on February 20, 2008, from http://www.news.com/Wikipedias-opensource-label-conundrum/2100-1038_3-5988267. html Webb, H. W., & Webb, L. A. (2004). SiteQual: An integrated measure of Web site quality. The Journal of Enterprise Information System, 17(6), 430–440. Wulf, K. D., Schillewaert, N., Muylle, S., & Rangarajan, D. (2006). The role of pleasure in Web site success. Information & Management, 43(4), 434–446. doi:10.1016/j.im.2005.10.005 Xiao, Z., Jones, M. J., & Lymer, A. (2002). Immediate trends in Internet reporting. European Accounting Review, 11(2), 245–275.

410

Zhang, Y., Zhu, H., Greenwood, S., & Huo, Q. (2001). Quality modeling for Web-based information systems. Paper presented at the the 8th IEEE Workshop on Future Trends of Distributed Computing Systems.

Key terMs And definitions Emotional Quality: Quality dimensions concerned with emotions that users may feel while using a website, assurance, empathy, interaction, playfulness, and emotion Information Quality: Quality dimensions related to information a website provides including completeness, timeliness, comprehensibility, trustworthy, presentation variability, architecture, and search capability Interface Quality: Quality dimensions with which a user may feel about interfaces and interactivity, including proximity, compatibility, navigation, appearance, and layout Service Quality: Quality dimensions that promote interactivity among website users and are concerned with responding users’ activities, including customization, support, channel diversity, responsiveness, incentive, and compensation System Quality: Quality dimensions based on technological factors for a website including availability, efficiency, reliability, security Website Quality Dimensions: Specific categories that are required to be considered in assessing website quality Website Quality Model: An integrative set of dimensions that comprehensively cover the entire aspect of website quality

411

Chapter 23

Electronic Reputation Systems Mario Paolucci LABSS-ISTC-CNR, Italy Stefano Picascia LABSS-ISTC-CNR, Italy Samuele Marmo LABSS-ISTC-CNR, Italy

AbstrACt Reputation is a social control artefact developed by human communities to encourage socially desirable behaviour in absence of a central authority. It is widely employed in online contexts to address a number of dilemmas that the interaction among strangers can raise. This chapter presents a social-cognitive theory as a framework to describe the dynamics of reputation formation and spreading. In Section 2, we examine the technology of reputation as implemented in some popular Web platforms, testing theory predictions about the tendency towards either a rule of courtesy or a rule of prudence in evaluation reporting, and thus trying to better understand the outcomes that each system promotes and inhibits.

introdUCtion Internet reputation systems are fascinating technologies. They employ an ancient artefact of mankind to enforce social order - based on traditional social remedies such as word of mouth and chatty talk (Dunbar, 1998)- to regulate a variety of digitally networked environments in absence of central authority.

Reputation appeared online as soon as the Internet became context for social interaction: the more the diffusion of the Net transferred online social problems once limited to the brick and mortar world, the more new flavours of that ancient artefact arose, shaped to fit in online settings to perform a distributed regulation role. First came electronic markets: the earliest widespread setting to feature an ad-hoc designed “reputation” technology was the eBay feedback forum, developed in 1996 by Pierre Omidyar (Li,

DOI: 10.4018/978-1-60566-384-5.ch023

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Electronic Reputation Systems

2006). No central authority exists, in online auction websites, to provide enforcement of contracts stipulated by perfect strangers, possibly located thousands of miles away. Sellers advertise goods the quality of which the buyers can’t verify: information asymmetry should drive the market to an adverse selection condition, turning it into a market for lemons, as Akerlof (1970) described in a classical paper. That has not been the case in eBay, and in our view this is due to the technology of reputation that helps addressing the problem by signalling cheaters and defecting users. The process of decentralization / user-empowerment that took place on the web, which represents the most remarkable Web2.0 effect, generated new contexts of sociality, and is often said to have turned the medium itself into a “public space” (Lovink, 2007) or a “network public sphere” opposed to the traditional mass-mediated public-sphere (Benkler, 2006). Social web applications, collaborative environments, let alone weblogs, projected self-motivated, variouslynetworked individualities over this public space interacting by the means of a digitised flavour of word of mouth. These individuals produce and deliver content, exchange ideas and goods, rate products and engage in a broad range of social activities. Reputation plays a fundamental role in the emerging social architectures built upon the wisdom of crowds principle. In the following, we summarise the socialcognitive theory of reputation developed by Conte and Paolucci (2002) that accounts for dynamics of evaluation circulation in natural societies. We will later employ such theory to investigate online

implementations of reputation, in order to gain insights on the possible biases that online reputation applications can undergo and their effects.

soCiAl Cognitive theory of rePUtAtion: the soCiAl Mind in rePUtAtion dynAMiCs reputation: An evolutionary social Control Artefact Consider a world of agents living in a common and dynamic environment. “Common” means that the environment contains all the agents, which consequently are sharing it. “Dynamic” means that the world is continuously changing. The agents are autonomous (self-interested) and endowed with limited resources (limited knowledge, limited memory, limited life-span, limited foresight). They act on the basis of their own goals, using beliefs about the world. A loose model of the world we live in, plagued by problems deriving from the combination of agents’ and environmental properties. In such a context social norms are an evolutionary artefact that would have emerged to ensure cooperation in early human societies (Ullman-Margalit, 1977). We propose that reputation can be an evolutionary cultural artefact as well. It is one that ensures an easier enforcement of social norms of cooperation, if compared to pure top-down solutions (e.g. court sanctioning), especially when implemented as an integration of them. In fact,

Table 1. Limited and selfish agents in a common and dynamic environment Agents

Environment Common

412

Dynamic

Limited

Interferences, coordination problems

Sudden Disasters

Autonomous

Collective and social dilemmas

Fragility of commitment

Electronic Reputation Systems

reputation in this sense can be seen as an agent property that results from transmission of beliefs about how this agent is evaluated with regard to a socially desirable conduct – be it cooperation or altruism, reciprocity, or law abiding. Although not deliberately designed to achieve social order, reputation-based systems in fact prescribe socially acceptable conducts, like benevolence or altruism, and forbid socially unacceptable ones, like cheating or free riding. As in the case of order emerging from simple local rules, reputation systems are decentralised, based upon distributed social control, because each participant wants other group members to comply with the group norms. Social control is known to be most effective when it is spontaneous and distributed, this is true in natural societies, and more generally in any complex social environment where socially desirable behaviour happens to contrast with individually successful behaviour, and where legal sanctions may be inapplicable or costly.

the Micro Aspect: reputation in the Mind We now try to model the micro level, that is reputation in the mind of the agent: the cognitive representation that, propagating in a group, lets the social property emerge. Mental actions involved in reputation gaining and spreading articulate at the following three levels: •

Epistemic: Accept the beliefs that form either a given image or acknowledge a given reputation. This implies that a believed evaluation gives rise to one’s direct evaluation. Suppose I know that the friend I mostly admire has a positive opinion about someone I despise. Even though puzzled by this dissonance inducing news, I may be convinced to accept this evaluation and share it out of friendship, or not.





Pragmatic-Strategic: Use image in order to decide whether and how to interact with the target. Once I have my own opinion (perhaps resulting from acceptance of others’ evaluations) about a target, I will use it to make decisions about my future actions concerning that target. Memetic: Transmit my (or others’) evaluative beliefs about a given target to others. Whether or not I act in conformity with a propagating evaluation, I may decide to spread the news to others in order to manipulate their beliefs.

image and reputation To better specify these levels of decision, we are naturally brought to introduce another ingredient, that is, a refining on the definition of reputation, distinguishing between a simple believed evaluation and its counterpart in the space of communication. Both concern the evaluation of a given object (the target), i.e. a social agent (which may be either individual or supraindividual), held by another social agent, the evaluator, but we only call the second one reputation. We call the first type of evaluation Image. It is an evaluation of a given target on the part of an agent and consists of a set of evaluative beliefs (Miceli and Castelfranchi, 2000) about the target, regarding her ability or possibility to fulfil one or more of the goals held by some agents, or of the norms present in the system; e.g. to behave responsibly in an economic transaction. Reputation is is an evaluation too, but instead of being concerned with a characteristic of the target, it describes the state of the target assessment in a communication space. Agents diffuse reputation when they report on what they have heard about the target, whether they believe this to be true or not. Reputation, in other words, is the process of circulating a belief about others’ minds, or more specifically, others’ evaluations

413

Electronic Reputation Systems

of the target; to accept it does not imply the acceptance of the nested belief. While transmitting image will imply a level of commitment from the transmitting agent on the evaluation content, agents spreading reputation would neither commit to the information truth value nor feel responsible for its consequences. An epiphenomenon is that, with reputation, agents are more capable to transmit uncertain information, and a given positive or negative reputation may circulate over a population of agents even if its content is not actually believed by the majority. In fact, reputation is a highly dynamic phenomenon in two distinct senses: it is subject to change, especially as an effect of corruption, errors, deception; it emerges as an effect of a multi-level process; and finally it immerges in the agent’s minds as new beliefs. That is, it proceeds from the level of individual cognition to the level of social propagation and from this level back to that of individual cognition again. What is more interesting, once it gets to the population level, it gives rise to a further property at the agent level, immerging in agents’minds as a meta-belief, which may influence their image of the target.

sets of Agents and reputational roles: who does what Now we can put together the different pieces, arriving to the definition of roles – presented as sets of agents - for reputation and image. Any given image implies the existence of three sets of agents: • • •

414

A set E of agents that own the evaluative belief (evaluators), A set T of evaluation targets, A set B of beneiciaries of the social norm implied in the evaluation. These are the agents sharing the goal with regard to which the elements of T are evaluated.

Often, evaluators and beneficiaries coincide, or at least have non empty intersection; since a given agent t needs to be socially evaluated when t is believed to be a good/bad means for a given goal of the set of agents B, beneficiaries are invited to evaluate. Reputation adds a fourth set of agents: •

a set M of memetic agents, or gossipers, who actively contribute to the diffusion of evaluations.

Often, set E can be taken as a subset of M; the evaluators are aware of the effect of evaluation, because they are often the first ones to transmit it. In most situations, the intersection between the two sets is at least nonempty, but exceptions exist. In real matters, agents may play more than one role simultaneously; they may even play all the four roles at the same time.

inaccurate reputation: directions of Memetic Agents’ benevolence, rule of Courtesy and rule of Prudence We may consider the case of inaccurate reputation only if we acknowledge: (a)

The possibility of agent(s) holding a different personal belief (image) and meta-belief (reputation) about the same target, and (b) The possibility that memetic agents transmit information in a not (completely) truthful way (Paolucci, 2000), that is partial or deformed, as a consequence of some of their goals. When circulating the voice, memetic agents may follow different strategies, according to the direction of their benevolence. Considering the agents’ autonomy (their self-interest), these directions are at least three:

Electronic Reputation Systems

1.

2.

3.

In the case of benevolence towards the set of beneficiaries, gossiping may be tendentially bitter and critic. As a result, the aggregate reputation will follow some prudence rule, like “pass on negative evaluation even if uncertain, pass on positive evaluation only if certain”. This may possibly give way to circulation of a reputation which would be cynical, even worse than the real characteristics of the target. On the contrary, when benevolence of transmitters is more target-oriented, it is possible to expect the application of some courtesy rule, like “pass on positive evaluation even if uncertain, negative evaluation only if certain”. This may give way to a courtesy equilibrium, where no-one expresses critique anymore, especially if fearing negative reciprocation, i.e. retaliation. If memetic agents do not have any benevolence towards any of the two cited groups (i.e. sets B and T), the theory predicts scarcity of reputation transmission. In those cases the production of information may be inducted by some institutional reinforcement of it. For example, the scholastic evaluation of scholars on the part of teachers.

We may reasonably suppose that de-responsabilization of memetic agents increments the quantity of circulating information, while benevolence direction is a consequence, all other factors being equal, of the overlapping of reputational roles of agents. For simplicity, we only consider the extreme situations of complete or inexistent overlapping of roles between sets E, T, B and M. In Table 2, terms “overestimation” and “underestimation” refer to the benevolence direction in selective transmission of reputation. Overestimation means the systematic application of a courtesy rule; underestimation refers to the opposite situation, i.e. the application of a prudence rule. “Provision” and “underprovision” are terms referring to the relative quantity of circulating information. Assuming the simplifying condition of equivalence between set E and set M, we get five different role overlapping situations. For a complete description of them we address to Conte and Paolucci (2002). We only consider two significant cases: 1. 2.

Systematic application of a courtesy or prudence rule in reputation spreading may bring along aggregate circulation of only some selected forms of evaluations, either positive or negative. Because this selective transmission is based on specific motivations of the set M (and E), we should understand the social cognitive conditions that determine the application of those rules.

general overlapping of evaluators (E), target (T), beneficiaries (B). overlapping of E ∩ B, while T has a separate status.

On case 1, we expect positive evaluations to prevail, and therefore surpass by far the number of critic evaluations. That would be determined by the overestimation coming from the overlapping of evaluators (and hence memetic agents by hypothesis) and targets, which would entail

Table 2. Expected impact on reputation spreading of different role overlapping Overlapping

M∩E

E∩T

E∩B

B∩T

high

underprovision

underprovision, overestimation

provision, underestimation

underprovision, overestimation

low

provision

provision, underestimation

underprovision

provision, underestimation

415

Electronic Reputation Systems

the application of a rule of courtesy in reputation transmission. We should also consider gossiper’s responsibilization when facing overlapping of beneficiaries and targets; these two effects are bound to neutralize and reverse the one effect of underestimation deriving from overlapping of evaluators and beneficiaries. A practical example of this is the case of gossip within very cohesive groups, like elite military units. Here people are unlikely to, and probably commanded not to, bad mouth each other. This should normally follow the rule: “don’t ask, don’t tell”. Gossip should then be scarce or inexistent, which may benefit discipline and goal reaching of the group as a whole. On case 2, inversely, we expect the emergence of a sort of “social alarm”, useful to warn the community of beneficiaries-evaluators about a possible danger coming from an external target. That is, we expect in this case “prudent” character of reputation spreading, with full expression of critic voices and possible expression of cinical judgement. An example of this case is the gossiping about teachers among students of a same class or level. It is well known that this is an intensely pursued student activity, and that the most part of evaluations is unlikely to be made of appraisals. We expect general adoption of such rules as a consequence of self-interested decisions of single memetic agents. Role overlapping, indeed, implies that the same people (or people having the same norm-related goals) are involved in more reputational roles, and then they may consider useful to be prudent or generous in their information spreading. Or they may have not enough motivation to circulate information of any sort. As we have seen, operationalization of cognitive properties of reputation and dynamics of reputational groups allows to express testable hypothesis, which can be investigated both by experimenting with human subjects and by designing software reputation agents.

416

In the following section we present empirical investigation on online reputation systems about the establishment of a prudence/courtesy rule in case of role overlapping.

wisdom of Crowds, web 2.0 and reputation based internet Applications The term ‘crowd’ never enjoyed a good fame in psychological and sociological literature. Since Gustave Le Bon published the classic The Crowd: A Study of the Popular Mind in 1895 the concept of de-individuation projected a sinister shadow over the interpretation of crowd behaviour. New Yorker’s James Suroweicki reverted this trend in 2004, when he published an elegant analysis of the conditions that lead to virtuous behaviour of individuals acting collectively through certain web platforms. According to Suroweicki these conditions are: 1. 2. 3. 4.

diversity of opinion (each person should have some private information) independence (people’s opinions are not determined by others) decentralization (people are able to draw on local knowledge) aggregation (presence of mechanisms that turn individual judgements in collective decisions)

This latter item refers to the very nature of the ‘social internet’. It is these particular aggregation mechanisms that represent a fundamental revolution from the perspective of history of culture. If we consider the internet as the main repository of the world’s cultural production, the algorithms employed for the storage, dissemination and retrieving of information constitute a sort of meta-memory, providing the means to extract information from the repository and select the contents.

Electronic Reputation Systems

From a social-cognitive perspective we could argue that most of the actual mechanisms that turn individual judgements in collective decisions rely on a common ground that is constituted by: 1.

2. 3.



The principle of rating. Implemented, explicitly or not, in the vast majority of web applications. A set of algorithms which implement a digital flavour of the artefact of reputation. The principle of ranking, which is consequent to the above and presides over the presentation of information.

These are epistemic tools and cognitive practices that emerged to provide a needful shortcut to information in the informationally-overloaded peer production web. Basically, the reputation of an item, that is how others value and rate the item - is the only way we have to extract information about it (Origgi, 2008). This applies - to different extents - to most of the platforms that are said to harness the power of collective intelligence: google, youtube, flickr, digg, del.icio.us, and others. In all these platforms the information is presented to the user in the form of a list of items ranked according to the presumed relevance. The algorithms that produce such an outcome differ one from the other, but all of them rely on these simple principles and, most important, only respond to engineering constraints: they aren’t based on scientific knowledge, nonetheless have a profound cultural impact. In the following1 we provide a basic taxonomy of mechanisms based on a number of sensible characteristics, then we try to illustrate how a difference in the reputational structure of a sysem can produce very different outcomes especially in the ratio of positive/negative evaluations expressed.

taxonomy of Mechanisms An initial taxonomy of user-oriented internet reputation mechanisms could include:







Factual mechanisms: Assign evaluation to a target as a predicate, which is relative to some factual event caused or embodied by the target. Users can have the role of targets and beneiciaries, but they are neither evaluators nor memetic agents. An example of this is the number of posts which is to be found by each author’s name on the discussion forum.These mechanisms do not produce bottom-up reputation, but still are attributed great importance in some systems, such as that of Amazon reviews, where a proliic reviewer has his reviews given particular prominence (David-Pinch, 2005). Rating mechanisms: All the mechanisms that let users express their evaluation in a rating fashion, i.e., picking one value on a scale that in one’s view correctly describes the target. They can be distinguished by rating type, and by type of the aggregation of ratings. Presentation of these types of output may highlight and stress either the positivity or the negativity of ratings, or try to keep a balanced view. Comment mechanisms: These are used as personalization to the rating mechanisms. Especially when rating happens on a single implicit parameter, a textual comment can help to give meaning to the rating. Examples of this are both text comments on eBay’s feedback forum and referrals given among connections on social networks. Connection list mechanisms: Characteristic of social networks (LinkedIn for example). Many connections means counting on many people’s support. The value of the reputation created by this mechanism is very personal: in one’s eyes the support from some people may have different value than in the eyes of another’s. An empty list can either mean that the user is new to the community, or that he had no success within it.

417

Electronic Reputation Systems

Unilateral Versus Bidirectional We argue that the functionality of a mechanism can deeply change depending on the implementing of unilateral versus bidirectional evaluations. In the former case (unilateral) the mechanism allows for evaluators and target not be overlapping. A mechanism to be unilateral must be such both when evaluation spreading is synchronous and asynchronous. Otherwise, it may be the case that an apparently unilateral mechanism is an asynchronous bidirectional one. We advance the hypothesis that unilateral implementation, permitting the separation of roles between agents of set E and agents of set T, may also help to avoid the emerging of a rule of courtesy in the spreading of social evaluation. On the contrary, it would allow the full expression of critics as well as of appraisals. In the second case (bidirectional) we hypothesise that the mechanism would bring along a situation of general overlapping of roles (assuming that evaluators are identified with beneficiaries as in the previous case). We advance the hypothesis that this would entail the application of a rule of courtesy in social evaluation spreading, endangering both the provision level and the informative quality of the evalutations. However, it should not be forgotten that courtesy is perhaps a good thing for the advertising of the system, as it would attract newcomers at a (probable) faster pace than a system where critic voices are found. Some examples: eBay has been using a bidirectional rating mechanism, recently switched to a unilateral one; Amazon has a unilateral one to let users evaluate products, while Amazon Auctions has a bidirectional one as well. Social networks have connection lists, which are bidirectional.

cess to evaluation about himself by others. In the natural setting agents transmitting social evaluation (set M) prefer to have protections for their gossiping, one of which is opacity towards the target. This would lower responsibility and it would favor a higher level of provision than the contrary. Likewise, on the Internet it may be found either opacity for the raters or absence of it. Opacity includes the target not knowing that the evaluation is being carried forth, not knowing who is evaluating or where and when to access that information. We advance the hypothesis that opaque mechanisms would lower responsibility of raters and transmitters and favor the spontaneous provision of evaluation, as well as it would favor the expression of all sorts of ratings, critics included. On the contrary, transparent mechanism would inhibit provision and problematic rating. Opaqueness of a mechanism is determined by the diffusion given to evalutions. A broadcasting mechanism is one that publishes evaluations on the website for all users to see, i.e., a mechanism that does not allow to select recipients of evaluation by the rater. A narrowcasting mechanism, on the other hand, permits transmission similar to that of natural gossip, giving users the possibility to decide who should be the recipient of the rating. This mechanism allows for the chain-like diffusion of social information that is typical of natural settings.

Dynamism The dynamism of a reputation mechanism is given by two parameters: •

Counting single pairwise evaluations versus counting multiple ones; Discounting evaluations by their age.

Broadcasting Versus Narrowcasting



Rating mechanisms can implement either partial or total evaluation opacity towards the target. Therefore, the latter may have or not have ac-

Single evaluation counting mechanisms report only the very last rating that each user gave upon another, discarding the older one. Multiple evalu-

418

Electronic Reputation Systems

ation counting mechanisms aggregate all ratings between pairs of users. If raters accumulate a rising level of evaluations from the same people, this would create very static “reputational” profiles, that would no more be useful for sanctioning cheating behavior, as any new evaluation would have very little impact on the total. Instead, accounting only the last evaluation from each specific user gives (a) more dynamism to the evaluation count and (b) it avoids evaluation inflating by collusion between pairs of agents (ballot stuffing). A mechanism can also be designed in order to “forget” older ratings in favour of new ones.

An empirical validation of the social Cognitive theory on internet reputation systems To show how our theoretical deductions could explain the differences seen in actual online reputation systems, we report and integrate some results from an empirical study (Paolucci et al., 2006). The theory suggests that different situations of overlapping could give different tendencies, included prudent reputation spreading. We tested online rating mechanisms for evidence of this. The task was made easier by the fact that,

unlike most natural reputational settings, online rating mechanisms must exactly define the roles of users. The implementation has to specify who is allowed to rate, on whom, when, who benefits of that, is evaluation public or private, signed or anonymous, are there gossipers that are not personally evaluating, and so on. This allowed to gather data about (a) the implementation of reputational roles and (b) the percentage of critic evaluations available to the users of the site. The goal was to find implementations of reputational roles which would give an output including critic voices, as it is not happening on eBay-like systems. This would prove that implementation of reputation mechanism can be fine-tuned in order to obtain high provision (dishinibition of transmitters) and informative value (completeness of voices, including critic ones) or to obtain a courtesy situation. In order to distinguish a situation where reputation effectively follows a rule of prudence from one where a courtesy rule applies we measured independently the proportion of positive and negative evaluations given by users on systems2. Our hypothesis is as follows: on systems where evaluator and target roles are implemented as overlapping, there will be a lower proportion of negative ratings than that in the opposite case; on systems where evaluator and target roles are

Table 3. Expected effects of different mechanism functional implementation Expected effects of mechanisms’ functionality Evaluation Tendency

Dynamism

Level of Provision

Unilateral rating

accuracy, prudence

-

-

Bidirectional rating

leniency, inhibition

-

-

Single user’s rating

-

dynamic

-

Multiple user’s rating

-

inertial

-

Time discount

-

dynamic

-

No time discount

-

inertial

-

Broadcasting

-

-

low

Narrowcasting

-

-

high

419

Electronic Reputation Systems

implemented as separated, there will be a higher level of negative ratings than that of the opposite situation, possibly bringing forth completeness of social information and accuracy. To identify the overlapping of roles we just observed if the implemented mechanisms were of the unilateral or bidirectional type.

Survey Method First we conducted a broad survey of existent sites hosting user-oriented rating systems. This brought to a list of about 70 sites from which 9 samples were picked up for in-depth examination. In this work we account for 5 of them: 3 electronic auctions (eBay, Guru, Amazon Auctions) and 2 social news websites (Digg, Slashdot). Afterwards we subscribed to those sites and conducted normal activity on them. Once direct experience with each one was gathered, a detailed description of each was compiled, including: • • • • • • • • • •

context of use for reputation; mechanisms implemented; type of social evaluation produced; opacity of evaluations to targets; possibility of information noise; eventual bias of the system’s information handling; anonymity for evaluation givers; reputational value of new proiles; reputational roles’ overlapping; proportion of extracted non positive evaluations.

Data regarding Digg were gathered through third-party websites that provide real-time statistics. In particular, the percentage of buries (negative evaluations) in comments was extracted searching 100 random profiles through http:// www.neaveru.com/digg/. Bury percentage was extracted searching 100 random stories in a four weeks time span through http://www.ajaxonomy.com/buryrecorder/

420

We considered negative and “neutral” evaluations to be similar because they are both the expression of some problematic feeling about the target, while they were considered as positive evaluations all those above the neutral one. The same applies to the lack of feedback provision in social news sites, especially in Digg: ignoring a story equals to preventing it from hitting the frontpage, which of course implies a negative evaluation of the story. Unfortunately statistics about feedback underprovision are not available, but the system is designed in order to encourage users to “digg” stories they like more than bury those they don’t (while comments are moderated equally trough diggs and buries). Under the general assumption that reputation should be inaccurate, we set an arbitrary threshold of 10% for the proportion of non positive evaluations. We searched for correlation between systems having both a lower proportion than 10% of problematic evaluations, overlapping of roles between targets and evaluators. While other secondary factors may be influencing that proportion, we assumed that it may be considered probable, if not certain, that situations with under 10% of problematic rating be the product of some sort of courtesy rule in information spreading. In other words, we postulated that the overlapping of roles be the primary factor determining the application of either courtesy or prudence rule in evaluation spreading.

data details For each site, the method of extracting users’ evaluations followed one or more of the following three methods, in order of preference: (i) alphabetical search of users’ profiles, (ii) temporal method, as users logged in to the system and were signaled to those already there, (iii) random extraction by browsing. The latter was only used as a last resort, and expresses only a minor share of the whole extracted data.

Electronic Reputation Systems

eBay (www.ebay.com) Context: (electronic auction) signaling quality of, and sanctioning moral hazard of “eBayers” Mechanisms: numerical rating with output consisting of the algebraic sum of ratings on scale [-1, 0, +1] and percent proportion of positives, on just one implicit satisfaction parameter, plus textual comments; both ratings and comments used to be bidirectional, became unidirectional as of May 2008 (see discussion for comments); broadcasting; rating is allowed only after assignment of an auction. Ratings are discounted by time: highlight is given to those referring to the last 12 months. Information type: only direct evaluation (shared image). Opacity of evaluation to targets: no opacity. Possibility of noise: ratings can only be withdrawn on a time-costly mutual agreement procedure, so that most ratings are permanent once given, and comments are never withdrawn. This introduces noise relative to the possibility of human error in rating. Bias in system’s information handling: though mathematical treatment of feedback is to be considered correct, great evidence is given to positive evaluations. This may have the effect of hiding problematic feedback or lower its visibility. There is no search function within comments. Description of items to which comments refer are shown only for a 90 days period after the comment date. “Neutral” feedback is not reported when older than 12 months in the numerical synthesis, while comments remain always active, even though almost unreachable when there is a high number of them. Anonymity: absent. Reputational value of new profiles: average, because it can be worsened by initial negative feedbacks. Reputational role overlapping: set M = set E = set T = set B; set E ≠ set T as of May 20, 2008.

Proportion of extracted non positive evaluations: about 1% of 7379 evaluations examined from 100 profiles with alphabetic method on May 2005.

Guru (www.guru.com) Context: (electronic auction) signaling competence and reliability of both professionals and employers in an auction for job offerings. These two roles are required separate profiles on the site. Mechanisms: numerical rating on a positive scale of 10 values with output consisting of arithmetic average, plus textual comment. Both ratings and comments are bidirectional, with multiple users-evaluation count, broadcasting, and are relative to multiple explicit parameters of service quality; rating is allowed only after assignment of an auction. No time discount on ratings. Information type: only direct evaluation (shared image). Opacity of evaluation to targets: absent. Possibility of noise: ratings can not be edited after release, thus noise due to rating errors is possibly present. Bias in system’s information handling: there a slight tendency to give more weight to non positive ratings, as the 10 values result from 5 levels (stars) with middle points between them, referring to 2 praising levels, one “neutral” middle level and 2 non praising levels. However, professionals are given 5 profiles or more to be used as they wish, exposing or withdrawing them depending on opportunity. This may bring to negative rating effectively disappear from the site. Anonymity: absent; rater is reported with every rating; however, they are only identified through serial numbers given by the site, so they remain unnamed, but can be searchable on the site. Reputational value of new profiles: average, as a low initial rating can lower its value. Reputational role overlapping: set M = set E

421

Electronic Reputation Systems

= set T = set B; in the site’s intent, there should be two reputational settings: users with an employers’ profile should rate on professionals to the benefit of other employers, and the same is true for professionals; however, there is no effective separation of this, as evaluations are given in a bidirectional way. This makes roles mix and overlap. Proportion of extracted non positive evaluations: 0.2% of a few thousands evaluations found on 2175 profiles (of which about 70% with no ratings) with random method on October 2005. User evaluations are very scarce and only available for the top listed profiles, and almost always have maximal values.

(one grey neutral, two red negatives). Anonymity: absent. Reputational value of new profiles: average, as an initial negative can lower an entrant’s profile value. Reputational role overlapping: set M = set E = set T = set B; conferring rating numbers just to the sellers is insufficient to realize separation between set E and set T, as the comments remain bidirectional. Proportion of extracted non positive evaluations: 2% out of 21539 evaluations from 50 profiles, with alphabetic method on October 2005.

Digg (www.digg.com) Amazon Auctions (http:// auctions.amazon.com) Context: (electronic auction) signaling quality of, and sanctioning moral hazard of buyers and sellers, i.e. all users of the electronic auction for merchandise items. Mechanisms: numeric rating on 5 positive levels, with output consisting of the arythmetic average and the total number of ratings; percent proportions of positive, “neutral” and negative rating are given to an additional click from user’s profile page; there is just one implicit satisfaction parameter; textual comments are added; numeric rating is unilateral, while comments are bidirectional. Rating is allowed only after assignment of an auction. Ratings are discounted by time: highlight is given to those referring to the last 12 months. Information type: only direct evaluation (shared image). Opacity of evaluation to targets: absent. Possibility of noise: ratings can not be edited after release,thus noise due to rating errors is possibly present. Bias in system’s information handling: there a slight tendence to give more weight to non positive ratings, as there are two positive rating values (marked in green), and three non positive ones

422

Context: social news website: users post links to news items discovered on the internet, along with a brief introduction. Fellow users can comment each post and “digg” it if they consider it interesting, relevant, or generally “worth reading”, or “bury” it if the post is not useful, spam, or inconsistent. The same applies to comments. Mechanism: Posts with a certain amount of “diggs” (binary positive evaluations) collected over a certain amount of time are featured in the homepage. Comments are evaluated in the same manner, those below a threshold defined by the user are not shown. Information type: only direct evaluation (shared image) Opacity of evaluation to targets: absent Possibility of noise: diggs can be always reverted, comments can be edited for 4 minutes after posting. Bias in system’s information handling: Non positive ratings count much more than positives when dealing with news postings: under certain circumstances, even a 5% of negatives can “bury” a story. In comment moderation no significant bias exists. Anonymity: only for negative ratings in stories, present in comments. Reputational value of new profiles: aver-

Electronic Reputation Systems

age. Reputational role overlapping: set E ≠ set T; set B includes both set E and set T; set E includes set M. Proportion of extracted non positive evaluations: 8% for stories, 40% for comments Notes: The amount of negative-rated stories is only indicative. (See “Survey method”)

Slashdot (http://slashdot.org) Context: a subset of readers of this online magazine is given the possibility to rate the posts on discussions where they can not intervene, as they would lose the rating privilege; the latter is assigned upon evaluation from the system of user’s profile (counting age of registration, level of activity, ratings received on posts from reviewers); the reading of the posts can be set in order to see only those with a certain “moderation level”; this is used both to lower the visibility of spam and unappropriate content as well as “flames” (personally offending comments); and to make quality content more visible. Rating is allowed at a specific moment (see additional notes). Mechanisms: unilateral multiple count broadcasting numerical rating mechanism (for stories); output is given as the algebric sum of all ratings on the post. Each rating has a comment of a single word attached, stating the motivation of rating (either “informative”, “insightful”, “funny”, “informative”; or “off topic”, “offensive”, “redundant”, etc.). No time discount. Unilateral, binary, transparent (anonymous in metamoderation only) Information type: only direct evaluation (shared image). Opacity of evaluation to targets: opaque. Possibility of noise: present; assigning moderating points to comments is not reversible, if not for “meta-moderating”, which a function introduced as a control of moderators, but not so wide-reaching as the former. Bias in system’s information handling:

neutral handling. Anonymity: absent. Even though difficult to reach, there is a page on the site where registered users can see who moderated a comment and how. Reputational value of new profiles: medium, as each post can be either moderated up or down; however, the post cannot be retreated upon receiving negative moderation points. Reputational role overlapping: set E ≠ set T; set B includes both set E and set T; set E includes set M. Proportion of extracted non positive evaluations: 34.15% of 101 evaluations from 50 moderators with the temporal method.

results As pointed out by the listed data, the main hypothesis is to be considered confirmed by our survey: implementation of reputational roles’ is a fundamental predictor of high or low level of critic evaluations’ spreading. Results of this are showed in Table 4 and 5, representing respectively situations with high set E∩T overlapping and situations where set E ≠ T.

discussion Electronic Auctions On the examined electronic auctions “reputation” mechanisms seem to convey mostly positive evaluations. Rather than the user’s future performance, this indicates the user’s experience in the past (cf. Dellarocas, Fan and Wood, 2004). We postulate that a strong courtesy environment is enabled by the overlapping of evaluators and targets, which is in turn caused by the bidirectionality of the implementation. Moreover, ratings are given in a non-anonymous broadcasting way, possibly adding further inhibition to problematic voices. Rating is only allowed when an auction is assigned, thus influencing scarcity of social information. New

423

Electronic Reputation Systems

Figure 1. Percentage of negative evaluations in the examined platforms

profiles have not a minimal value, but an average one, that provides incentive towards abandoning a deteriorated profile for a new one. We therefore argue that, while eBay-like feedback mechanisms are driving strong reciprocation at the rating level, it does not protect effectively from malicious sellers that build up a high “reputation” to start then cash it away with low quality service, or straightforward fraud. This situation is, anyway, already changing: starting from May 2008, in the eBay system the sellers are no longer able to give buyers negative feedback, basically making them no more evaluators. The mechanism suddenly became unilateral, so set E was to be separated from Set T. The shift will be an occasion to test the theory prediction that separation of targets and evaluators leads to the establishment of a prudence rule. More accurate evaluation by buyers with less fear of retaliation should increase, in the near future, the share of negative evaluations. This should make the system

424

more reliable, and probably that will be the case, even though there is now a higher risk of “bad mouthing” (sellers collude with buyers in order to bad mouth other sellers; see Dellarocas, 2003). As of May 2008 eBay’s reputation infrastructure also underwent other changes: Positive repeat customer Feedback now counts (up to 1 Feedback from the same buyer per week). Repeated ratings didn’t count at all before. This, of course, elevates the chances of collusion between evaluators and targets in order to inlate a seller’s rating: ballot stufing, which reinforces the effects of bad mouthing + positive discrimination (sellers provide exceptionally good service to a few selected individuals and average service to the rest) Feedback more than 12-months old doesn’t count anymore towards Feedback percentage. Feedback will last less, making the

Electronic Reputation Systems

Table 4. Summing up results about electronic auctions Electronic Auctions Mechanisms

Information type Opacity towards set T

Ebay

Guru

Amazon Auctions

Numerical rating, comment (bidirectional, single count, time discount, broadcasting)

Numerical rating, comment bidirectional, multiple count, no time discout, broadcasting

Comment (bidirectional) Numerical Rating (unilateral) single count, time discount, broadcasting

shared image

shared image

shared image

absent

absent

absent

Anonymity

absent

partial

absent

Choice of recipient

absent

absent

absent

Type of system

centralized image

centralized image

centralized image

Noise

possible

possible

possible

Information handling

positive preference

Negative preference

negative preference

Provision estimate

High (50-60%)

Low

10-20% by site’s source, low

Overlapping of roles

E=T=B E includes M

E=T=B E includes M

E=T=B E includes M

Resulting Effect

courtesy

Scarcity, courtesy

scarcity, courtesy

Non positives

1,00%

0,20%

2,00%

New id value

average

average

average

system more dynamic (symmetrically), thus discounting the past and darkening the shadow of the future over ebayers’ heads. According to the theory this should make good reputation more effective, bad reputation less. It will be more expensive to keep a good reputation once gained, whereas an occasional old fraud will be soon forgotten.

social Content websites Here, the sample analysed included two sites with similar characteristics. Slashdot has a very elegant unilateral rating mechanism, as evaluators are strictly kept separate from the targets, and non positive evaluations are about 34%. An extra level of rating (meta-moderation) exists: random “senior” users are asked to evaluate the fairness of comment evaluation performed by selected moderators. If the meta-moderation is negative a

user can be prevented from evaluating again. The system seems to work quite well, as reading its posts with a high threshold for their moderating points can effectively rule out uninformative and low quality posts. New usernames have average value, but only long-time users are allowed to moderate (ie rate stories and comments), which discourages abandoning a deteriorated username for a new one. The architecture of Digg is not so elaborate, Evaluators and Targets are not separate with the same accuracy as in Slashdot, but the outcome is very similar. The system amends the substantial bias towards underprovision of explicit negative feedback in stories by ultra vires hiding posts that get a certain amount of negatives in a short time. Thus, negative evaluation counts much more than positive: sometimes as little as 5% negatives are sufficient to “bury” a story, if they are expressed by users with different profiles in little time span.

425

Electronic Reputation Systems

ConClUsion Our brief analysis of websites implementing user-oriented reputational mechanisms suggests a number of conclusions that can be drawn: The social cognitive theory of reputation, developed to describe dynamics from natural societies, explains part the of the functioning of internet reputation systems. However the ‘social web’, by transporting online the social artefacts upon which it builds, substantially modifies their characteristics3. Electronic reputation systems only implement a subset of the features of social reputation systems: we rarely have reputation as voices following one another often without explicit reference to their origin, as it is in society. Often, mechanisms are like “black boards” where everything is published for all to see. Only direct evaluations are permitted, and voice reporting has no primary place or no place at all.

In this simplified form, online reputation has been considered as a solution for the sanctioning of moral hazard as well as the signalling of quality in many different areas: electronic auctions, recommender systems, social networks, discussion fora, search engines, peer-to-peer networks. Boundaries between these contexts have been fading with the advent of Web2.0: social networking capabilities are rather ubiquitous in all modern web platforms. We should expect a similar development in the employment of reputation systems: they are bound to become the cornerstone of the social web, filtering content and signalling quality. The design of internet reputation systems can be fine-tuned towards different goals such as courtesy, provision, completeness of information, prudence. The possibility of assigning roles, setting anonymity levels, providing or not opacity towards targets gives the designer a wide range of choices

Table 5. Summing up results about social news filters Social news filters

Mechanisms

Information type

Digg.com

Slashdot

Numeric Rating unilateral, binary broadcasting

Numeric Rating unilateral, multiple count, broadcasting

shared image

shared image

Opacity towards set T

absent

absent

Anonymity

present

present

Choice of recipient

absent

absent

Type of system

centralized image

centralized image

Noise

very limited

possible

Information handling

neutral

neutral

Provision estimate

high

high

Overlapping of roles

E≠T

Resulting effects

prudence

prudence

Non positives

8,40% for stories (but see “Survey method”), 38% for comments

34,15%

New id value

average

average

426

E≠T B includes E and T E >= M

Electronic Reputation Systems

for the implementation of a system that produces the desired outcome in term of user’s behaviour. The theory provides operationalisation of predictions that can be used towards that goal. More testing of online users’ epistemic, pragmatic and memetic decisions is indispensable for providing precise data about the importance given to each information mechanism that should be implemented. However, our survey shows that: •



implementation of overlapping evaluators and target brings to very low expression of problematic voices (well under 10% of the total), with a possible effect of rule of courtesy; implementation of separated evaluators and targets (while evaluators are part of the beneiciaries) avoids courtesy, provides prudence in evaluation spreading and possibly the complete expression of evaluations.

More research is needed in order to state precisely which parameters other than role overlapping are determining the free expression of evaluations or inhibition. However, our survey pointed out a clear tendency to inhibition on systems implementing overlapping of set E and T, as non positive evaluations remain under 2% on different electronic auction sites with the same implementation. On systems that do not present such overlapping, even though there is no opacity nor anonymity, evaluations seem to show a more complete face. Courtesy or prudence are legitimate online systems’ goals. Indeed, courtesy and prudence can be seen as having both advantages and disadvantages for the online domain. Prudence is a prerequisite for obtaining complete and possible accurate reputation, and hence it is the best means to overcome information asymmetry in the markets. However, the full expression of critics on

electronic markets, especially on newly started sites can be hindering if compared to courtesy evaluations. Indeed, the latter can be a powerful attractor of newcomers, because it tends to present things as very bright, possibly brighter than the reality. This is in our view a suitable explanation of the success of the reputation system of sites like eBay. However, when such sites are grown to have a big number of users, attention should be devoted to keep those users well informed, especially giving them a rating mechanism that provides reliable information on the potential partners. Studies at the macro level about eBay show that a higher percentage of sellers leave the market than buyers, while a higher rate of new buyers entered the market than sellers (Lin, Li, Janamanchi and Huang, 2006). A different case is that of reputation systems in non-economic contexts. Social content websites like Digg.com, Reddit.com or the Google PageRank itself build on a strong reputational factor, but here the evaluations are expressed with no reference to an explicit set of norms. In other words, what is evaluated, in Digg, is not clear, and the same applies to Google’s PageRank, which is unable to discern between different semantics of a web link4. Digg, Reddit and likewise sites operate on a semantic layer above PageRank, by allowing users to consciously vote for pages, and even to express negative ratings, but still the nature of what is evaluated remains unclear. Further research should investigate the underpinnings of these systems which will increase their influence on how we extract knowledge from the common repository that is the Web. Theory predictions linking the reputational role structure and the proportion of positive/negative evaluations also apply to these websites, as shown. However many more biases that certainly exist need to be addressed, if we wish to understand how the web is affecting our relationship with culture.

427

Electronic Reputation Systems

referenCes Akerlof, G. (1970). The market for “lemons”: Quality uncertainty and the market mechanism. The Quarterly Journal of Economics, 84, 488–500. doi:10.2307/1879431 Conte, R., & Castelfranchi, C. (1995). Cognitive and social action. London: London University College of London Press. Conte, R., & Paolucci, M. (2002). Reputation in artificial societies. Social beliefs for social order. Boston: Kluwer. David, S., & Pinch, J. (2006). Six degrees of reputation: The use and abuse of online review and recommendation systems. First Monday–Peer reviewed Journal of the Internet. Retrieved from http://www.firstmonday.org/issues/issue11_3/ david/index.html Dellarocas, C. (n.d.). Efficiency through feedbackcontingent fees and rewards in auction marketplaces with adverse selection and moral hazard. ACM Conference on Electronic Commerce2003 (pp. 11-18). Dellarocas, C., Fan, M., & Wood, C. (2004). Selfinterest, reciprocity, and participation in online reputation systems. Working paper no. 4500-04. MIT Sloan School of Management Dunbar, R. (1998). Grooming, gossip, and the evolution of language. Cambridge, MA: Harvard University Press. Le Bon, G. (1895). The crowd: A study of the popular mind. Li, L. (2006, October 27). Reputation, trust, & rebates: How online markets can improve their feedback mechanisms. Paper 55. Institute for Mathematical Behavioral Sciences. Retrieved from http://repositories.cdlib.org/imbs/55

428

Lin, Z., Li, D., Janamanchi, B., & Huang, W. (2006). Reputation distribution and consumerto-consumer online auction market structure: an exploratory study. Decision Support Systems, 41(2), 435–448. doi:10.1016/j.dss.2004.07.006 Lovink, G. (2007). Zero comments: Blogging and critical Internet culture. New York: Routledge. Miceli, M., & Castelfranchi, C. (2000). The role of evaluation in cognition and social interaction. In K. Dautenhahn (Ed.), Human cognition and social agent technology. Amsterdam: Benjamins. Origgi, G. (2008). Designing wisdom through the web. The passion of ranking. Presented at the Workshop on Collective wisdom, Collège de France, Paris 22-23 May 2008. Paolucci, M. (2000). False reputation in social control. Advances in Complex Systems, 3(1-4), 39–51. doi:10.1142/S0219525900000042 Paolucci, M. Balke, T., Conte, R., Eymann, T. and Marmo, S., Review of Internet User-Oriented Reputation Applications and Application Layer Networks (September 16, 2006). Available at SSRN: http://ssrn.com/abstract=1475424 Surowiecki, J. (2004). The wisdom of crowds. Doubleday Ullman-Margalit, E. (1977). The emergence of norms. Oxford, UK: Oxford University Press.

endnotes 1

2

3

This section extends previous work reported on Paolucci et al. (2006) Being positive ratings much more present than negatives, we focused on the proportion of the latter ones Also think about the way the concept of “friendship” has been redefined and reified by social networking applications

Electronic Reputation Systems

4

For example, when linking to a fraudulent web page as a warning to other users, PageRank will count the link as a positive vote. The practice known as Google bombing relies on this particular characteristic of the search engine.

429

430

Chapter 24

Improving the Information Security of Collaborative Web Portals via Fine-Grained Role-Based Access Control S. Demurjian University of Connecticut, USA H. Ren University of Connecticut, USA S. Berhe University of Connecticut, USA M. Devineni Serebrum Cooperation, USA Sushil Vegad Serebrum Cooperation, USA K. Polineni Serebrum Cooperation, USA

AbstrACt Collaborative portals are emerging as a viable technology to allow groups of individuals to easily author, create, update, and share content via easy-to-use Web-based interfaces, for example, MediaWiki, Microsoft’s Sharepoint, and so forth. From a security perspective, these products are often limited and coarse grained in their authorization and authentication. For example, in a Wiki, the security model is often at two ends of the spectrum: anonymous users with no authorization and limited access via readonly browsing vs. registered users with full-range of access and limited oversight in content creation and modiication. However, in practice, such full and unfettered access may not be appropriate for all users and for all applications, particularly as the collaborative technology moves into commercial usage (where copyright and intellectual property are vital) or sensitive domains such as healthcare (which DOI: 10.4018/978-1-60566-384-5.ch024

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Improving the Information Security of Collaborative Web Portals

have stringent HIPAA requirements). In this chapter, we report on our research and development effort of a role-based access control for collaborative Web portals that encompasses and realizes security at the application level, the document level (authoring and viewing), and the look-and-feel of the portal itself.

introdUCtion Over the past decade, the World Wide Web (WWW) has come to the forefront as a viable means to allow individuals and organizations to collaborate. Consequently, web portals have emerged as a means to facilitate these interactions, ranging from information repositories to full-fledged authoring and document content collaboration. For instance, WebMD (www.webmd. com) and Wikipedia (www.wikipedia.org) are utilized by unregistered users to browse content via easy-to-use web-based interfaces. For registered users, these web portals provide a means to author, create, modify, and track documents of all types within a consistent framework or infrastructure. A registered user of Wikipedia has the ability to create new document content and modify existing content. Open source products such as Mediawiki (http://www.mediawiki.org) or a commercial solution such as Microsoft’s Sharepoint (http:// www.microsoft.com/sharepoint/default.mspx) allows any individual with sufficient expertise to generate their own web portal to meet specific purposes and needs. However, from a security perspective, these products are often very limited in the level of protection that is offered to information content that is created and uploaded using these various portals. For example, a registered Wikipedia user could create and upload intentionally erroneous content (e.g., a document that says that the world is flat). Some of these web sites depend on the community of users themselves to monitor document content; as the volume of content at these sites grows, it becomes problematic to attempt to maintain information in this fashion. Due to the

lack of security and control, many corporate and governmental users are hesitant to utilize such technologies for content creation and collaboration, restricting their usage to an information repository; these same users have serious confidentiality, copyright, and intellectual property concerns as well. For example, an emerging usage of collaborative portals is in patient and physician collaboration on day-to-day health care (https:// www.relayhealth.com/rh/default.aspx) where confidentiality is governed in the United States by the Health Insurance Portability and Accountability Act (HIPAA, http://www.hhs.gov/ocr/ hipaa/). Utilizing existing collaborative portals in health care are likely to violate HIPPA, given the coarse level of access and limited accountability to content creation and modification; the security of patient/physician interactions simply could not be assured. For commercial viability, collaborative portals must have more rigorous security capabilities than the coarse-grained authorization and authentication (user names and passwords) that are typically offered by Wikis/web portals. As a web application, a collaborative portal must prevent inherent vulnerabilities. As characterized by the Open Web Application Security Project (OWASP), the top ten web application vulnerabilities have been identified to assist developers, designers, and organizations in protecting their web applications from intrusion and malicious access (http:// www.owasp.org/index.php/Top_10_2007). These vulnerabilities include: SQL injection flaws where SQL code typed into say a name or address data field alters a command to the database possibly resulting in the release of information; insecure communications as reflected by the lack of us-

431

Improving the Information Security of Collaborative Web Portals

age of https (secure http) for all interactions that have sensitive data; inadequate cryptography algorithms and credentials that do not adequately protect data stored in a database; and so on. In addition to this targeted discussion, there is a comprehensive primer of WWW security issues that classifies security concerns according to: client side via a web browser, server side in the web and application servers themselves, CGI scripts and their potential security holes, protection of confidential documents from access and misuse, and denial of service attacks (http://www.w3.org/ Security/Faq/). Collaborative portals, by their nature, are intended to promote a high degree of interaction, and provide administrative users with a high degree of access (via privileges); as a result, all of these aforementioned security vulnerabilities and WWW security issues are paramount for their commercial viability. In this chapter, we report on our research and development effort of applying role-based access control (RBAC) (Sandhu, 1996) to web portals as part of a funded research effort (NSF, 2006) for collaborative software requirements elicitation (Pia, Demurjian, Vegad, Kopparti & Polineni, 2007). In this effort, the Axon Wiki (http://www.serebrum.com/axon/index.html) has been prototyped with RBAC security at the application level, the document level (authoring and viewing), and the look-and-feel of the portal itself. Axon is a Java-based, Ajax (http://developers.sun.com/ajax/) Wiki that offers document authoring, collaboration, publishing, versioning, and other capabilities. The intent is to provide a full-capability Wiki that has fine-grained RBAC in terms of security requirements, flexibility and administration, with more security capabilities than available open source and commercial products; a report on the adaptation and evaluation capabilities of Axon is also available (Berhe, Demurjian, Ren, Devineni, Vegad & Polineni, 2008). Please note that the work presented in this chapter is being practically applied by another group (overlap of authors) in a real-world setting to allow faculty

432

researchers and health care providers to collaborate with one another to make decisions on health information technology (Crowell, Agresta, Cook, Fifield, Demurjian, Carter, Becerra-Ortiz, Tracey, Vegad & Polineni, 2009). In particular, that work illustrates the practical need of a collaborative portal with fine-grained access control capabilities in the health care domain, enabling different individuals with specific roles to work together on a particular medical topic of interest. Moreover, Axon allows each member to securely distribute medical information (education material, treatment feedbacks, research material, etc.), and constrain access (read and/or write) on a role-by-role basis to authorized users. The remainder of this chapter has five sections. First, to provide an important context to the work presented in this chapter, we review alternative security techniques, highlighting the security limitations faced by collaborative web portals/platforms. Second, we provide background concepts on Axon’s capabilities and architecture. Third, we detail Axon’s security, focusing on: assumptions and permissions (specific to Axon but applicable to collaborative portals); permissions that support RBAC at the application level, the document level (authoring and viewing), and the look-and-feel of the Wiki itself; the relational database tables that are utilized to realize the RBAC in Axon; and, limitations of Axon. Fourth, we review related research efforts and highlight future trends. Finally, concluding remarks are presented.

seCUrity And its liMitAtions This section has three objectives. First, we briefly introduce and review the three classic access control models: mandatory access control, MAC (Bell & La Padula, 1975), discretionary access control, DAC (Linn & Nystrom, 1999), and rolebased access control, RBAC (Sandhu, 1996). As part of this discussion, we highlight the relevance of each access control model for collaborative

Improving the Information Security of Collaborative Web Portals

portals. Second, we discuss information security limitations and their risks of breaches and misuse, including: unintentional and overt, focusing on client and server side issues that must be handled for any web-based application. Third, we review a select set of open source and commercial collaborative portals, focusing on information security support/limitations. In MAC (Bell & La Padula, 1975), security levels (SL’s) such as unclassified (U), confidential (C), secret (S), and top secret (T) where U < C < S < T form a lattice structure and are assigned to each subject (clearance - CLR) and object (classification - CLS). The permission of the subject to perform some operation on the object depends on the CLR and CLS relation as dictated by: Simple Security Property (read down - no read up) (Bell & La Padula, 1975); Simple Integrity Property (write down - no write up) (Biba, 1977); and, Liberal *-Property (write up - no write down’) (Bell & La Padula, 1975). Even careful usage of MAC can lead to problems. For example, a user with Simple Security for reads and Liberal * for writes cannot see sensitive information, but can write information to more secure levels; other users reading that information must be skeptical concerning its content. The usage of MAC for governmental secure computing is required in many of their applications; its adoption in collaborative web portals is problematic. The main difficulty is granularity; in MAC, the sensitivity of relational data can be at the row, column, or even attribute-value levels. For collaborative portals, the granularity is often a document; supporting security levels to specific portions of the document is extremely difficult. In RBAC (Sandhu, Coyne, Feinstein, & Youman, 1996), the concept of a role is utilized to represent common user responsibilities with respect to the application, e.g., a university would have roles for faculty, administrators, students, etc. Roles are authorized to perform some operations on certain objects, and are assigned to users to specify the named assignments that those users

can perform in the organization. RBAC is wellsuited to support least privilege, which allows for access to only that information which is necessary to accomplish one’s tasks (Ferraiolo, 2001). The NIST RBAC Standard is ideal for collaborative web portals (http://csrc.nist.gov/groups/SNS/rbac/ documents/towards-std.pdf), since the granularity of the security permission can be fine and custom. As noted in the introduction, in a portal such as MediaWiki, unregistered users access the portal via their web browser in a read-only mode, while registered users have broad access to create, modify, and delete content. The collaborative community itself (i.e., registered users) monitor the site for incorrect, inaccurate, or unacceptable (graphic) content; however, as these repositories grow to billions of pages and more, the ability to monitor all of the content becomes an impossible task. Thus, for true acceptance in commercial settings, there needs to be finer-grained control from a security perspective, that not only grants access by role to users, but limits the content (allowable pages or portions of pages) itself. In DAC, the emphasis is on the delegation of authority, where an authorized individual (not the security officer) may delegate all or part of his/her authority to another individual, increasing security risk, and raising interesting security assurance implications (Linn & Nystrom, 1999). Large organizations often require delegation to meet demands on individuals in specific roles for certain periods of time. In DAC, role delegation is a user-to-user relationship that allows one user to transfer responsibility for a particular role to another authorized individual, and can be classified as: administratively-directed delegation, where an administrative infrastructure outside the direct control of a user mediates delegation (Linn & Nystrom, 1999); and, user-directed delegation where an individual (playing a role) determines if and when to delegate responsibilities to another individual to perform the role’s permissions (Na 2000). If collaborative web portals are to have wide-spread commercial acceptance, delegation

433

Improving the Information Security of Collaborative Web Portals

of authority must be supported. When a user delegates their role, they delegate the authority to do the task, which leads to a set of critical identifying characteristics of delegation (Barka & Sandhu, 2000): monotonicity which is who controls the delegated role; permanence which is the time duration of the delegation (until revoked or time constrained); totality which is the amount of permissions to be delegated (all or partial); and, cascading revocation which is the revocation of the role to the delegated user and anyone else that user may have delegated the role to. To our knowledge, these delegation capabilities are not available in any collaborative portal. As a web application, collaborative portals must also deal with information security issues related to: client side, server side, CGI scripts, protecting confidential documents, and denial of service attacks (http://www.w3.org/Security/ Faq/). From this list, we highlight some of the major security issues that are of interest to this chapter. Since collaborative web portals often run within web browsers, on the client side, there are many potential problems including: Active-X or Java Applets that can breach privacy, misuse of supplied personal information, network eavesdropping, the true level of encryption provided by SSL, the privacy of requests of web documents, known problems or security holes of popular browsers, and so on. On the server side, any content (documents or data) that moves through the web/application server on the way to a document repository or database may be vulnerable; rightly or not, a great deal of trust is placed not only in the secure transmission of information, but that the information once transmitted, is being handled securely by web, application, database, and document repository servers. In terms of documents themselves, there needs to be protection of content such as: insuring no lost updates when multiple users are allowed to edit the same document at the same time; providing a write one/read many model to limit who can check out and edit a document; offering a detailed history/versioning capability

434

to insure that all content is recoverable; and so on. From a web services perspective, messaging within a collaborative portal via SOAP (http:// www.oasis-open.org/specs/index.php#wssv1.0) will need to support message integrity, message confidentiality, and single message authentication via some type of general purpose mechanism that would associate a security token (user ID and role) with messages that are passed from client to server. Such a token could provide the guarantee that the content create/modify/delete request is being made from an authorized individual; this is essentially a re-authentication of the user for each major “action” that is taken in the portal with respect to content changes. Finally, to conclude this section, we revise four select portals focusing on their security capabilities and limitations. MediaWiki, launched in 2003, had the initial goal to create and maintain a free encyclopedia, where anyone can contribute (Lund, 2006). MediaWiki provides features that are intended to protect content in two ways: tracking all content modifications via versioning (to recover to an earlier version) and IP blocking from malicious users (creating fraudulent content). In addition to unregistered users, MediaWiki has two other roles: registered users who can create, edit, rename, and add content; and, administrators who may lock pages to prevent content changes. Clearly, in practice, these two roles (along with versioning and IP blocking) are very limiting for many applications (e.g., health care, engineering, banking/commerce, etc.) and their complex requirements. Next, consider the JBoss Portal (http:// www.jboss.org/jbossportal/), an open source platform that provides simple document protection and limits the change of document content to its author; this of course limits collaboration. Like MediaWiki, JBoss has two fixed roles that limit access to and modification of content. As a contract to JBoss, consider OpenGroupware (http://www. opengroupware.org), one of the first collaborative open source web portals (since 2000) that is document centric and allows owners to set permissions

Improving the Information Security of Collaborative Web Portals

(e.g., read, write, upload, delete, etc.) for each of their documents. Like the previous two, OpenGroupWare has two roles: system administrators and document owners. Lastly, SharePoint2007, a proprietary Microsoft collaborative web portal, provides a more fine-grained access control to an individual document or groups of documents. Nevertheless, authentication is checked against the windows user account, which constraints its usage to a Windows environment. While SharePoint is the most advanced of the four reviewed, like many Microsoft products, it suffers from feature creep; its wide range of capabilities make it difficult to utilize, particularly for stakeholders in non-technical domains.

bACKgroUnd ConCePts Axon is a full-function Wiki for content creation (WYSIWYG), document publishing (Web, PDF, RTF), document distribution (Email, Print, Fax), mobile access (limited via a BlackBerry), and role-based access control to allow collaboration among users that are sharing a particular task. As shown in Figure 1, the Axon Wiki is loaded from a web server into a multi-framed structure that includes a number of features. First, there is a top bar of functions (box A with Hide, History, Import, Export, Email, and Print); tabs for Spaces and Index (box B) where the Spaces tab is organized as accordions (e.g., Sales and Marketing, Finance, etc.) with parent topics (Status Reports), child topics (e.g., 2008 and 2007) and grandchild topics (e.g., Driver testing, Driver review, etc. in box C) with icons to create, edit, etc. topics (box D). The spaces, accordions, and their topic trees are customizable based on domain. Second, there is a main window with Topic and Doc (box E) tabs: the Topic tab is editable (XHTML) for the selected topic (i.e., AMSS Project Plan) with Edit, History, Intralink, etc. (box F); and, the Docs tab tracks the attached documents (e.g., PDF, Word, MPEG, etc.) for the topic, which is at the bottom

of Figure 1 and includes the ability to Attach, Copy, Paste, Cut, Check-Out, Check-In, Replace, Delete, History, and Email (box F). In addition, Axon has a number of other capabilities related to document creation, publishing, and viewing, as shown in Figure 2. At the top of the figure, the WYSIWYG editor is shown that is very MS-Word like to allow the easy creation of content (the XHTML document in the main window). This has been achieved using Ajax. Documents that are created in this fashion can be viewed (and/or changed) by other authorized users; a detailed version history is maintained. At the bottom of Figure 2, the documents of a topic can be assembled in various ways to publish a new combined document in different formats (e.g., PDF, RTF, etc.). This provides a very powerful capability in the Axon portal to create customized content from existing content without being concerned with cutting and pasting and differing file formats. Finally, in Figure 3, the Axon architecture is shown. The involved technologies are indicated for the reader’s interest. The clients can either be connected via workstations, laptops, or mobile devices. The Presentation Layer provides the typical means to access the underlying application. Brainstorm, which is the name of our software requirements elicitation effort (NSF, 2006; Pia, Demurjian, Vegad, Kopparti & Polineni 2007), contains two grayed boxes (application specific, changeable) and two ungrayed boxes (representing core Axon functionality). The Application Layer embodies many of the various underlying technologies to support Axon core functionality. The Data Layer allows Axon to be configured with any relational database as a backend. As shown in Figure 3, the realization of security in Axon is achieved in the Application layer via a combination of LDAP and our own custom RBAC implementation. LDAP, the Lightweight Directory Access Protocol, is utilized to track directory information on users during interactive sessions. Our focus in this chapter is on achieving RBAC,

435

Improving the Information Security of Collaborative Web Portals

Figure 1. The Axon WIKI main screen and topic/DOCs tabs

to allow the look-and-feel of Axon to be customized according to the permissions defined in the next Section on a role-by-role basis. Given the tight time constraints of the Phase I SBIR NSF grant (6 months), we prototyped a basic RBAC and other security using relational database tables to capture permissions, and once captured. These same tables are consulted to customize Axon based on the chosen role of a user; this will be described in the next section of this chapter in detail.

436

seCUrity CAPAbilities of AXon In this section, we begin by detailing the Axon security assumptions and concepts as a means to define the relevant RBAC security permissions that Axon will support. The intent is to delineate the security granularity level in terms of the different portions of the Axon Wiki that need to be controlled. While the permissions that are presented are specific to Axon, the reader will note throughout the discussion that the concepts and permissions can be generalized to apply to

Improving the Information Security of Collaborative Web Portals

Figure 2. Additional Axon document capabilities

a collaborative setting where there is a need to control, on a role-by-role basis, application content, access to documents, and GUI look-and-feel. To get started, in Figure 4, we provide a table of permissions for Axon. Basic assumptions regarding users and roles are: •





A User is identiied by a UserName (unique), UserID (unique), and User duration (UserStartTime and UserEndTime that the User is active). A role can be deined for any capability: Guest, Author, Manager, and Admin in Figure 4 are typical roles that would be available in Axon across application domains. For each role, there is an indication of the access to topics. A user can be associated with one or more roles in Axon. When a user starts a session with a tool, then the user must be authenticated (user name, password), and





once authenticated, the user is given a set of authorized roles to choose from. Once the user selects one of these roles, Axon customizes itself based on the chosen role using the permissions in the database. You can change role during your session, and log on to multiple sessions each with its own role. To isolate the user from the role, we introduce a group abstraction. Each User is the member of one or more Groups, where each Group is identiied by: GroupName (unique), GroupID (unique), and Group duration (GroupStartTime and GroupEndTime that the Group is active). Users can be in multiple groups and have multiple roles. Each group can have zero or more users. An active Session for a User limits the User to a particular Group . For any active Session, a User is limited to being in a particular group from a

437

Improving the Information Security of Collaborative Web Portals

Figure 3. The Axon architecture and its components



permission perspective. A User may have multiple simultaneously open Axon Wiki sessions with independent logons at any point in time. Permissions (as given in Figure 4) will be assigned to Roles, Roles will be assigned to Users/Groups, and Users/Groups will be assigned to Accordions.



In addition, there are other concepts that are relevant to define the security permissions as given in Figure 4 with respect to Axon. •

438

The Axon Wiki has a Project that contains multiple Accordions (i.e., Sales and Marketing, Finance, and Project AMSS in Figure 1), and for each Accordion, a Topic Tree and an Index is maintained. As deined,



each Accordion can have one or more Users, and each Accordion can have zero or more Groups (with each Group having zero or more Users as deined previously). The Axon Accordions represent different categories of topics to be maintained for each User. For example, the Sales and Marketing, and Finance accordions in Figure 1 could clearly be targeted to different user roles (Salespersons and CPAs, respectively), which are very type oriented and shared by multiple individuals. As shown for the Project AMSS accordion, there are speciic topics dedicated to the users (with roles) who will be authorized to access. The Topic Tree (see Figure 1 again) contains three levels of parent, child, and grandchild topics, where: each topic in this

Improving the Information Security of Collaborative Web Portals

tree is associated with exactly one XHTML page; each topic in this tree is associated with zero or more documents of all types (Word, PPT, PDF, GIF, etc.); and, the DOCS tab contains a list of documents (see lower half of Figure 1 again). Speciically, for the selected topic - all documents for the topic and its descendants are shown. Lastly, given the discussion so far, there are many detailed permissions that can be defined to realize alternative look-and-feel and usage security

of Axon on a role-by-role (user-by-user) basis. In terms of the topics in the Topic Tree, there are three permission assignments that are maintained: each Role can have one or more topics; each Group can have zero or more topics; and, each Accordion can have zero or more topics. Thus, when a user in a group logs onto Axon, the Accordions that are displayed are determined by the topics assigned to the authorized Accordions (which each have a list of zero or more topics), the Groups that the User is a member of (which each have a list of zero or more topics), and the specific Role (which

Figure 4. Axon security privileges

439

Improving the Information Security of Collaborative Web Portals

each have a list of zero or more topics) that the User is playing at logon time. The permissions in Axon, given in Figure 4, are explained using Figure 1 (with boxes A to F) by reviewing permissions for a set of sample roles. These four different roles are assigned typical permissions for different Wiki users: Guest is a user with very limited permissions; Author is a user able to create and manage topics and content, and perhaps even have limited capabilities for assigning permissions to other Users; Manager is the Author with additional capabilities regarding topics and content and more wide scale User management; and, Admin is a system administrator with access everywhere to all aspects of a Wiki. Specifically, for Axon: •





440

Permissions Related to Global Menu (see A in Figure 4): There are permissions on the Global Menu for Hide, History, Import, Export, Email, and Print. These permissions are Yes/No assigned on a role-by-role basis. The assignment of No means that the associated icon does not appear. Permissions Related to Tree Menu (see D in Figure 1): There are permissions on icons such as: New Topic to Create a new Topic; Copy to Make a Copy of an Existing Topic; Paste to Paste a Copy of an Existing Topic; Rename to Change the Name of a Topic; and, Archive to Store a new Version of the topic (XHTML page associated with the topic) and all of its associated documents. These permissions are Yes/No on a role-by-role basis. The assignment of No means that the associated icon does not appear. Permissions Related to Topic (or DOCS) Buttons (see E and F in Figure1): There are permissions on buttons such as: Edit, History, Intralink, etc. for Topic and Attach, Copy, Paste, etc. for DOCS. Most of these buttons are self explanatory except: Intralink lets you link topic to another



topic (in a different accordion); and, Publish which corresponds to the document assembler in Figure 2. These permissions are Yes/No assigned on a role-by-role basis. The assignment of No means that the associated icon does not appear. Permissions Related to Accordion Topics (see C in Figure 1): View means that the User (via his/her Role) has permission to view the Topic and XHTML page for the topic. Edit means that the User (via his/her Role) has permission to modify, delete, update, etc., the XHTML page for the topic.

The remainder of this section contains the set of relational database tables to handle the assumptions of Axon and its permissions as defined in Figure 4 in order to realize RBAC for Axon. To begin, we need to track the different Project configurations for Axon that contain the set of Accordions for each Project. Then, we can assign a particular Project configuration to a User. The ProjectInfo and AccordionInfo tables given below keep track of Projects and Accordions. ProjectAccordions maps Accordions into Projects. Start and End Times have been included for the ProjectAccordions table - that means that the Accordion is only visible for that time period. Basically, ProjectAccordions establishes the Accordions (e.g., in Figure 1, US Travel, Project Brainstorm, etc.) that are in a particular Axon configuration for a given application.

ProjectInfo AccordionInfo ProjectAccordions Next, we need to model the Topics (parent, child, and grandchild), and associate Topics with

Improving the Information Security of Collaborative Web Portals

Projects and Accordions to form a tree. Since a Project contains one or more Accordions, it makes sense to track the topics per Project/Accordions. Note that in this case, both ProjectID and AccordionID must be non-null. Regardless of permissions, for a given ProjectID, there are topics defined for each Accordion as identified by a AccordionID. The topics (and their subtopics and sub-subtopics) are all associated with a Accordion; all Topics (parents), SubTopic1 (children), and SubTopic2 (grandchildren) associated with a Accordion can be found by joining these three tables on TopicID and then selecting (or sorting) by AccordionID (or AccordionName if you join Topic and AccordionInfo tables). Thus, using Topics, SubTopic1, and SubTopic2, we are able to have a master list of all possible topics (and their subtopics and sub-subtopics) that are associated with each Accordion in a particular Project (ProjectID, AccordionID in combination).

Topic SubTopic1 SubTopic2 The TopicVersion table tracks different versions of a topic (as related to the XHTML page that is associated with each topic). The two tables, Attachment and AttachmentVersions, track the various documents (PDF, Word, etc.) associated with a Topic and their versions.

TopicVersion Attachment

AttachmentVersion Given these tables, we can proceed to define a set of tables to support permissions. For Users, Groups, and Roles, three tables are defined as below; all three have start and end times to delineate the duration of the User, Group, or Role. For authorizations there are two tables: one for User-to-Group authorization and a second for User –to-Role authorization. In this case, the start and end times are the durations of these authorizations and these times are constrained by the involved tables, e.g., the User-to-Role authorization is constrained by the start and end times of the User and of the Role. In addition, a table of PermissionInfo is defined, for the various types of permissions in Figure 4 (e.g., View, Edit, Archive, Replace, etc.).

UserInfo PermissionInfo GroupInfo RoleInfo UserGroupAuthorization UserRoleAuthorization Using these tables, we defined two possible options to model Topic authorization. First, in Option A below, the tables TopicUserAuth, TopicGroupAuth, and TopicRoleAuth are defined, to allow permissions to be established with respect

441

Improving the Information Security of Collaborative Web Portals

to Topics, by either UserID, GroupID or RoleID, respectively, for a given ProjectID/AccordionID (which are needed to clearly differentiate between the same Topic that may be defined in multiple ProjectID/AccordionID combinations). The idea is to utilize the three tables together to establish the Topics that are actually listed (under a Project/Accordions) for each User (as limited by Role and/or Group). The Topic table defined previously contains all Topics (subtopics and sub-subtopics) for all Accordions of a Project (a superset); the TopicUserAuth, TopicGroupAuth, and TopicRoleAuth customizes this superset to a subset (which may be the entire superset) of the Topics authorized to a User belonging to a Group and also playing a Role.

TopicUserAuth TopicGroupAuth TopicRoleAuth The advantage of Option A is that permissions logically and physically separate. Next, in Option B, we defined a generic TopicAuth table as:

TopicAuth This table would be used as follows: for User authorizations to Topics, ProjectID, and Accordi-

442

onID would be defined with GroupID and RoleID null; for Group authorizations to Topics, GroupID would be defined with UserID and RoleID null; and, for Role authorizations to Topics, RoleID would be defined with UserID and GroupID null. The advantage to Option B is a central location for permissions; the disadvantage is that we must keep track of all IDs (nulls) for all permissions and as a result, changes that we make impact all three types of permissions. In the short term, we selected Option B since it allows us to expand the authorizations to Accordions and Groups without introducing any new tables. For Wiki look-and-feel security, a set of three tables are defined to identify the widgets (buttons, etc.) of the Wiki to be controlled, the privileges of those widgets, and then to define the Wiki lookand-feel security on a role-by-role basis.

WikiLookandFeelAuthorization

Widget WidgetPrivilegeType

Widget is being used to refer to a button, icon, link, or any other aspect of the Wiki GUI that needs to be controlled. To illustrate these tables, let’s consider some actual tuples. First, there are all of the Widgets that are present in the Axon Wiki, which are uniquely identified (W1 to W6) and are all buttons (Table 1). Next, there are the privileges for each widget - Table 2 shows buttons that can either be Yes/No (for buttons) or ActiveIcon or NoIcon (for icons like Email in Figure 2). Lastly, on a role-by-role basis, for each role, we identify the widget and the allowable permission, as given by the tuples in Table 3 for role R1. To summarize, Widget is the list of all of the

Improving the Information Security of Collaborative Web Portals

Table 1. The Widget Table: Controllable Features of Axon WidgetID W1 W2 W3 W4 W5 W6

WidgetType Button Button Button Button Button Button

WidgetCategory

WidgetName

TopicLinks TopicLinks TopicLinks GlobalMenu TreeMenu TreeMenu

History Edit IntraLink History NewTopic Copy

Table 2. The WidgetPrivilegeType Table: Security Actions for each Privilege WidgetPrivilegeID

WidgetPrivilegeName

P1 P2 P3 P4

Yes No ActiveIcon NoIcon

look-and-feel components of the Wiki that need to be controlled, WidgetPrivilegeType tracks the status of each widget, while the WikiLookandFeelAuthorization tracks the actual authorization by role to each pair. Once the user has established a role for the session, the customization of the Wiki is controlled by lookups and joins using these tables. While Axon is a commercially product, there are limitations and enhancements that are still under investigation. First, in the current release, scalability has an impact on a number of features: system performance, and user, accordion, and topic document administration and their management. In its usage to-date, Axon’s document and project management system has not as yet been stressed with a high volume of users (and associated documents), which may impact performance. Therefore, ongoing work is focusing on

improving critical system design features. This includes the way that domains could be used, the way that folders must be organized and the structure of the entire system for locating topics and their associated documents. This may have a considerably impact on the response time, such a user navigates between accordions and their topics (expanding the tree – see A in Fig. 1). Scalability can impact system administration; realistic enterprise applications with thousands of accounts need to be precisely managed. To address this issue, we are carefully re-designing the user administration design and architecture, so that system wide functions (e.g., setting up new users, resetting passwords, disabling accounts, delete topics and/or documents loaded by mistake), can be easily performed. Scalability at the user level must also be addressed. For example, a user may have tens of accordions and each of these ac-

Table 3. The WikiLookandFeelAuthorization Table: Privileges by Role RoleID R1 R1 R1 R1 R1

WidgetID W1 W2 W3 W4 W5

WidgetPrivilegeID P1 P2 P3 P4 P5

443

Improving the Information Security of Collaborative Web Portals

cordions could have dozens of parent, child, and grandchild topics. When the user is interested in making this information available to other users by role, privilege definition and maintenance become monumental tasks, since permissions are at the topic level. Thus, we are currently working on a mechanism that seamlessly facilitates project definition. One approach under consideration is having permission-predefined topic trees that can be included in an accordion and then customized by name, privilege, etc. Finally, in the case where Axon is being used to manage a specific repository of shared documents for a defined timeframe, there must be a mechanism to archive and/or delete when a time limit is reached or an employee leaves, e.g., at the end of a contract term, all of the documents and information stored on the system must be transferred or deleted automatically.

relAted worK/fUtUre trends Our work in this chapter has been influenced by many different areas of research. To start, the role-based access control that we have designed for Axon has been influenced by our own past work (Demurjian, Ting & Thuraisingham 1993), as well as foundational work in RBAC (Sandhu, 1996) and standards (Ferrari, 2001). However, we have, for this version of Axon, kept the RBAC rather limited in its scope. In terms of security for collaborative computing, it is interesting to note that non-web-based computer supported cooperative work, proposed in the 1980s and explored into the early 1990s, addressed many issues on individual and group behaviors that are still relevant today (Grudin, 1991). This included: work in dynamic collaboration for a work group over space and time constraints (Ishii & Miyake, 1991); multimedia communication systems that support widely distributed work groups (Francik, Rudman, Cooper & Levine, 1991); work on security for computer supported collaborative work (Foley & Jacob, 1995); and, our own work (Demurjian,

444

Ting & Thuraisingham, 1993) on RBAC for collaborative computing environments. Much of this work has been supplanted by web-based research such as: an effort that seeks to construct a security architecture that is capable of being tailored for customer needs with a services-based approach for authentication, authorization, and RBAC (Demurjian, Ting, Balthazar, Ren, Phillips & P. Barr, 2001); a survey effort that seeks to identify the myriad of security issues that must be considered when multiple individuals collaborate and share information (Attalah, 2006); and, controlling access to information in a repository where there is a dynamic and unknown user population (Ray & Chakraborty, 2006) using a trust-based approach. There is a clear trend emerging towards exploring security solutions for collaborative applications and settings. In emerging paradigms in web computing, there have been a number of efforts that employ service-oriented architectures (SOA) or webservices as a basis for achieving security. For instance, (Bhatti, Joshi, Bertino & Ghafoor, 2003) proposes X-RBAC, an approach that leverages XML as a means to model RBAC and enforce defined security for an application by interacting with a security processor via SOAP (or other XML messaging). In a related effort (Bhatti, Ghafoor, Bertino & Joshi, 2005), X-GTRBAC provides a larger-scale infrastructure based on X-RBAC for handling security for enterprise (business) applications. The underlying security infrastructure in both of these efforts must take advantage of SOA and Web services in order to offer guarantees in terms of security assurance, data integrity, confidentiality, etc. (Bertino & Martino, 2006). There has also been an effort to provide a flexible, customer-driven security architecture for an open collaborative environment with authentication, authorization, accounting, and RBAC for Web and Grid settings (Phillips, Demurjian & Bessette, 2005). We agree that SOA and Web Services will be critical to security solutions that easily operate in a web-and-collaborative setting such as Axon.

Improving the Information Security of Collaborative Web Portals

In arriving at the decision of developing a straightforward database model and associated implementation to realize RBAC for Axon, we also considered other emerging technologies that support RBAC. First, XACML, the eXtensible Access Control Markup Language (http://sunxacml. sourceforge.net) is a web services policy constraint language that provides a standard infrastructure for security policy definition in a web context. There are many different implementations that have begun to emerge for XACML, e.g., one open source implementation was available, but in our timeframe (September 2006), the associated releases seemed premature and incomplete. As another example, consider the Bandit Role Engine (www. bandit-project.org/index.php/Role_Engine), an open source RBAC solution based on the available RBAC standard from NIST and XACML. However, it too had a to-be-announced Version 2 (no due date posted) that made it less attractive for our use. Thus, given our time constraint, and the fluidness of these products, we decided to not take the route of using an existing product; however, future updates may incorporate these solutions into the current prototype. Our security at the document level (the XHTML document AMSS Project Plan in Figure 1) is limited to controlling access to the entire document via the buttons Edit, History, Intralink, etc. However, in true collaboration, this document itself, being based on XML, may be partitioned into components, with RBAC utilized to control who can see/edit each of the components. In that regard, we have explored efforts that involve security for the semantic web, in general, and controlling access to XML documents, in particular. For the semantic web, there has been an effort that essentially established a road-map for security for the semantic web by identifying the key security issues (Thuraisingham, 2003), as well as an effort that has focused the discussion of security for the semantic web from the perspective of web databases and services that are appropriate (Ferrari & Thuraisingham, 2004).

In terms of potential solutions to control access to information at the document level, there have been two efforts of note: (Bertino & Ferrari, 2002) proposed a model for access control of XML documents with the policy including credentials and privileges; and, (Fan, Chan & Garofalakis, 2004) extended the concept of security views so that they are applicable to XML DTDs in order to screen information for XML instances before they are displayed. Both of these efforts of our interest as we proceed to fine-tune our security in Axon to control access to the components of the XHTML documents associated with topics on a role-by-role basis.

ConClUsion In this chapter, we have presented our effort on fine-grained RBAC for collaborative web portals. First, we reviewed three dominant access control models (RBAC, MAC, and DAC) and their suitability for collaborative portals, explored webclient-side, server-side, and document-centric information security issues and limitations, and addressed the capabilities and limitations of a select set of collaborative portals. Second, we described Axon collaborative portal, focusing on its architecture, components and content creation/ management capabilities (please see Figures 1 to 3 again). Third, using this as a basis, we detailed Axon’s attainment of a fine-grained, role-based access control security solution (please see Figure 3 again) at the application level, the document level (authoring and viewing), and the look-and-feel of the Wiki itself, as realized using a relational database; this discussion also included limitations of Axon. Finally, we reviewed related work and its influence on our effort, and using this as a basis, detailed emerging future trends in collaborative security. The resulting Axon wiki is being utilized in an actual collaboration to allow faculty researchers and health care providers to interact with one another to make information health information

445

Improving the Information Security of Collaborative Web Portals

technology decisions (Crowell, Agresta, Cook, Fifield, Demurjian, Carter, Becerra-Ortiz, Tracey, Vegad, & Polineni, 2009).

referenCes Attalah, M. (2000). Security issues in collaborative computing. In D. Chen & D. Lee (Eds.), Computing and combinatorics. (LNCS 4112, 2). Springer. NSF Award #0611053 to Serebrum. (2006). BrainStorm-collaborative customer requirements elicitation for distributed software development. Barka, E., & Sandhu, R. (2000). Framework for role-based delegation models. In Proceedings of the 16th Annual Computer Security Applications Conference (pp. 168-176). IEEE Computer Society Bell, D., & La Padula, L. (1975). Secure computer systems: Mathematical foundations model. (Tech. Rep. M74-244). Bedford, MA: The Mitre Corporation. Berhe, S., Demurjian, S., Ren, H., Devineni, R., Vegad, S., & Polineni, K. (n.d.). Axon–an adaptive collaborative Web portal. In Proceedings of 3rd International Workshop on Adaptation and Evolution in Web Systems Engineering. Retrieved from http://icwe2008.webengineering.org/Program/Workshops/ Bertino, E., & Ferrari, E. (2002). Secure and selective dissemination of XML documents. [ACM Press.]. ACM Transactions on Information and System Security, 5(3), 290–331. doi:10.1145/545186.545190 Bertino, E., & Martino, L. (2006). Security in SOA and Web services. In Proceedings of 2006 IEEE International Conference on Services Computing, 41. IEEE Computer Society

446

Bhatti, R., Ghafoor, A., Bertino, E., & Joshi, J. (2005). X-GTRBAC: An XML-based policy specification framework and architecture for enterprise-wide access control. [ACM Press.]. ACM Transactions on Information and System Security, 8(2), 187–227. doi:10.1145/1065545.1065547 Bhatti, R., Joshi, J., Bertino, E., & Ghafoor, A. (2003). Access control in dynamic XML-based Web services with X-RBAC. In Proceedings of International Conference on Web Services (pp. 234-249). Biba, K. (1977). Integrity considerations for secure computer systems. (Tech. Rep. TR-3153). Bedford, MA: The Mitre Corporation. Cheon Na, S. (2000). Role delegation in rolebased access control. In Proceedings of 5th ACM Workshop on Role-Based Access Control (pp. 39-44). ACM Press. Crowell, R., Agresta, T., Cook, M., Fifield, J., Demurjian, K., Carter, S., et al. (2009). Using a collaborative Web-based portal and wiki for making health information technology decisions. In S. Murugensan (Ed.), Handbook of research on Web 2.0, 3.0, and X.0: Technologies, business and social applications. Hershey, PA: IGI Global. Demchenko, Y., Gommans, L., de Laat, C., Oudenaarde, B., Tokmakoof, A., Snijders, M., & Van Buuren, R. (2005). Security architecture for open collaborative environment. In P. Sloot, A. Hoekstra, T. Priol, A. Reinefeld & M. Bubak (Eds.), Advances in grid computing. (LNCS 3470, pp. 589-599). Springer. Demurjian, S., Ting, T. C., Balthazar, J., Ren, H., Phillips, C., & Barr, P. (2001). A user role-based security model for a distributed environment. In B. Thuraisingham, R. van de Riet, K. Dittrich & Z. Tari (Eds.), Data and applications security: Developments and directions (Vol. IFIP 73, pp. 259-270). Springer

Improving the Information Security of Collaborative Web Portals

Demurjian, S., Ting, T. C., & Thuraisingham, B. (1993). User-role based security for collaborative computing environments. Journal of Multi-Media Review, 4(2), 40–47. Fan, W., Chan, C.-Y., & Garofalakis, M. (2004). Secure XML querying with security views. In Proceedings of the 2004 ACM SIGMOD International Conference on Management of Data (pp. 587-598). ACM Press. Ferraiolo, D. (2001). An argument for the rolebased access control model. In Proceedings of 6th ACM Symposium on Access Control Models and Technologies (pp. 142-143). ACM Press. Ferraiolo, D., Sandhu, R., Gavrilla, S., Kuhn, D., & Chandramouli, R. (2001). Proposed NIST standard for role-based access control. [ACM Press.]. ACM Transactions on Information and System Security, 4(3), 224–274. doi:10.1145/501978.501980 Ferrari, E., & Thuraisingham, B. (2004). Security and privacy for Web databases and services. In G. Goos, J. Hartmanis & J. van Leeuwen (Eds.), Advances in database technology. (LNCS 2992, p. 3923). Springer. Foley, S., & Jacob, J. (1995). Specifying security for computer supported collaborative working. Journal of Computer Security, 3(4), 233–254. Francik, E., Rudman, S., Cooper, D., & Levine, S. (1991). Putting innovation to work: Adoption strategies for multimedia communication systems. [ACM Press.]. Communications of the ACM, 34(12), 52–63. doi:10.1145/125319.125322 Grudin, J. (1991). CSCW-introduction to the special section. [ACM Press.]. Communications of the ACM, 34(12), 30–34. doi:10.1145/125319.125320 Ishii, H., & Miyake, N. (1991). Towards an open shared workspace: Computer and video fusion approach to teamworkstation. [ACM Press.]. Communications of the ACM, 34(12), 37–50. doi:10.1145/125319.125321

Linn, J., & Nystrom, M. (1999). Attribute certification: An enabling technology for delegation and role-based controls in distributed environments. In Proceedings of 4th ACM Workshop on Role-Based Access Control (pp. 121-130). ACM Press. Osborn, S., Sandhu, R., & Munawer, Q. (2000). Configuring role-based access control to enforce mandatory and discretionary access control policies. [ACM Press.]. ACM Transactions on Information and System Security, 3(2), 85–106. doi:10.1145/354876.354878 Phillips, C., Demurjian, S., & Bessette, K. (2005). A service-based approach for RBAC and MAC security. In Z. Stojanovic & A. Dahanayake (Eds.), Service-oriented software system engineering: Challenges and Practices (pp. 317-339). Hershey, PA: Idea Group. Pia, P., Demurjian, S., Vegad, S., Kopparti, S., & Polineni, K. (2007). BrainStorm: Collaborative customer requirements elicitation for distributed software development. In Proceedings of 31st Annual Software Engineering Workshop. Ray, I., & Chakraborty, S. (2006). A framework for flexible access control in digital library systems. In E. Damiani & P. Liu (Eds.), Data and applications security XX. (LNCS 4127, pp. 252266). Springer. Sandhu, R., Coyne, E., Feinstein, H., & Youman, C. (1996). Role-based access control models. [IEEE Computer Society.]. IEEE Computer, 29(2), 38–47. Thuraisingham, B. (2003). Security issues for the Semantic Web. In Proceedings of the 27th Annual International Conference on Computer Software and Applications (p. 632). IEEE Computer Society.

447

Improving the Information Security of Collaborative Web Portals

AdditionAl reAdings

Key terMs And definitions

Bullock, A., & Benford, S. (1999). An access control framework for multi-user collaborative environments. In Proceedings of the International ACM SIGGROUP Conference on Supporting Group Work, ACM Press, 140-149.

Axon: Axon is a collaborative web portal that supports fine-grained access control, multiple channel publication, business workflows, and an advanced document search. Brainstorm: Brainstorm is a toolkit for software requirements elicitation efforts. Collaborative Web Portals: Collaborative Web Portals allow multiple users to work together on a particular subject. There are many web portals that support collaboration in various domains. Of particular note is Wikipedia, the largest free online encyclopedia. Discretionary Access Control (DAC): DAC is an access control model mainly used in environments where the owners of a resource are permitted to pass their permissions to other subjects. DAC also incorporates the idea of group permissions. Mandatory Access Control (MAC): MAC is a system of access control that assigns security labels or classifications to system resources and allows access only to entities with distinct levels of authorization or clearance. Role-base Access Control (RBAC): RBAC is an access control model that reduces the administration overhead compared to other traditional access control models. In RBAC, permissions are assigned directly to roles, and then roles are assigned to users. As a result, permissions can change without changing user authorization. XHTML: XHTML is an application of XML, a more restrictive subset of SGML. XHTML documents allow for automated processing to be performed using standard XML tools unlike complex parsers for HTML.

Jaeger, T., & Prakash, A. (1996). Requirements of role-based access control for collaborative systems. In Proceedings of the 1st ACM Workshop on Role-Based Access Control, ACM Press, 53-64. Lin, D., Rao, P., Bertino, E., Li, N., & Lobo, J. (2008). Policy decomposition for collaborative access control. Estes Park, Colorado: In Proceedings of the 13th ACM Symposium on Access Control Models and Technologies, ACM Press, 103-112. Park, J. S., & Hwang, J. (2003). Role-based access control for collaborative enterprise in peer-to-peer computing environments. In Proceedings of the 8th ACM Symposium on Access Control Models and Technologies, ACM Press, 93-99. Shen, H., & Dewan, P. (1992). Access control for collaborative environments. In Proceedings of the 1992 ACM Conference on Computer-Supported Cooperative Work, ACM Press, 51-59. Tolone, W., Ahn, G., Pai, T., & Hong, S. (2005). Access control in collaborative systems. ACM Computing Surveys, ACM Press, 37(1), 29–41. doi:10.1145/1057977.1057979 Zhang, Z., Haffner, E., Heuer, A., Engel, T., & Meinel, T. (1999). Role-based access control in online authoring and publishing systems vs. document hierarchy. In Proceedings of the 17th Annual International Conference on Computer Documentation, ACM Press, 193-198.

448

449

Chapter 25

Web 2.0 Effort Estimation Emilia Mendes The University of Auckland, New Zealand

AbstrACt Web effort models and techniques provide the means for Web companies to formalise the way they estimate effort for their projects, and potentially help in obtaining more accurate estimates. Accurate estimates are fundamental to help project managers allocate resources more adequately, thus supporting projects to be inished on time and within budget. The aim of this chapter is to introduce the concepts related to Web effort estimation and effort forecasting techniques, and to discuss effort prediction within the context of Web 2.0 applications.

introdUCtion The Web is used as a delivery platform for numerous types of Web applications and services, ranging from complex e-commerce solutions with back-end databases to on-line personal static Web pages, blogs and wikis. Recently, the standards from Web 1.0 enabled the implementation of new technologies that allowed the use of the Web as originally envisaged by Tim Berners-Lee, making it a social Web. This represented a paradigm shift where the authoring of content moved from being controlled by just a few (and read by many), to the collaborative authoring DOI: 10.4018/978-1-60566-384-5.ch025

where all participate (Anderson, 2007). The applications and services that provide functionality that “aims to facilitate creativity, information sharing, and, most notably, collaboration among users” fall under the banner of Web 2.01. Regardless of being under the banner Web 1.0 or Web 2.0, the reality is that we currently have available a sheer diversity of Web application types, Web technologies and services, and such diversity is likely to continue growing. However, such diversity entails many challenges to those who develop/propose such applications, technologies and services. Complementary to the abovementioned scenario, there are many small to medium Web companies

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Web 2.0 Effort Estimation

Figure 1. Steps used to obtain an effort estimate (Mendes, 2007d)

worldwide, all bidding for as many Web projects as they can accommodate, delivering applications in domains often unfamiliar to developers (e.g. social networking applications, aggregation services, data ‘mash-ups’), and that use technologies and services with which these companies had no previous experience. This scenario only adds up to the current situation where most Web development projects suffer from unrealistic project schedules, leading to applications that are rarely developed on time and within budget (Reifer, 2000). In essence, regardless of the existing number of different Web technologies, services and application domains, Web companies need to have sound effort estimation in other to manage projects in a way that enables them to be delivered on time and within budget. The purpose of estimating effort is to predict the amount of effort (person/time) required to develop an application (and possibly also a service within the Web context), often based on knowledge of ‘similar’ applications/services previously developed. Figure 1 provides a general overview of an effort estimation process. Estimated characteristics of the new application/service to be developed, and its context (project) are the input, and effort is the output we wish to predict. For example, a given Web company may find that to

450

predict the effort necessary to implement a new e-commerce Web application, it will need to estimate early on in the development project the following characteristics: • •

• •



Estimated number of new Web pages. The number of functions/features (e.g. shopping cart, on-line forum) to be offered by the new Web application. Total number of developers who will help develop the new Web application Developers’ average number of years of experience with the development tools employed. The choice of main programming language used.

Of these variables, estimated number of new Web pages and the number of functions/features to be offered by the new Web application characterise the size of the new Web application; the other three, total number of developers who will help develop the new Web application, developers’ average number of years of experience with the development tools employed, and main programming language used, characterise the project - the context for the development of the new application, and are also believed to influence the amount of

Web 2.0 Effort Estimation

effort necessary to develop this new application. The project-related characteristics are co-jointly named ‘cost drivers’. No matter what the Web development is (application or service), in general the one consistent input found to have the strongest effect on the amount of effort needed to develop an application or service is size (i.e. the total number of server side scripts, the total number of Web pages), with cost drivers also playing an influential role. In most cases, effort estimation is based on past experience, where knowledge or data from past finished applications & projects are used to estimate effort for new applications & projects not yet initiated. The assumption here is that previous projects are similar to new projects to be developed, and therefore knowledge and/or data from past projects can be useful in estimating effort for future projects. The steps presented in Figure 1 can be repeated throughout a given Web development cycle, depending on the process model adopted by a Web company. For example, if the process model used by a Web company complies with a waterfall model this means that most probably there will be an initial effort estimate for the project, which will remain unchanged throughout the project. If a Web company’s process model complies with the spiral model, this means that for each cycle within the spiral process a new/updated effort estimate is obtained, and used to update the current project’s plan and effort estimate. If a Web company uses an agile process model, an effort estimate is obtained for each of the project’s iterations. In summary, a company’s process model will determine the amount of visibility an effort estimate has, and if the estimate is to be revisited or not at some point during the Web development life cycle. One point that is also important, however outside the scope of this Chapter, is the use of a value-based approach to software development and management (Boehm and Sullivan, 2000). In a nutshell what this means is that every management and/or development decision in an organisation

needs to be associated with this organisation’s value proposition and business goals. For example, if time-to-market is part of an organisation’s value proposition, this needs to be reflected in its software development and management processes. A detailed discussion on value-based software engineering (which equally applies to Web engineering) is given in (Boehm, 2003). It should be noted that cost and effort are often used interchangeably within the context of effort estimation since effort is taken as the main component of project costs. However, given that project costs also take into account other factors such as contingency and profit (Kitchenham et al., 2003) we will use the word “effort” and not “cost” throughout this chapter. As will be detailed later, our view regarding Web 2.0 coincides with that of Tim Berners-Lee, who asserted the following (Laningham, 2006): “Web 1.0 was all about connecting people. It was an interactive space, and I think Web 2.0 is of course a piece of jargon, nobody even knows what it means. If Web 2.0 for you is blogs and wikis, then that is people to people. But that was what the Web was supposed to be all along. And in fact, you know, this ‘Web 2.0’, it means using the standards which have been produced by all these people working on Web 1.0.”2 Therefore, we believe that the effort estimation techniques that have been used to date for Web effort estimation, in addition to the overall process of effort prediction, are equally applicable to Web 2.0 applications and services. For this reason, the next four Sections provide an introduction to the four most commonly used Web effort estimation techniques, namely Regression Analysis, Case-based Reasoning, Classification and Regression Trees, and Bayesian Networks. The diagram presented in Figure 1 is repeated for each of the effort estimation techniques in order to highlight the sequence of effort prediction steps that characterise each technique. Once these

451

Web 2.0 Effort Estimation

Figure 2. Example of a regression line (Mendes, 2007d)

techniques are described, the following Section presents a brief literature review of Web effort estimation studies, and finally our last Section provides our understanding of what represents Web 2.0 applications and services, and discusses the implications of the Web 2.0 paradigm for Web effort estimation.

regression AnAlysis Regression analysis is a technique where, using a dataset containing data on past finished Web

projects, an Equation is generated, representing the relationship between size, cost drivers, and effort. Such Equation is generated using a procedure that determines the “best” straight-line fit (see Figure 2) to a set of project data that represents the relationship between effort and size & cost drivers (Schofield, 1998). Figure 2 shows, using real data on Web projects from the Tukutuku Benchmarking project3, an example of a regression line that describes the relationship between log(Effort) and log(totalWebPages). It should be noted that the original data for the variables Effort and

Figure 3. Steps used to obtain an effort estimate using regression analysis

452

Web 2.0 Effort Estimation

totalWebPages have been transformed using the natural logarithmic scale to comply more closely with the assumptions of the regression analysis techniques. Details on these assumptions, how to identify variables that need transformation, and further information on regression analysis techniques are provided in (Mendes, 2007d). The Equation represented by the regression line in Figure 2 is as follows: log Effort = log a + b log totalWebPages

(1)

where log a is the point in which the regression line intercepts the Y-axis, known simply as the intercept, and b represents the slope of the regression line, i.e. its inclination, generically represented by the form, y = mx + c

(2)

Equation 1 shows a linear relationship between log(Effort) and log(totalWebPages). However, since the original variables have been transformed before the regression technique was employed, this equation needs to be transformed back such that it uses the original variables. The resultant equation is:

and cost drivers to be zero), a0 ... an are constants derived from past data and CD1…CDn are cost drivers that have an impact on effort. Regarding the regression analysis itself, two of the most widely used techniques are multiple regression (MR) and stepwise regression (SWR). The difference between these two techniques is that MR obtains a regression line using all the independent variables at the same time, whereas SWR is a technique that examines different combinations of independent variables, looking for the best grouping to explain the greatest amount of variation in effort. Both use least squares regression, where the regression line selected is the one that reflects the minimum values of the sum of the squared errors. Errors are calculated as the difference between actual and estimated effort and are known as ‘residuals’ (Schofield, 1998). A regression technique uses constant scalar values based on past project data; however, for anyone wishing to use this technique, and its generated model, the sequence of steps to use are 1, 2, 3 and 4. The sequence of steps (see Figure 3) is as follows: a) b)

Effort = a totalWebPagesb

(3) c)

Other examples of equations representing regression lines are given in Equations 4 and 5: EstimatedEffort = C + a0EstSizeNewproj + a1CD1... + anCDn

(4)

EstimatedEffort = C EstSizeNewproj 0 CD1 1 CDn a

a

an

(5)

where C is the regression line’s intercept, a constant denoting the initial estimated effort (assuming size

d)

Past data is used to generate a regression line (Step 1). An Equation (algorithmic model) is built from past data obtained in a) (Step 2). The model created in b) then receives, as input, values for the estimated size and cost drivers relative to the new project for which effort is to be estimated (Step 3). The model generates an estimated effort (Step 4).

Steps 1 and 2 are used to generate a regression model (Equation) for the first time, and later on whenever it is necessary to re-calibrate the original model. Recalibration may be needed after several new projects are finished and incorporated to the company’s data base of past finished projects. However, a company may also decide to

453

Web 2.0 Effort Estimation

Figure 4. Steps used to obtain an effort estimate using CBR

re-calibrate a model after every new project is finished, or to use the initial model for a longer time period. If the development team remains unchanged (and assumes that the team does not have an excessive learning curve for each new project) and new projects are similar to past projects there is no pressing need to re-calibrate a regression model too often.

Using CBR involves (Angelis & Stamelos, 2000): i.

CAse-bAsed reAsoning Case-based Reasoning (CBR) uses the assumption that ‘similar problems provide similar solutions’. It provides effort estimates by comparing the characteristics of the current project to be estimated, against a library of historical data from completed projects with known effort (case base).

ii.

Figure 5. Euclidean distance using two size features (n=2)

454

Characterising a new project p, for which an effort estimate is required, with variables (features) common to those completed projects stored in the case base. In terms of Web effort estimation, features represent size measures and cost drivers that have a bearing on effort. This means that, if a Web company has stored data on past projects where, for example, data represent the features effort, size, development team size, and tools used, the data used as input to obtaining an effort estimate will also need to include these same features. Use of this characterisation as a basis for finding similar (analogous) completed

Web 2.0 Effort Estimation

iii.

projects, for which effort is known. This process can be achieved by measuring the “distance” between two projects at a time (project p and one finished project), based on the features’ values, for all features (k) characterising these projects. Each finished project is compared to project p, and the finished project presenting the shortest distance overall is the ‘most similar project’ to project p. Although numerous techniques can be used to measure similarity, nearest neighbour algorithms using the unweighted Euclidean distance measure have been the most widely used to date in Web engineering. Generation of a predicted effort value for project p based on the effort for those completed projects that are similar to p. The number of similar projects taken into account to obtain an effort estimate will depend on the size of the case base. For small case bases (e.g. up to 90 cases), typical values use the most similar finished project, or the two or three most similar finished projects (1, 2, and 3 closest neighbours/analogues). For larger case bases no conclusions have been reached regarding the best number of similar projects to use. The calculation of estimated effort is obtained using the same effort value as the closest neighbour, or the mean effort for two or more closest neighbours. This is the common choice in Web engineering.

With reference to Figure 4, the sequence of steps used with CBR is as follows: a)

b)

The estimated size and cost drivers relating to a new project p are used as input to retrieve similar projects from the case base, for which actual effort is known (Step 1). Using the data from a) a suitable CBR tool retrieves similar projects to project p, and ranks these similar projects in ascending

c)

order of similarity, i.e. from ‘most similar’ to ‘least similar’ (Step 2). A suitable CBR tool calculates estimated effort for project p (Step 3).

Note that here no explicit model is built. When using CBR there are six parameters that need to be considered, which are as follows (Selby & Porter, 1998):

feature subset selection Feature subset selection involves determining the optimum subset of features that yields the most accurate estimation. Some existing CBR tools, e.g. ANGEL (Shepperd & Kadoda, 2001) optionally offer this functionality using a brute force algorithm, searching for all possible feature subsets. Other CBR tools (e.g. CBR-Works from tec:inno) have no such functionality, and therefore to obtain estimated effort, we must use all of the known features of a new project to retrieve the most similar finished projects, or reduce the number of features based on input from correlation analysis techniques.

similarity Measure The similarity measure records the level of similarity between different cases. Several similarity measures have been proposed in the literature to date, the three most popular currently used in the Web engineering literature (Angelis & Stamelos, 2000; Mendes et al., 2000; Selby & Porter, 1998) are the unweighted Euclidean distance, the weighted Euclidean distance, and the maximum distance. However, there are also other similarity measures available, and presented in (Angelis and Stamelos, 2000). Each of the three similarity measures aforementioned is described below. Unweighted Euclidean distance: The unweighted Euclidean distance measures the Euclidean (straight-line) distance d between two cases, where each case has n features. The equation used

455

Web 2.0 Effort Estimation

to calculate the distance between two cases x and y is the following: 2

d (x , y ) =

2

2

x 0 - y 0 + x 1 - y1 + ... + x n -1 - yn -1 + x n - yn

2

(6) where x0 to xn represent features 0 to n of case x; y0 to yn represent features 0 to n of case y. This measure has a geometrical meaning as the shortest distance between two points in an n-dimensional Euclidean space (Angelis & Stamelos, 2000) (see Figure 5). Figure 5 illustrates the unweighted Euclidean distance by representing coordinates in a twodimensional space, E2 as the number of features employed determines the number of dimensions, En. Given the following example (Table 1): The unweighted Euclidean distance between the new project 1 and finished project 2 would be calculated using the following equation: d=

2

100 - 350 + 20 - 12

2

= 250.128 (7)

The unweighted Euclidean distance between the new project 1 and finished project 3 would be calculated using the following equation: d=

2

100 - 220 + 20 - 25

2

= 120.104 (8)

Weighted Euclidean distance: The weighted Euclidean distance is used when features are given weights that reflect the relative importance of each feature. The weighted Euclidean distance measures the Euclidean distance d between two cases, where each case has n features and each feature has a weight w. The equation used to calculate the distance between two cases x and y is the following: 2

2

2

(9) where x0 to xn represent features 0 to n of case x; y0 to yn represent features 0 to n of case y; w0 to wn are the weights for features 0 to n. Maximum distance: The maximum distance computes the highest feature similarity, i.e. the one to define the closest analogy. For two points (x0,y0) and (x1,y1), the maximum measure d is equivalent to the Equation: d = max((x 0 - y 0 )2 ,(x 1 - y1 )2 )

(10)

This effectively reduces the similarity measure down to a single feature, although this feature may differ for each retrieval episode. So, for a given “new” project Pnew, the closest project in the case base will be the one that has at least one size feature with the most similar value to the same feature in project Pnew.

Using the weighted Euclidean distance, the distance between projects 1 and 3 is smaller than the distance between projects 1 and 2, thus project 3 is more similar to project 1 than project 2.

Table 1. project

totalWebPages

totalImages

1 (new)

100

20

2

350

12

3

220

25

456

2

d (x , y ) = w 0 x 0 - y 0 + w1 x 1 - y1 + ... + wn -1 x n -1 - yn -1 + wn x n - yn

Web 2.0 Effort Estimation

scaling Scaling (also known as standardisation) represents the transformation of a feature’s values according to a defined rule, such that all features present values within the same range and as a consequence have the same degree of influence on the result (Angelis & Stamelos, 2000). A common method of scaling is to assign zero to the observed minimum value and one to the maximum observed value (Kadoda et al., 2000), a strategy used by ANGEL and CBR-Works. Original feature values are normalised (between 0 and 1) by case-based reasoning tools to guarantee that they all influence the results in a similar way.

number of Analogies The number of analogies refers to the number of most similar cases that will be used to generate an effort estimate. With small sets of data, it is reasonable to consider only a small number of most similar analogues (Angelis & Stamelos, 2000). Some studies in Web engineering have only used the closest case/analogue(k = 1) to obtain an estimated effort for a new project (Mendes et al., 2002b), while others have also used the two closest and the three closest analogues (Mendes et al., 2000; Mendes et al., 2001; Mendes et al., 2003a).

Analogy Adaptation Once the most similar cases have been selected the next step is to identify how to generate (adapt) an effort estimate for project Pnew. Choices of analogy adaptation techniques presented in the literature vary from the nearest neighbour (Briand et al., 1999; Jeffery et al., 2001), the mean of the closest analogues (Shepperd & Kadoda, 2001), the median of the closest analogues (Angelis & Stamelos, 2000), the inverse distance weighted mean and inverse rank weighted mean (Kadoda et al., 2000), to illustrate just a few. The adapta-

tions used to date in Web engineering are the nearest neighbour, mean of the closest analogues (Mendes et al., 2000; 2001), and the inverse rank weighted mean (Mendes et al., 2002a, 2002b; 2003a; 2003b; 2003c). Each adaptation is explained below: Nearest neighbour: For the estimated effort Pnew , this type of adaptation uses the same effort of its closest analogue. Mean effort: For the estimated effort Pnew , this type of adaptation uses the average of its closest k analogues, when k > 1. This is a typical measure of central tendency, and treats all analogues as being equally important towards the outcome – the estimated effort. Median effort: For the estimated effort Pnew , this type of adaptation uses the median of the closest k analogues, when k > 2. This is also a measure of central tendency, and has been used in the literature when the number of selected closest projects is > 2 (Mendes et al. 2002a; 2002b). Inverse rank weighted mean: This type of adaptation allows higher ranked analogues to have more influence than lower ones over the outcome. For example, if we use three analogues, then the closest analogue (CA) would have weight = 3, the second closest analogue (SC) would have weight = 2, and the third closest analogue (LA) would have weight = 1. The estimated effort would then be calculated as: Inverse RankWeighed Mean =

3CA + 2SC + LA 6

(11)

Adaptation rules Adaptation rules are used to adapt the estimated effort, according to a given criterion, such that it reflects the characteristics of the target project (new project) more closely. For example, the estimated effort to develop an application a would be adapted such that it would also take into

457

Web 2.0 Effort Estimation

Table 2. project

totalWebPages (size)

totalEffort (effort)

Target (new)

100 (estimated value)

20 (estimated and adapted value)

Closest analogue

350 (actual value)

70 (actual value)

account the estimated size of application a. The adaptation rule that has been employed to date in Web engineering is based on the linear size adjustment to the estimated effort (Mendes et al. 2003a; 2003b), obtained as follows: •



Once the most similar analogue in the case base has been retrieved, its effort value is adjusted and used as the effort estimate for the target project (new project). A linear extrapolation is performed along the dimension of a single measure, which is a size measure strongly correlated with effort. The linear size adjustment is calculated using the equation presented below.

Effortnew Pr oject =

Effort finished Pr oject Size finished Pr oject

Sizenew Pr oject (12)

Given the following example (Table 2): The estimated effort for the target project will be calculated as:

Effortnew Pr oject =

(13)

When we use more than one size measure as feature, the equation changes to:

Eest .P

æ ö 1 çç q =x Eact Sest .q ÷÷÷ = ççå ÷ q çç q =1 Sact .q ÷÷÷ è >0 ø

(14)

where: q is the number of size measures used as features. Eest.P is the Total Effort estimated for the new Web project P. Eact is the Total Effort for the closest analogue obtained from the case base. Sest.q is the estimated value for the size measure q, which is obtained from the client. Sact.q is the actual value for the size measure q, for the closest analogue obtained from the case base. This type of adaptation assumes that all projects present similar productivity, which may not

Figure 6. Example of a regression tree for Web effort estimation

458

70 100 = 20 350

Web 2.0 Effort Estimation

reflect the Web development context of numerous Web companies worldwide.

ClAssifiCAtion And regression trees Classification and Regression Trees (CART) (Brieman et al., 1984) use independent variables (predictors) to build binary trees, where each leaf node represents either a category to which an estimate belongs, or a value for an estimate. The former situation occurs with classification trees and the latter occurs with regression trees, i.e. whenever predictors are categorical (e.g. Yes/ No) the CART tree is called a classification tree and whenever predictors are numerical the CART tree is called a regression tree. In order to obtain an estimate one has to traverse tree nodes from root to leaf by selecting the nodes that represent the category or value for the independent variables associated with the project to be estimated. For example, assume we wish to obtain an effort estimate for a new Web project using as its basis the simple regression tree structure presented in Figure 6. This regression tree was generated from data obtained from past completed Web applications, taking into account their existing values of effort and independent variables (e.g. new Web pages (WP), new images (IM), and new features/functions (FN)). The data used to build a CART model is called a ‘learning sample’, and once a tree has been built it can be used to estimate effort for new projects. Assuming that the estimated values for WP, IM and FN for a new Web project are 25, 15 and 3, respectively, we would obtain an estimated effort of 35 person hours after navigating the tree from its root down to leaf ‘Effort = 35’. If we now assume that the estimated values for WP, IM and FN for a new Web project are 56, 34 and 22, respectively, we would obtain an estimated effort of 85 person hours after

navigating the tree from its root down to leaf ‘Effort = 85’. A simple example of a classification tree for Web effort estimation is depicted in Figure 7. It uses the same variable names as that shown in Figure 6; however these variables are now all categorical, where possible categories (classes) are “Yes” and “No”. The effort estimate obtained using this classification tree is also categorical, where possible categories are “High effort” and “Low effort”. A CART model constructs a binary tree by recursively partitioning the predictor space (set of all values or categories for the size variables and cost drivers judged relevant) into subsets where the distribution of values or categories for effort is successively more uniform. The partition (split) of a subset S1 is decided on the basis that the data in each of the descendant subsets should be “purer” than the data in S1. Thus node “impurity” is directly related to the amount of different values or classes in a node, i.e. the greater the mix of classes or values, the higher the node “impurity”. A “pure” node means that all the cases (e.g. Web projects) belong to the same class, or have the same value. The partition of subsets continues until a node contains only one class or value. Note that sometimes not all the initial size variables and cost drivers are used to build a CART model; rather, only those variables that are related to effort are selected by the model. This means that a CART model can be used not only to produce a model that can be applicable for effort prediction, but also to obtain insight and understanding of the factors relevant to estimate effort. A detailed description of CART and its use for Web effort estimation is presented in (Mendes, 2007d). The sequence of steps (see Figure 8) followed here are as follows: a) b)

Past data is used to generate a CART model (Step 1). A CART model is built based on the past data obtained in a) (Step 2).

459

Web 2.0 Effort Estimation

Figure 7. Example of a classification tree for Web effort estimation (Mendes, 2007d)

Figure 8. Steps used to obtain an effort estimate using CART

c)

d)

The model created in b) then is traversed using as input, values/categories for the estimated size and cost drivers relative to the new project to which effort is to be estimated (Step 3). The leaf that results from the traversal provides an effort estimate (Step 4).

bAyesiAn networKs A Bayesian Network (BN) is a model that supports reasoning with uncertainty due to the way in which it incorporates existing knowledge of a complex domain (Pearl, 1988). This knowledge is represented using two parts. The first, the qualitative part, represents the structure of a BN

460

as depicted by a directed acyclic graph (digraph) (see Figure 9). The digraph’s nodes represent the relevant variables (factors) in the domain being modelled, which can be of different types (e.g. observable or latent, categorical). The digraph’s arcs represent the causal relationships between variables, where relationships are quantified probabilistically (Woodberry et al., 2004). The second, the quantitative part, associates a node probability table (NPT) to each node, its probability distribution. A parent node’s NPT describes the relative probability of each state (value) (Figure 9, nodes ‘Pages complexity’ and ‘Functionality complexity’); a child node’s NPT describes the relative probability of each state conditional on every combination of states of its parents (Figure 9, node ‘Total Effort’). So, for

Web 2.0 Effort Estimation

Figure 9. A small Bayesian Network and three NPTs

Figure 10. A real Bayesian Network Structure

example, the relative probability of ‘Total Effort’ being ‘Low’ conditional on ‘Pages complexity’ and ‘Functionality complexity’ being both ‘Low’ is 0.7. Each row in a NPT represents a conditional probability distribution and therefore its values sum up to 1 (Pearl, 1988). Once a BN is specified, evidence (e.g. values) can be entered into any node, and probabilities for the remaining nodes automatically calculated using Bayes’ rule (Pearl, 1988). Therefore BNs can be used for different types of reasoning, such as predictive, diagnostic, and “what-if” analyses

to investigate the impact that changes on some nodes have on others (Stamelos et al., 2003). Bayesian Networks have only recently been used for Web effort estimation (Mendes, 2007a; 2007b; 2007c), and some of the BN models built were found to produce significantly better prediction than regression-based, CBR-based, CART-based, mean- and median-based models. One of them, built using a combination of data from the Tukutuku database and expert opinion from a Web developer (domain expert) with more than 10 years of Web development

461

Web 2.0 Effort Estimation

and management experience, is presented in Figure 10. It is our view that BNs provide more flexibility than the other techniques aforementioned, due to the following reasons:

a)

b) c)







A BN model, i.e. a BN structure and its NPTs, can be automatically built from data (data-driven model). The downside is that this choice requires a reasonable number of project data in order to derive as many of the NPTs’ probabilities as possible. A BN structure and its NPTs can be elicited from a combination of data and feedback from Domain Experts (hybrid model). A Domain Expert within this context can be one or more Web developers, and/or one or more Web project managers who have expertise in effort estimation. A BN structure and its NPTs can be completely elicited from Domain Experts (expert-driven model).

This means that if a Web company has only a small amount of past data on finished Web projects, it can either build a hybrid or expertdriven model. The sequence of steps (see Figure 11) followed here are as follows:

d)

Past data and/or knowledge of past finished projects is used to build a BN model (Step 1). A BN model is built based on the past data and/or expertise obtained in a) (Step 2). Evidence is entered on some of the nodes that are part of the model created in b). Such evidence corresponds to values/categories for the estimated size and cost drivers relative to the new project to which effort is to be estimated (Step 3). The model generates a set of probabilities associated with each of the effort states (Step 4).

In the case of BNs, the estimated effort will have an associated probability distribution over a set of states. So assuming that estimated effort was measured using two states – high and low, the BN model will provide the probability that estimated effort will be high, and the probability the estimated effort will be low. There are techniques that can be used in order to obtain estimated effort as a discrete value; however they are outside the scope of this chapter. Interested readers, please refer to (Mendes, 2007a).

Figure 11. Steps used to obtain an effort estimate using a BN model

462

Web 2.0 Effort Estimation

web effort estiMAtion literAtUre review There have been numerous attempts to model effort estimation for Web projects, but, except for (Mendes, 2007a; 2007b; 2007c), none have used a probabilistic model beyond the use of a single probability distribution. Table 1 presents a summary of previous studies. Whenever two or more studies compare different effort estimation techniques using the same dataset, we only include the study that uses the greatest number of effort estimation techniques. Mendes and Counsell (2000) were the first to empirically investigate Web effort prediction. They estimated effort using machine-learning techniques with data from student-based Web projects, and size measures harvested late in the project’s life cycle. Mendes and collaborators also carried out a series of consecutive studies (Fewster & Mendes, 2001; Mendes & Kitchenham, 2004; Mendes & Counsell, 2000; Mendes & Mosley, 2002; Mendes et al., 2001; Mendes et al., 2002a; 2002b; 2002c; 2003a; 2003b; 2005a; 2005b) where models were built using multivariate regression and machinelearning techniques using data on student-based and industrial Web projects. Recently Mendes (2007a; 2007b, 2007c) investigated the use of Bayesian Networks for Web effort estimation, using data on industrial Web projects from the Tukutuku database. Other researchers have also investigated effort estimation for Web projects: Reifer (2000; 2002) proposed an extension of the COCOMO model, and a single size measure harvested late in the project’s life cycle. None were validated empirically. This size measure was later used by Ruhe et al. (2003), who further extended a software engineering hybrid estimation technique, named CoBRA© (Briand et al., 1998), to Web projects, using a small data set of industrial projects, mixing expert judgement and multivariate regression. Later, Baresi et al. (2003), and Mangia et al. (2003) investigated ef-

fort estimation models and size measures for Web projects based on a specific Web development method, namely the W2000. Finally, Costagliola et al. (2006) compared two sets of existing Webbased size measures for effort estimation. Table 3 shows that most Web effort estimation studies to date used data on student-based projects; estimates obtained by applying Stepwise regression or Case-based reasoning techniques; accuracy measured using the Mean Magnitude of Relative Error (MMRE), followed by the Median Magnitude of Relative Error (MdMRE) and Prediction at 25% (Pred(25)) (Mendes et al., 2003b).

web 2.0 And effort estiMAtion In terms of our understanding of what constitutes Web 2.0 applications and services, we employ the proposal of Paul Anderson (2007), inspired by Tim O’Reilly’s detailed discussion of what O’Reilly believed Web 2.0 meant (O’Reilly, 2005). Anderson (2007) describes a set of Web-based services and applications he believes express the fundamentals of the Web 2.0 concept. These services and applications are not technologies per se; however were built “using the building blocks of the technologies and open standards that underpin the Internet and the Web. These include blogs, wikis, multimedia sharing services, content syndication, podcasting, and content tagging services.” A detailed introduction to each of these services and example applications are provided in Section 2 of (Anderson, 2007). A strong characteristic of Web 2.0 services and applications seems to be to provide Rich Internet Applications where Web browser-based applications and services provide functionality, graphics and usability services similar to, or better than, desktop applications. Here a specific group of technologies has been adopted for the delivery of Web 2.0 applications and services: Ajax, which stands for Asynchronous Javascript + XML. In

463

464

Case study

Not detailed

Case study

Case study

Case study

Case study

Case study

Case study

Formal experiment

Not detailed

Case study

Case study

2nd (Reifer, 2002)

3rd (Mendes et al., 2001)

4th (Fewster & Mendes, 2001)

5th (Mendes et al., 2002a)

6th (Mendes et al., 2002b)

7th (Ruhe et al., 2003)

8th (Mendes et al., 2003)

9th (Baresi et al. 2003)

10th (Mangia & Paiano, 2003)

11th (Costagliola et al., 2006)

12th (Mendes, 2007a; 2007b; 2007c)

Type

1st (Mendes & Counsell, 2000)

Study

1 – (150)

1 – (15)

unknown

1 - (30)

2 - (37 and 25)

1 - (12)

1 - (37)

1 - (25)

1 - (37)

1 - (37)

1 - (46)

2 - (29 and 41)

# datasets (# datapoints)

professionals

professionals

unknown

Computer Science students

Honours and postgraduate CS students

An exponential model named Metrics Model for Web Applications (MMWA) Linear regression, Stepwise regression, Case-based reasoning, Classification and Regression Trees

Bayesian Networks, Stepwise Regression, Mean and Median effort, Case-based reasoning, Classification and regression Trees

Functional, Navigational Structures, Publishing and Multimedia sizing measures Web pages, New Web pages, Multimedia elements, New multimedia elements, Client side Scripts and Applications, Server side Scripts and Applications, All the elements that are part of the Web Objects size measure Total Web pages, New Web pages, Total Images,New Images, Features off-the-shelf (Fots), High & Low effort Fots-Adapted, High & Low effort New Features, Total High & Low Effort Features

Ordinary least squares regression

Case-based reasoning

Page Count, Media Count, Program Count, Reused Media Count (only one dataset), Reused Program Count (only one dataset), Connectivity Density, Total Page Complexity Information, Navigation and Presentation model measures

COBRA, Expert opinion, Linear regression

Web Objects

Case-based reasoning, Linear regression, Stepwise regression, Classification and Regression Trees

Page Count, Media Count, Program Count, Reused Media Count, Reused Program Count, Connectivity Density, Total Page Complexity

Honours and postgraduate Computer Science students professionals

Case-based reasoning

Requirements and Design measures, Application measures

Honours and postgraduate Computer Science students

Generalised Linear Model

Linear regression, Stepwise regression

Structure metrics, Complexity metrics, Reuse metrics, Size metrics

Length size, Reusability, Complexity, Size

Honours and postgraduate Computer Science students

WEBMO (parameters generated using linear regression)

Case based reasoning, Linear regression, Stepwise regression

Prediction techniques

Honours and postgraduate Computer Science students

Web objects

Page Count, Reused Page Count, Connectivity, Compactness, Stratum, Structure

Size Measures

professionals

2nd year Computer Science students

Subjects

Table 3. Summary Literature Review (Mendes, 2007a)

Bayesian Networks provided superior predictions

All techniques provided similar prediction accuracy

-

-

-

COBRA

Linear/stepwise regression or case-based reasoning (depends on the measure of accuracy employed)

-

Linear Regression

-

Case based reasoning for high experience group

Best technique(s)

MMRE, MdMRE, MEMRE, MdEMRE, Pred(25), Boxplots of residuals, boxplots of z

MMRE, MdMRE, Pred(25), Boxplots of residuals, boxplots of z

-

-

MMRE, Pred(25), Boxplots of absolute residuals

MMRE, Pred(25), Boxplots of absolute residuals

MMRE, MdMRE, Pred(25), Boxplots of absolute residuals

MMRE, MdMRE, Pred(25), Boxplots of absolute residuals

Goodness of fit

MMRE

Pred(n)

MMRE

Measure Prediction Accuracy

Web 2.0 Effort Estimation

Web 2.0 Effort Estimation

addition to Ajax, Flash is also a technology often used to deliver such services and applications (Anderson, 2007). Another important characteristics of Web 2.0 services and applications is to incorporate “collaboration, contribution and community” in the functionality offered, aiming to make the Web into a “social” Web (Anderson, 2007). This means that the design of services and applications must provide the means to improve and assist large user participation, where some services are also designed to capture users’ interactions and make use of them for self improvement (Anderson, 2007). Such characteristics affect Web effort estimation differently, as will be discussed later. As outlined in the ‘Introduction’ Section of this Chapter, a Web effort estimation process receives as input to the process size measures and cost drivers, and produces as output an effort estimate for a new project (see Figure 1). The way in which inputs are transformed into an output can vary widely: i) it can often be done subjectively, for example, when based on expert-based estimation; ii) however, sometimes it can also be based on the use of algorithmic and machinelearning techniques (an example of the former is multivariate regression, and examples of the latter are Case-based reasoning, Classification and Regression Trees, Bayesian Networks); and finally iii) occasionally it can also be based on the combination of expert opinion & algorithmic or machine-learning techniques. In terms of the estimation process itself, it is our view and experience that its main steps remain unchanged regardless of whether the effort estimate relates to a new Web 1.0 or Web 2.0 application. In other words, size measures and cost drivers still need to be provided as input to this process, a mechanism to obtaining an estimate still needs to be applied, and an effort estimate still needs to be produced as output. There are, however, numerous factors (listed below) that can affect the success of a Web effort

estimation process, which we believe are currently exacerbated when applied to estimating effort for Web 2.0 applications. Note that many of these factors are quite challenging to a large number of Web companies worldwide, including those that develop applications that would be nowadays classified as Web 1.0: 1. 2.

3. 4.

5. 6.

7.

8.

Expertise in the application domain relating to the new application to be developed. Expertise in the interface design and type of application (e.g. collaborative) relating to the new application to be developed. Expertise in the technologies that will be used with the new application. Expertise in the programming languages that will be used to implement the new application. Mature processes where all those involved in a project know their roles and duties. Quality mechanisms in place aiming to solve any early problems that arise during a Web development project, thus mitigating future problems. Understanding of the set of factors that have a bearing on effort. Such factors may represent characteristics of the new application to be developed (e.g. size), project characteristics (e.g. number of developers working on the project, developers’ previous experience with the programming language, tools being used), customer characteristics (e.g. customer’s knowledge with using Web applications, customer’s technical knowledge). Understanding of how to measure the size of the ‘problem’ to be developed, where ‘problem’ within this context represents the new Web application to be implemented.

Although all the abovementioned factors can be challenging when estimating effort for Web 2.0 applications, we believe that the three factors that currently make the estimation of accurate

465

Web 2.0 Effort Estimation

effort estimates for Web 2.0 applications are the following: 1.

2.

3.

The lack of expertise in the interface design and type of application (e.g. collaborative) to be developed. The lack of Web developers’ and project managers’ knowledge of some of the technologies and programming languages that characterise Web 2.0 applications. The absence of standards in regards to how to size a Web application in general.

In relation to point 1 above, the characteristics of Web 2.0 applications require developers to implement applications that enable collaboration & co-authorship, that are highly interactive and that provide a high level of usability. Some of these characteristics, such as usability, are also common to Web 1.0 applications. However, even usability takes on another dimension when we refer to Web 2.0 applications, since these applications must enable users to be active participants. Therefore, when for Web 1.0 applications a usability requirement would be, for example, a shopping cart to make clear where the ‘add to basket’ button is, changes to a usability requirement such as ‘provide real-time map movement and scaling’, as available with Google maps, a Web 2.0 application. There is a large body of knowledge on collaboration, co-authorship, interactive tools and environments that can be of benefit to Web companies and Web developers. However, it is our view that it may be necessary that experienced Web 2.0 developers and/or researchers help bridge the gap between practitioners and the new development paradigm inherent to Web 2.0 applications. The solution to point 2 above is similar to that we suggest for point 1: the participation of Web 2.0 developers and/or researchers in helping practitioners understand the Web 2.0 technology and programming languages they are not familiar with.

466

In terms of point 3, the literature has several proposals of Web measures for effort estimation, some of which have already been referenced in the Literature Review Section. The set of measures that to date has been used the most was proposed as part of the Tukutuku benchmarking project (Mendes et al. 2005a). 25 variables characterising a Web application and its development process were identified. These size measures and cost drivers were obtained from the results of a survey investigation (Mendes et al. 2005a), using data from 133 on-line Web forms that provided quotes on Web development projects. They were also confirmed by an established Web company and a second survey involving 33 Web companies in New Zealand, so making them very likely a suitable set of measures to use for very early Web effort estimation. The issue here is that when these measures were proposed they were based on feedback from Web companies that developed Web 1.0 applications. Therefore it is necessary to either revisit these measures in light of Web 2.0 applications. In addition, it is also our belief that points 1 and 2 above may be fundamental when determining the size measures and cost drivers that have a bearing on effort. In summary, it is our view and experience that the process used to estimate effort for Web 1.0 applications does not change when estimation now focuses on Web 2.0 applications. One still needs to identify size measures and cost drivers believed to have a bearing on effort, and use data on past finished projects, expert opinion, or a combination of data and developers’ expertise, to predict an effort estimate for a new project.

ConClUsion Effort estimation is the process by which a company estimates early on in the development life cycle the amount of effort needed to complete a Web development project on time and within budget.

Web 2.0 Effort Estimation

There are numerous challenges to providing a sound effort estimate, some of which were discussed in this chapter. In addition, numerous techniques and models have been used for Web effort estimation. These techniques were briefly introduced, and a literature survey of previous Web effort estimation studies was also presented. Finally, we provided our views on the challenges faced by companies developing Web 2.0 applications, and suggestions on what is needed to make effort estimation successful when developing Web 2.0 applications.

referenCes Anderson, P. (2007). What is Web 2.0? Ideas, technologies, and implications for education. JISC Technology and Standards Watch. Retrieved in September 2008, from http://www.jisc.ac.uk/ media/documents/techwatch/tsw0701b.pdf Angelis, L., & Stamelos, I. (2000). A simulation tool for efficient analogy based cost estimation. Empirical Software Engineering, 5, 35–68. doi:10.1023/A:1009897800559 Baresi, L., Morasca, S., & Paolini, P. (2003, September 3-5). Estimating the design effort of Web applications. In Proceedings Ninth International Software Measures Symposium (pp. 62-72). Boehm, B. (2003, March). Value-based software engineering. In ACM Software Engineering Notes. Boehm, B., & Sullivan, K. J. (2000). Software economics: A roadmap. In the Future of Software Engineering, International Conference on Software Engineering, Limerick, Ireland.

Briand, L. C., El Emam, K., & Bomarius, F. (1998). COBRA: A hybrid method for software cost estimation, benchmarking, and risk assessment. In Proceedings of the 20th International Conference on Software Engineering (pp. 390-399). Briand, L. C., El-Emam, K., Surmann, D., Wieczorek, I., & Maxwell, K. D. (1999). An assessment and comparison of common cost estimation modeling techniques. In Proceedings of ICSE 1999, Los Angeles (pp. 313-322). Brieman, L., Friedman, J., Olshen, R., & Stone, C. (1984). Classification and regression trees. Belmont: Wadsworth. Costagliola, G., Di Martino, S., Ferrucci, F., Gravino, C., Tortora, G., & Vitiello, G. (2006). Effort estimation modeling techniques: A case study for Web applications. In Proceedings of the Intl. Conference on Web Engineering (ICWE’06) (pp. 9-16). Fewster, R. M., & Mendes, E. (2001). Measurement, prediction, and risk analysis for Web applications. In Proceedings of the IEEE METRICS Symposium (pp. 338-348). Jeffery, R., Ruhe, M., & Wieczorek, I. (2001). Using public domain metrics to estimate software development effort. In Proceedings of the 7th IEEE Metrics Symposium, London (pp. 16-27). Kadoda, G., Cartwright, M., Chen, L., & Shepperd, M. J. (2000). Experiences using case-based reasoning to predict software project effort. In Proceedings of the EASE 2000 Conference, Keele, UK. Kitchenham, B. A., Pickard, L. M., Linkman, S., & Jones, P. (2003). Modelling software bidding risks. IEEE Transactions on Software Engineering, 29(6), 542–554. doi:10.1109/ TSE.2003.1205181

467

Web 2.0 Effort Estimation

Laningham, S. (Ed.). (2006, August 22). Tim Berners-Lee. Podcast, developerworks interviews, IBM Web site. Retrieved in September 2008, from http://www-128.ibm.com/developerworks/ podcast/

Mendes, E., & Mosley, N. (2002). Further investigation into the use of CBR and stepwise regression to predict development effort for Web hypermedia applications. In Proceedings ACM/IEEE ISESE (pp. 79-90), Nara, Japan.

Mangia, L., & Paiano, R. (2003) MMWA: A software sizing model for Web applications. In Proceedings of the Fourth International Conference on Web Information Systems Engineering (pp. 53-63).

Mendes, E., Mosley, N., & Counsell, S. (2001). Web measures–estimating design and authoring effort. IEEE MultiMedia, 8(1), 50–57. doi:10.1109/93.923953

Mendes, E. (2007a). Predicting Web development effort using a Bayesian network. In . Proceedings of EASE, 07, 83–93. Mendes, E. (2007b). The use of a Bayesian network for Web effort estimation. In Proceedings of International Conference on Web Engineering (pp. 90-104). (LNCS 4607). Mendes, E. (2007c). A comparison of techniques for Web effort estimation. In Proceedings of the ACM/IEEE International Symposium on Empirical Software Engineering (pp. 334-343). Mendes, E. (2007d). Cost estimation techniques for Web projects. Hershey, PA: IGI Global. Mendes, E., & Counsell, S. (2000). Web development effort estimation using analogy. In Proceedings of the 2000 Australian Software Engineering Conference (pp. 203-212). Mendes, E., Counsell, S., & Mosley, N. (2000, June). Measurement and effort prediction of Web applications. In Proceedings of 2nd ICSE Workshop on Web Engineering (pp. 57-74), Limerick, Ireland. Mendes, E., & Kitchenham, B. A. (2004). Further comparison of cross-company and within-company effort estimation models for Web applications. In Proceedings IEEE Metrics Symposium (pp. 348-357).

468

Mendes, E., Mosley, N., & Counsell, S. (2002a). The application of case-based reasoning to early Web project cost estimation. In Proceedings of IEEE COMPSAC (pp. 393-398). Mendes, E., Mosley, N., & Counsell, S. (2002c, June). Comparison of length, complexity, and functionality as size measures for predicting Web design and authoring effort. IEE Proceedings. Software, 149(3), 86–92. doi:10.1049/ipsen:20020337 Mendes, E., Mosley, N., & Counsell, S. (2003a). Do adaptation rules improve Web cost estimation? In Proceedings of the ACM Hypertext Conference 2003 (pp. 173-183), Nottingham, UK. Mendes, E., Mosley, N., & Counsell, S. (2003b). A replicated assessment of the use of adaptation rules to improve Web cost estimation. In Proceedings of the ACM and IEEE International Symposium on Empirical Software Engineering (pp. 100-109), Rome, Italy. Mendes, E., Mosley, N., & Counsell, S. (2005a). Investigating Web size metrics for early Web cost estimation. Journal of Systems and Software, 77(2), 157–172. doi:10.1016/j.jss.2004.08.034 Mendes, E., Mosley, N., & Counsell, S. (2005b). The need for Web engineering: An introduction. In E. Mendes & N. Mosley (Eds.), Web engineering (pp. 1-26). Springer-Verlag.

Web 2.0 Effort Estimation

Mendes, E., Watson, I., Triggs, C., Mosley, N., & Counsell, S. (2002b, June). A comparison of development effort estimation techniques for Web hypermedia applications. In Proceedings IEEE Metrics Symposium (pp. 141-151), Ottawa, Canada. Mendes, E., Watson, I., Triggs, C., Mosley, N., & Counsell, S. (2003c). A comparative study of cost estimation models for Web hypermedia applications. Empirical Software Engineering Journal, 8(2), 163–196. doi:10.1023/A:1023062629183 O’reilly. T. (2005). What is Web 2.0: Design patterns and business models for the next generation of software. O’Reilly Web site. O’Reilly Media Inc. Retrieved on September 30, 2005, from http://www.oreillynet.com/pub/a/oreilly/tim/ news/2005/09/30/what-is-web-20.html Pearl, J. (1988). Probabilistic reasoning in intelligent systems. San Mateo, CA: Morgan Kaufmann. Reifer, D. J. (2000). Web development: Estimating quick-to-market software. IEEE Software, 17(6), 57–64. doi:10.1109/52.895169 Reifer, D. J. (2002). Ten deadly risks in Internet and intranet software development. IEEE Software, (Mar-Apr): 12–14. doi:10.1109/52.991324 Ruhe, M., Jeffery, R., & Wieczorek, I. (2003). Cost estimation for Web applications. [Portland, OR.]. Proceedings of ICSE, 2003, 285–294. Schofield, C. (1998). An empirical investigation into software estimation by analogy. Unpublished doctoral dissertation, Dept. of Computing, Bournemouth University. Selby, R. W., & Porter, A. A. (1998). Learning from examples: Generation and evaluation of decision trees for software resource analysis. IEEE Transactions on Software Engineering, 14, 1743–1757. doi:10.1109/32.9061

Shepperd, M. J., & Kadoda, G. (2001). Using simulation to evaluate prediction techniques. In Proceedings of the IEEE 7th International Software Metrics Symposium (pp. 349-358), London, UK. Stamelos, I., Angelis, L., Dimou, P., & Sakellaris, E. (2003). On the use of Bayesian belief networks for the prediction of software productivity. Information and Software Technology, 45(1), 51–60. doi:10.1016/S0950-5849(02)00163-5 Woodberry, O., Nicholson, A., Korb, K., & Pollino, C. (2004) Parameterising Bayesian networks. In Proceedings of the Australian Conference on Artificial Intelligence (pp. 1101-1107).

Key terMs And definitions Bayesian Networks: A technique that enables the construction of a model that supports reasoning with uncertainty due to the way in which it incorporates existing knowledge of a complex domain (Pearl, 1988). This knowledge is represented using two parts. The first, the qualitative part, represents the structure of a BN as depicted by a directed acyclic graph (digraph). The digraph’s nodes represent the relevant variables (factors) in the domain being modelled, which can be of different types (e.g. observable or latent, categorical). The digraph’s arcs represent the causal relationships between variables, where relationships are quantified probabilistically (Woodberry et al., 2004). The second, the quantitative part, associates a node probability table (NPT) to each node, its probability distribution. A parent node’s NPT describes the relative probability of each state (value). A child node’s NPT describes the relative probability of each state conditional on every combination of states of its parents. Each row in a NPT represents a conditional probability distribution and therefore its values sum up to 1 (Pearl: 1988)

469

Web 2.0 Effort Estimation

Case-Based Reasoning: A technique that assumes that similar problems present similar solutions. It provides estimates by comparing the characteristics of the current project to be estimated against a library of historical information from completed projects with a known effort (case base). Effort Estimation: Process employed to predict the necessary amount of labour units to accomplish a given task, based on knowledge of previous similar projects and other project characteristics that are believed to be related to effort. Project characteristics (independent variables) are the input, and effort (dependent variable) is the output we wish to predict Expert-Based Effort Estimation: Represents the process of estimating effort by subjective means, and is often based on previous experience from developing/managing similar projects. This is by far the mostly used technique for Web effort estimation. Within this context, the attainment of accurate effort estimates relies on the competence and experience of individuals (e.g. project manager, developer) Regression Analysis: A technique where, using a dataset containing data on past finished projects, an Equation is generated, representing the relationship between size, cost drivers, and effort. Such Equation is generated using a procedure that determines the “best” straight-line fit to a set of project data that represents the relationship between effort and size & cost drivers Web 2.0: Set of new technologies, implemented using as basis the standards from Web

470

1.0, that allow the use of the Web as originally envisaged by Tim Berners-Lee, making it a social Web. This represented a paradigm shift where the authoring of content moved from being controlled by just a few (and read by many), to the collaborative authoring where all participate (Anderson, 2007). The applications and services that provide functionality that “aims to facilitate creativity, information sharing, and, most notably, collaboration among users” fall under the banner of Web 2.04 Web Effort Estimation: Process employed to predict the necessary amount of labour units to accomplish a given task, based on knowledge of previous similar Web projects and other Web project characteristics that are believed to be related to effort. Web project characteristics (independent variables) are the input, and effort (dependent variable) is the output we wish to predict. Typical inputs are the number of Web pages to be developed, number of scripts to be programmed, number of new images and animations

endnotes 1 2

3 4

http://en.wikipedia.org/wiki/Web_2.0 A transcript of the podcast is available at: http://www-128.ibm.com/developerworks/ podcast/dwi/cmint082206.txt http://www.metriq.biz/tukutuku http://en.wikipedia.org/wiki/Web_2.0

472

Chapter 26

A Social Web Perspective of Software Engineering Education Pankaj Kamthan Concordia University, Canada

AbstrACt The discipline of software engineering has been gaining increasing signiicance in computer science and engineering education. A technological revitalization of software engineering education requires a considerate examination from both human and social perspectives. The goal of this chapter is to adopt a systematic approach towards integrating Social Web technologies/applications in software engineering education, both inside and outside the classroom. To that regard, a pedagogical patterns-assisted methodology for incorporating Social Web technologies/applications in software engineering education is proposed and explored. The potential prospects of such integration and related concerns are illustrated by practical examples. The directions for future research are briely outlined.

introdUCtion It could be said that today’s civilization runs on software and will likely continue to do so in the foreseeable future. It is therefore natural to devote much attention to the life of software from its inception to its operation and eventually its retirement, and software engineering is the discipline that does that. As software engineering matures, its body of DOI: 10.4018/978-1-60566-384-5.ch026

knowledge is shared, communicated, and consumed. Indeed, in the last decade, software engineering has been playing an increasingly prominent role in computer science and engineering undergraduate and graduate curricula of Universities around the world (Rezaei, 2005; Surakka, 2007). As with other disciplines, software engineering education (SEE) needs to be sensitive to the variations and evolution of the social and technical environment around it. In particular, any changes in the information technology (IT) environment and the

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

A Social Web Perspective of Software Engineering Education

generation of students making use of it (Palfrey & Gasser, 2008) need to be reflected in SEE, if it leads to viable opportunities and demonstrated benefits. There have been calls for a reform of SEE in which technology is given a prominent place (Frailey, 1998; Shaw, 2000; Lethbridge, 2000). However, there have been relatively few efforts (Kamthan, 2008b) in the direction of precisely and objectively articulating the integration of IT in SEE. The Social Web, or as it is more commonly referred to by the pseudonym Web 2.0 (O’Reilly, 2005; Shuen, 2008), is the perceived evolution of the Web in a direction that is driven by ‘collective intelligence,’ realized by IT, and characterized by user participation, openness, and network effects. Web 2.0 differs from its predecessor, the so-called Web 1.0, in many ways (Cormode & Krishnamurthy, 2008) including legal, social, and technical dimensions. The focus of this chapter is to assess the implications of the Social Web as it pertains to teaching and learning of software engineering (Kamthan, 2009), including synergistic interactions between teachers and students. The focus is on pedagogical affordances of the Social Web for SEE. The rest of the chapter is organized as follows. First, the background necessary for later discussion is provided and related work is presented. This is followed by a proposal for a systematic introduction of the Social Web environment consisting of technologies/applications for SEE, labeled as SW4SE2. The prospects of SW4SE2 are illustrated using practical examples. Next, challenges and directions for future research are outlined. Finally, concluding remarks are given.

bACKgroUnd In this section, the human and the social aspect of software engineering is briefly traced, and the role of IT in realizing it in practice is outlined.

An overview of the human and social Aspects of software engineering In the past decade, there has been shift in the theory and practice of software engineering in a few notable directions, one of which is the acknowledgement of the significance of the role of non-technical aspects. In particular, it has been realized that there is a need to foster a social environment in software engineering at several different levels, and this is also increasingly being seen as significant to SEE (Layman et al., 2005). This acknowledgement has come at multiple levels, of which people and process are two critical, interrelated elements.

People In general, large-scale software development is conducted in teams. The nature of software teams is fundamentally social in which the principles for fostering cooperative behavior (Tenenberg, 2008) including that for face-to-face communication, repeatability and reciprocity of interactions, mutual monitoring, and acquiescence/sanctions apply. The premise to any software development is ethics, which is a human value. Indeed, all pertinent decisions, including software quality concerns (Kamthan, 2008a), are ‘driven’ by ethical considerations aiming to develop software for the benefit of the society at-large.

Process For development of software aimed for the general public, the industry has essentially rejected bureaucratic process environments that discourage social interaction. In recent years, the software process environments have become increasingly collaborative (Williams, 2000), embracing the client and user involvement. Indeed, agile methodologies and Open Source Software (OSS) ecosystems are exemplars of this movement.

473

A Social Web Perspective of Software Engineering Education

The human aspect, and indeed the social aspect, of software engineering trickle down to process workflows. It has long been recognized that requirements elicitation is a social process (Macaulay, 1993). For example, ethnomethodology and interviews are two socially-inclined requirements elicitation techniques. The views of a given software architecture are often derived from viewpoints of different stakeholder types. The crucial design decisions, such as selection and application of architectural styles or design patterns, often depend upon collaboration, deliberation, and mutual cooperation. The success of Pair Programming, one of the core practices of Extreme Programming (XP), strongly depends on the collaboration between the pair (Aiken, 2004) and an acknowledgement of its social nature (Chong & Hurlbutt, 2007). Finally, acceptance testing, particularly usability testing, requires collaboration and cooperation of representative user groups.

software engineering education and evolution of the web The computing environment enabling the social component of software engineering has taken time to get established. In the 1970s, the technologies/ applications to support the social aspect of software engineering were not mature, and in the 1980s, they were largely limited to the use of electronic mail (e-mail). It was the 1990s, particularly the ascent of the Web, that opened new vistas for people to communicate in a variety of ways on a global scale, and strengthened the foundations of electronic learning (e-learning) as a field of study (Anderson, 2008). This led to new directions in teaching and learning, and eventually to permeation of e-learning in computer science education. For example, the use of Java applets in illustrating the dynamics of complex algorithms in a classroom has been emphasized (Kamthan, 1999) and the benefits of hypertext for relating and navigating through

474

software artifacts have been shown (Bompani, Ciancarini, & Vitali, 2002). For the sake of this chapter, IT is a single encompassing term that describes any technology that helps to produce, archive, manipulate, communicate, and/or disseminate information. In the last few years, the movement in IT as it pertains to the Web has been characterized by an apparent ‘humanization’ and even ‘socialization.’ Indeed, the TIME Magazine’s naming the Person of the Year for 2006 to be ‘You,’ and the increasingly number of products pre-fixed with an ‘i’ and post-fixed with a ‘2.0’are a sign of this paradigmatic change. The term ‘Social Web’ is relatively new. However, its origins can be traced back to commercial Web applications such as Web portals of the mid-to-late 1990s that pioneered the concept of user-generated information, electronic feedback from customers, and personalization of users. The examples of these include dmoz.org, imdb.com, and amazon.com, respectively. There appear to be three primary factors that can be attributed to bringing the vision of the Social Web to a mainstream realization: 1.

2.

3.

Focus on Humans. The Social Web has enabled a many-to-many bidirectional communication paradigm in which people are primary and technology is secondary. This has led to a shift from the conventional humancomputer interaction to human-information interaction and computer-mediated social interaction. Viability of Technology. The underlying device and technological infrastructure of the Social Web has evolved and matured, and implementations of its technologies are available as open source. Involvement of Collective. There is awareness, interest, and large-scale participation by the public in general in the Social Web environment. This has led to a convergence of conventional social networks

A Social Web Perspective of Software Engineering Education

Table 1. An outline of SW4SE2 1. Deciding the Scope of Software Engineering Knowledge 2. Implications from Theories of Learning, Theories of Pedagogy, and Pedagogical Patterns 3. Selecting and Applying Suitable Social Web Technologies/Applications to Software Engineering Educational Activities

and contemporary technological networks (Kleinberg, 2008).

Related Work on the Use of Social Web in Education There have relatively few initiatives so far for integrating Social Web technologies/applications in education (Bernsteiner, Ostermann, & Staudinger, 2008; Hsu, 2008; Sigala, 2007). The uses of Wiki as a teaching tool in software engineering have been reported (Gotel et al., 2007; Parker & Chao, 2007). However, their theoretical foundation, specifically their correspondence to any theories of teaching and learning, is unclear. A learning process based on the Socialization, Externalization, Combination, Internalization (SECI) model of knowledge management that uses Social Web technologies has been suggested (Chatti et al., 2007). However, the treatment is largely peripheral and one-sided: the critical issue of ‘Ba’ (shared context for knowledge creation and transformation in the SECI model) is not discussed, the precise advantages of Social Web towards teaching and learning are not given, and the corresponding limitations have not been pointed out. The potential of Social Web for elearning has been highlighted (Shank, 2008) but the focus is more on technology than on teaching or learning. Finally, a systematic approach for the use of Wiki in preparing e-learning course materials in the area of grid computing has been suggested (Shih, Tseng, & Yang, 2008). However, the discussion is limited by the promotion of an unsubstantiated list of advantages and absence of any limitations of Wiki.

Feasibility

integrAting soCiAl web teChnologies/ APPliCAtions in softwAre engineering edUCAtion SW4SE2 is a slight variation of a specialization of IT4SE2 (Kamthan, 2008b), a methodology for integrating IT in SEE. It consists of a sequence of steps as shown in Table 1. The steps 1–3 in Table 1 are nonlinear, non-mutually exclusive, and could be granularized further if necessary. The steps of SW4SE2 must also be feasible in order to be practical. This is hardly automatic and needs to be part of the overall instructional design and resource management policy.

step one: deciding the scope of software engineering Knowledge This step is about scoping the software engineering knowledge in need for technological impetus for the purpose of teaching and learning. These topics need to be communicated and, for that, certain educational activities are normally put into practice. The software engineering topics could correspond to the knowledge areas of the Guide to the Software Engineering Body of Knowledge (SWEBOK) and the Software Engineering Education Knowledge (SEEK). It is beyond the scope of this chapter to suggest an authoritative list of such topics, which is likely to depend on step 2 and vary across courses and across educational institutions. The examples in the later sections provide a glimpse into some topics and activities that lend themselves to a technological treatment.

475

A Social Web Perspective of Software Engineering Education

step two: implications from theories of learning, theories of Pedagogy, and Pedagogical Patterns This step is about adopting an educational perspective towards teaching and learning. There are multiple theories of teaching and learning that are applicable. The two theories of educational psychology on which theories of pedagogy in general and theories of instructional design and pedagogical strategies in particular are being modeled today are objectivism and constructivism (Smith & Ragan, 1999). From an objectivist view, knowledge is external to an individual (and therefore objective), and therefore learning involves a ‘transfer of knowledge’ from the instructor to the learner. From a constructivist view, knowledge is not external to an individual, and therefore learning involves constructing one’s own knowledge from one’s own experiences. In recent years, constructivism has received attention in SEE (Hadjerrouit, 2005). There has been much debate over the years in the educational community on the virtues and drawbacks of objectivism and constructivism; however, there are signs of reconciliation (Cronjé, 2006). Indeed, it is the author’s contention that the two views should be seen as complementary rather than competing and, in certain cases, nonmutually exclusive rather than conflicting. The theory of constructivism has been broadly classified into the categories of individual, radical, and social. In social constructivism, social interaction plays a fundamental role in cognitive development. The two related notions in social constructivism are the More Knowledgeable Other (MKO) and the Zone of Proximal Development (ZPD). The MKO possesses more knowledge about a particular topic and/or better skills for a particular task than the student and, can, for example, be a teacher, more capable peer, or a computer. The ZPD is the area between what the student can do and can not do, even with assis-

476

tance. During the time a student is in the ZPD for a particular task, the MKO provides scaffolding to assist the student in acquiring knowledge and/ or skills necessary to carry out the task. The theory of constructivism has also been an inspiration for constructinism. There are similarities between social constructivism and social constructinism: in both cases, groups collaboratively construct knowledge, leading to the creation of a small culture of shared artifacts with shared meanings. The spirit of creative experimentation (Staron, 2007) has a crucial place in a constructivist and constructionist approaches to SEE, both inside and outside the classroom, and is especially relevant to the context of this chapter. However, there are also notable differences: in social constructivism, the focus is on a student’s learning that takes place due to their interactions in a group; in social constructinism the focus is on the artifacts that are created through the social interactions of a group (such as during a course project). A pedagogical strategy (teaching approach) must be sensitive to the theories of learning that have been adopted and currently in practice but should not be constrained by any one of them. A classroom use of Social Web technologies/applications in SEE could be more objectivist than constructivist where the educator plays the role of an ‘instructor.’ This could, for example, entail preparing Social Web technologies/applicationsbased lesson plans and lectures, and encouraging questions from students on a timely basis without severely interrupting the flow of the lectures. The use of Social Web technologies/applications in assignments and course projects could be more socially constructivist and/or socially constructionist than objectivist where the educator plays the role of a ‘mediator’ or a ‘guide.’ This could, for example, entail providing a balance between discipline and flexibility to the students in carrying out a software project with minimal guidance and timely feedback by the educator as

A Social Web Perspective of Software Engineering Education

and when needed: the crucial aspect being that the students play the primary role and the educator plays the secondary role.

Pattern-Assisted Software Engineering Education A pattern is an empirically proven solution to a recurring problem in a given context (Buschmann, Henney, & Schmidt, 2007). The reliance on conceptually reusable knowledge such as patterns that is garnered from past experience and expertise can be useful in any endeavor, including SEE. For novice teachers, pedagogical patterns can serve as a source of guidance and/or reference. There can be a number uses of pedagogical patterns in a SEE setting including the following: 1.

2.

3.

Patterns Related to Course Information Delivery. There are patterns applicable to curriculum design (Bergin, 2003), classroom demonstrations (Schmolitzky, 2007), and classroom teaching (Eckstein et al., 2003). Patterns Related to Connectivity. There are patterns for soliciting feedback (Bergin, 2006; Eckstein, Bergin, & Sharp, 2002a) from students. Patterns Related to Course Project. There are patterns for administering course projects (Bergin, 2006; Eckstein, Bergin, & Sharp, 2002b; Hayes et al., 2006; Naruse et al., 2008), selecting and adopting a suitable process model (Coplien & Harrison, 2005; Elssamadisy & West, 2006), and realizing team collaboration (Schümmer & Lukosch, 2007).

It is evident that pedagogically-inclined patterns are implicitly or explicitly based on real-world practices of theories of teaching and learning. For example, the SHOW IT RUNNING pattern (Schmolitzky, 2007) is aligned with an objectivist view as it advocates illustrating the use of software in the class to the students by the

teacher, while the MAKE IT THEIR PROBLEM pattern is aligned with a constructivist view. It should however be noted that, in general, such an approach needs to take into account several factors including availability of suitable pedagogical patterns that can sufficiently ‘map’ learning activities, the selection of pedagogical patterns based on their alignment with the adopted pedagogical strategy (Bennedsen & Eriksen, 2006), and clearly identified value to learners (Fincher, 1999).

step three: selecting and Applying suitable social web technologies/ Applications to software engineering educational Activities Table 2 highlights the relationship between common types of activities resulting from student– student or teacher–student interactions in SEE and Social Web technologies/applications. Table 2 is by no means exhaustive. Also, for a given collaboration context, there may be more than one applicable Social Web technologies/ applications, and they may not necessarily be equally suitable. The following criteria could be used for the selection of a Social Web technology/application: (1) nature of information (such as sensory modality) being communicated, (2) alignment with teaching and learning goals, (3) considerations for openness (proprietary versus non-proprietary), (4) maturity (stability), and (5) feasibility (availability and affordability). The criteria are minimal and non-mutually exclusive. An objective, third-party review of a candidate technology/application can also help in making the decision for adoption.

Situating Social Web-Based Educational Activities in Learning Taxonomies and Theories of Pedagogy A learning taxonomy is a classification of learning objectives. There are various proposals for

477

A Social Web Perspective of Software Engineering Education

Table 2. A mapping between activities in SEE and corresponding Social Web technologies/applications Activity Syndication Classroom Demonstrations, Presentation of Audio/Video Interviews

Social Web Technology/Application News Feed (RSS, Atom) Mashup, Podcast, Shared Presentation (YouTube, Google Video)

Acquisition of Core Lecture Material or Supplementary Course Material

Collaborative Note Taking (NoteMesh), Folksnomy, Wiki, Social Bookmarking (del.icio.us, Google Bookmarks, vi.sualize.us)

Participation in Asynchronous and Synchronous Communication, Conducting a Discussion

Blog, Mailing List, News Group (Yahoo! Groups, Google Groups), Podcast

Researching for Assignment or Software Project

Collaborative Annotation (Google Notebook, Microsoft OneNote, Similarr, Xoost)

Managing Software Project

Collaborative Project Management (FogBugz, Mindquarry)

Scheduling

Web Calendar (Google Calendar, Jiffle)

Brainstorming Developing Software Process Artifacts Managing Software Source Files

learning taxonomies and several uses of learning taxonomies (Fuller et al., 2007) that apply to SEE: they help understand the learning process, provide a shared ‘language’ for describing learning outcomes, provide a description of learning stages at which a student ‘operates’ on a topic, define curriculum objectives of a course in terms of the desired level of understanding of each topic being covered, design a course at different levels of granularity in time, structure modes of assessment, and so on. The categories of cognitive, affective, and psychomotor are commonly used as the basis for classifying learning objectives in a learning taxonomy. The educational activities in Table 2 belong to one or more of these categories. For example, brainstorming using a mind map belongs to all categories. The educational activities in Table 2 also correspond to certain theories of instructional design. The Attention, Relevance, Confidence, Satisfaction (ARCS) model of motivational design (Keller & Suzuki, 1988) is aligned with social constructivism, and several activities in Table 2 help operationalize the specific examples, active

478

Mind Map (bubbl.us, comapping, Mindomo) Collaborative Read/Write, Traceable, Versioned Application (Wiki, Google Docs) Collaborative, Distributed Source File Sharing (SourceForge)

participation, future usefulness, modeling, and learner control elements of the ARCS model. For example, collaborative note taking enables learner control. It is possible, as needed, to conceal/reveal information included in Wikis. This style of presentation supports the progressive disclosure of information that is known to assist learning (Lieberman, 2007); if each information fragment presented on the Wiki is considered as an epitome, then this in turn is in agreement with the elaboration theory of instruction (Reigeluth & Stein, 1983). Finally, the educational activities in Table 2 also correspond to some of the proposed nine events of instruction (Gagné et al., 2005), namely (1) Gain Attention; (2) Inform Learner of Objective; (3) Recall Prior Knowledge; (4) Present Material; (5) Provide Guided Learning; (6) Elicit Performance; (7) Provide Feedback; (8) Assess Performance; and (9) Enhance Retention and Transfer. For example, classroom demonstrations and interviews correspond to (4), providing comments regarding the assignment to a posting on the course mailing

A Social Web Perspective of Software Engineering Education

list by students or providing suggestions on the course project Wiki run by students correspond to (7), and so on.

Situating Social Web-Based Educational Activities in E-Learning Taxonomies For the sake of this chapter, the terms presence and electronic communication (or e-communication) can be characterized as follows. There is presence if there is physical or virtual availability of both the teacher and the student during the interchange of information for the purpose of pedagogy. There is e-communication if (1) there is e-communication between the teacher and the student at the time of instruction or (2) the e-communication is the primary communication medium for completing the course. Then, the type of e-learning that can take place in a Social Web environment can be placed into an adaptation of a known classification of e-learning (Negash & Wilcox, 2008): 1.

2.

Type I E-Learning: Physical/Virtual Presence and E-Communication (Faceto-Face). In this case, there is a traditional face-to-face classroom setting in which both the teacher and the student are physically present in the classroom at the time of transmission of information, and therefore there is presence. There are certain Social Web technologies/applications that enable type I e-learning including verbal annotations on illustrations of behavioral/dynamical phenomena communicated via animations, videos, and podcasts. This model can be extended outside the classroom where there is interaction between the teacher and the student and among students. Type II E-Learning: Virtual Presence and Synchronous E-Communication. In this case, the teacher and the student do not meet physically, however, they always meet

3.

virtually during transmission of information, and therefore there is presence. There are a variety of Social Web technologies/ applications that enable type II e-learning including use of collaborative software and interactive audio/video. Type III E-Learning: No Presence and Asynchronous E-Communication. In this case, the teacher and the student do not meet during transmission of information, physically or virtually, and therefore there is no presence. For example, the teacher can prepare the course-related information in advance, and subsequently deliver it; the student then accesses this information at a later time. Thus, there is an evident time delay between delivery and access of information. There are a variety of Social Web technologies/applications that enable type III e-learning including blogs, newsgroups, podcasts, and Wiki.

The other types of e-learning that are possible are hybrid combinations of types I−III. It is important to note that the type of e-learning in which there is no presence and there is no e-communication (such as dissemination of information via electronic media including compact discs, digital video discs, and universal serial bus-based flash drives) is not considered as relevant in the context of the Social Web.

examples of integrating social web technologies/Applications in software engineering education As evident from Table 2, the Social Web lends various opportunities for teacher−student and student−student communication. A subset of these is considered in the rest of the section.

479

A Social Web Perspective of Software Engineering Education

Figure 1. A tag cloud embedded in the lecture notes on responsibility-driven design

Collaborative Learning In an objectivist approach to SEE, lectures and tutorials are still the norm where a teacher (or a teaching assistant) often makes use of black/ white board or overhead projector for delivery. It is (at least theoretically) expected that each student will attend all of these lectures and tutorials from beginning to the end, and be attentive all the time during the session. However, in practice, this need not be the case. The author has come across dedicated students who for one reason or another had to come in late, had to leave early, or for reasons of fatigue or otherwise, missed the essence of the session. A partial solution is to make the slides available for download, however, at times, there is implicit knowledge being communicated by the teacher during the lecture that is not always made explicit in any form. In such cases or even otherwise, students could benefit from their peers. There are Social Web applications such as mynoteIT and NoteMesh that allows students in the same courses to share notes with each other

480

as well as edit each others notes. The motto of NoteMesh is ‘collaborate to graduate.’ Folksonomy For the reason of time constraints or otherwise, the introduction of a topic during a lecture or tutorial is often relatively ‘local’ and from a single viewpoint. However, during assignments or tests, the students are expected to see the ‘big picture.’ Using the notion of folksonomy or social tagging (Reichel & Kohlhase, 2008; Smith, 2008), the students could associate other relationships with the lecture as they see fit. For example, phrases from past lecture(s) or the textbook could be candidates for tags. A collection of tags can lead to the formation of a tag cloud. A tag cloud is set of related tags with associated weights that represent frequency of use of each tag. The frequency of use of each tag within a tag cloud is illustrated by visual cues such as distinct font color and size. Figure 1 shows a tag cloud for responsibility-driven design (RDD) (Wirfs-Brock & McKean, 2003), a behaviordriven approach to object-oriented design.

A Social Web Perspective of Software Engineering Education

It should be noted that folksonomy (as opposed to taxonomy) is an uncontrolled vocabulary, and the lack of terminological control can have linguistic implications due to synonymy, homonymy, and polysemy. It is also not automatic that all tags that are created by the students may be relevant to the context. For example, unless defined explicitly, a tag labeled ‘framework’ or even ‘RDD’ may not be related to RDD as depicted in Figure 1. Syndication Every so often a teacher needs to keep the students informed of the latest developments, including critical announcements, related to the course. However, individually informing each student of the developments via instant messaging or otherwise is inconvenient for a teacher; arbitrarily visiting the course Web Site is somewhat unsystematic and time consuming for a student. The subscription to periodically refreshable news feeds via syndication helps ameliorate this issue. A practical realization of syndication is a type of metadata implemented in form of channels that the students can subscribe to. There are a variety of syndication technologies of which Really Simple Syndication (RSS) and Atom are beginning to find broad support in conventional user agents and news feed readers. For example, the following RSS markup fragment represents news for a specific day from a single channel:



Object-Oriented Design Education Channel http://www.see.ca// link>

This is a channel for news on developments related to OOD.

News for January 30, 2008 http://www.see. ca/2008/01/30//link>

In a recent interview with Software Engineering Radio, Rebecca Wirfs-Brock highlights the significance of roles and responsibilities ...



The fragment could, for instance, be stored in a file named ood_education.rss and linked from a place that channel readers could readily discover.

Collaborative Researching The Social Web can be an indispensable source for students researching for information for assignments, or during the realization of a software project. Searching has traditionally been one of the most common ways of researching. There are Social Web search engines (such as Similarr and Xoost) that enrich the searching experience by allowing, for example, to communicate with others who are using the same query string during searching or to annotate the search results. Bookmarking has traditionally been one of the most common ways of remembering the resources of interest visited while browsing on the Web. However, these bookmarks reside on the user’s computer and are not accessible by other devices (and therefore are not shareable). Social bookmarking goes beyond traditional bookmarking, and enables management (for example, storage, organization, search, and sharing) of bookmarks

481

A Social Web Perspective of Software Engineering Education

residing remotely at third-party services. There are several social bookmarking services in use today including del.icio.us, Google Bookmarks, and vi.sualize.us that allow bookmarking textual and non-textual information. By unifying their knowledge base, social bookmarking can help both teachers and students to collaborate and share their links to resources. There are Social Web applications (like Google Notebook and Microsoft OneNote) that allow one to attach notes to and clip text, graphics, and links during researching. These ‘notebooks’ can be saved, and can subsequently be used for collaboration and sharing with others. Furthermore, the ‘notebooks’ in Google Notebook can be exported to Google Docs.

Social Scheduling To be able to communicate in person is a critical component in SEE. For example, a team working on a software project has to often schedule a face-to-face meeting with each other or with the teacher. In general, it can be difficult to manage a schedule that is agreeable and flexible to all. Furthermore, seeking consensus can become increasingly difficult as the number of persons involved increases. The use of Social Web applications that facilitate calendar sharing (such as the Google Calendar or Jiffle) can reduce some of the tedium involved in scheduling a meeting agenda. These applications move the focus away from one person (say, meeting chair) being in-charge of gauging others’ preferences via several bi-directional contacts to each person interacting with the calendar to seek an optimal solution. Furthermore, these applications offer other conveniences such as being reachable at any time of a day, access to the latest schedule, privacy by access to restricted/registered users, and so on.

482

Brainstorming In a collaborative approach to discussing an assignment or the details of a software project, students often engage in brainstorming. One way to brainstorm is through visualization, and mind mapping is a graphically-oriented approach to realize it. A mind map is a diagram that represents goals, tasks, or other concepts linked to and arranged radially around a central theme or an idea. It is used to generate, visualize, and organize ideas, and as an aid to understanding, problem solving, and decision making. The students can share these mind maps over the Web and, depending on the permissions, read and/or edit others’ maps. There are Social Web applications such as bubbl.us, comapping, and Mindomo that enable collaborative mind mapping. Figure 2 illustrates a snapshot in time (work in progress) of a mind map using bubbl.us. In it, three students, namely S1, S2, and S3 are in a brainstorming session on the feasibility of a proposed implementation language. The ‘bubbles’ reflect respective inputs by students.

Collaborative Modeling and Prototyping The activities of conceptual modeling (Cowling, 2005) and prototyping are critical to the success of software projects and becoming increasing common in SEE. They are also seldom carried out in isolation (Kamthan, 2008c). There are Social Web applications such as Gliffy and Protonotes that support collaborative modeling and prototyping. Figure 3 illustrates a domain model in Gliffy. Gliffy allows sharing of diagrams, supports version control, and appears to have a low learning curve. However, there are certain challenges to be overcome in collaborative diagramming. For example, the intentions of the geographically dispersed collaborating authors may not be the same and this can lead to interference that may not be easily communicated (Campbell, 2004).

A Social Web Perspective of Software Engineering Education

Figure 2. An example of a partial mind map reflecting a brainstorming session on the viability of an implementation language

The collaborative diagramming tools available over the Web also have to reach the expected level of maturity compared to their desktop counterparts. For example, Gliffy has limited capabilities compared to Microsoft Visio or IBM Rational Rose XDE. The support for the Unified Modeling Language (UML) (Booch, Jacobson, & Rumbaugh, 2005), a standard language for modeling object-oriented software systems, is partial. Therefore, certain desirable constraints

on relationships can not be expressed in Figure 3. It is also not possible for users to extend a given template.

Collaborative Authoring The Social Web presents a suitable environment for collaborative authoring of software process artifacts using various means including Google Docs and Wiki.

Figure 3. The construction in progress of a domain model for a file system showing three concepts and three relationships

483

A Social Web Perspective of Software Engineering Education

Google Docs is a Social Web application that provides capability to create word processing documents, spreadsheets, and presentations, as well as there import and export in various commonly-used formats. It also allows real-time collaboration and sharing of these resources using the Web. However, Google Docs has yet to completely replace a conventional office suite. The support is limited to certain user agents and there are currently physical limits on files sizes and designated storage space that may be constraining. The concept of Wiki was invented in the mid-1990s as a group communication utility. It allowed open editing of information as well as the organization of the contributions and, with various enhancements, continues to serve well in that vein. There are several, opens source flavors of Wiki available today addressing different target groups and organizational needs. Most flavors of Wiki, including MediaWiki and TinyWiki, can be easily acquired, installed, and administered under commonly-deployed computing platforms (Ebersbach, Glaser, & Heigl, 2006). In a Wiki, it is possible to present information in a heterogeneous combination of multiple modalities. For example, a document under Wiki could include text, graphics, mathematical symbols, and/or animations. There are various uses of a Wiki in a software engineering course. A Wiki can be used both by teachers in a course and by students for a course project. Teacher Uses of Wiki A teacher could administer a Wiki for a course (namely, the “Home Page” along with related resources) and could also invite students to provide open feedback (anonymous or otherwise) on the progression of the course. The accumulated feedback can be useful in a course retrospective and may lead to improvements in teaching. Even if administering a Wiki is not an option, course material pertaining to the lectures could

484

be placed on the Web and Wiki could be used to supplement them. For example, as shown in Figure 4, key topics and terms in a classroom lesson could point to resources from the projects of the Wikipedia Foundation (such as Wikibooks, Wikipedia, Wikisource, Wikiversity, Wiktionary, and so on). This example can be extended to a convergence and harmonic use of multiple Social Web applications. For instance, the elements of requirements specification and test documentation using one Social Web application (say, Wiki) could point to the use case model using one Social Web application (say, Gliffy). The previous example enables teachers to demonstrate to the students that the discipline of software engineering is not ‘closed’ (in form of books) and belongs to an ecosystem that has a social aspect outside the classroom. Indeed, this supports the notion of open e-books based specifically on Wiki (Ravid, Kalman, & Rafaeli, 2008). Furthermore, such a use somewhat alleviates the burden of developing every aspect of the course material from scratch and opens the potential for staying current. Finally, it also allows teachers to present the course material at different levels of abstraction. For example, a single document D under Wiki may include information at a high level of abstraction and the details of relevant concepts therein could be available in external resources {R1,…,Rn} linked from D. Student Uses of Wiki A team of students could run its own Wiki specific to course project and limited to its members (and perhaps to the teacher and to the teaching assistants). This has numerous benefits. For example, it has been suggested that requirements specifications must evolve from their paper-based legacy (Firesmith, 2003). The deployment of Social Web technologies/applications is one way to realize that. Indeed, Wiki has been used for managing use cases (Decker et al., 2006), requirements negotiation (Damian et al., 2000), and for the evolution

A Social Web Perspective of Software Engineering Education

Figure 4. A classroom lesson on object-oriented design metrics using external resources in the Wiki environment

of the requirements specification (Decker et al., 2007). A common problem in software projects is that the details of actually carrying out the process often remain suppressed. However, this implicit knowledge is critical for future projects, particularly those in the same or a similar domain. The versioning and feedback information in software documents based on a Wiki helps make some of the experience of the team members in carrying out the process explicit. This experience, including triumphs and tribulations, is relevant to social constructivism and can be useful in a project retrospective. Finally, teacher feedback on postings by students is one way to realize scaffolding. The disadvantages of a Wiki are that student participation is not guaranteed (Cole, 2009) and, while some students may be extrovert prolific writers, others may not. The assessment of individual, original work becomes ever more challenging (Graya et al., 2008). Wikis are also known for ‘noise’ (impertinent information), ‘casual writing’ (presence of phonetic and 1337 style of writing), and ‘editing wars’ (endless, multi-directional

debates). These, however, can be attributed to human usage rather than to inherent limitations of the underlying technology/application. Therefore, some form of monitoring and control perhaps initiated by the teacher is essential.

Collaborative Presentations In a course project, it is often the case that a team has to give presentations as part of the project. For example, there can be presentations during the semester to report on the status of the project and/or individual deliverables and at the end of the project to report on the product. Since the project itself is a result of collaborative effort, the author recommends to students that the presentations also be prepared and delivered in collaboration. There are Social Web services such as SciVee, SlideShare, and 280 Slides that facilitate preparing, storing, sharing, presenting, and exporting slides to other formats. Using hypertext, these slides can point to other artifacts that can be spawned during presentation.

485

A Social Web Perspective of Software Engineering Education

Course Feedback and Project Retrospectives

and therefore, participation in blogging is not guaranteed a priori.

A blog provides an outlet for anybody to publicly express opinion, in writing, on just about any subject. In general, blogs are published on the Web in a chronological order. A blog is not necessarily isolated: in a side bar of a personal blog, the author of a blog, namely the blogger, can provide a list of other blogs (or a blogroll). The universe of all blogs along with their relationships has been termed as blogspace or blogosphere. There are a number of blogging services on the Web including Blogger and Blogspot. There are a few synergistic benefits of blogging for the teachers and the students. To teachers, blogging lends an opportunity to respond to students’ concerns in an informal environment, inform them of their scholarly activities (like upcoming presentations or publications), or state their position on a topic. This is especially crucial for a subjective discipline like software engineering. To students, blogging presents an opportunity to ask questions, and make their agreements and disagreements known in a public forum, both during and after the course. This could be used by teachers to improve the course. Blogs can also be used by members of a software project team as a daily ‘diary’ in which to document their experience about the project that would otherwise not appear in process artifacts. This personal experience can at times be emotional (Kim, 2008) highlighting a variety of human aspects including triumph and tribulation, and elation and frustration. In carrying out a software project, the goal is not only to successfully complete the project but also to learn how to successfully complete projects. A blogroll pertaining to a single team can be invaluable while conducting a project retrospective. However, there are side-effects of blogging. For example, the personality traits of those involved in blogging can be significant (Guadagno, Okdie, & Eno, 2008; Leslie & Murphy, 2008)

Software Engineering is ‘Living’

486

The fact that software engineering is still a young, constantly evolving discipline is not always evident, particularly to undergraduate students. The textbooks, once published, attempt to merely communicate static knowledge. The teachers who wish to communicate new developments to students are faced with the dilemma of keeping the focus on the core software engineering knowledge while still being able to allude to new developments in the field that are within the scope of the curriculum. Utilities such as Google Trends could be used (Rech, 2007a) by teachers to convey some of the dynamism and excitement as well as help form a ‘bridge’ between academia and the real-world, and thereby demonstrate to the students that the discipline of software engineering is indeed ‘alive and well’. Google Trends analyzes a portion of Google searches to compute the number of searches for a term, relative to the total number of searches done on Google over time. It then shows a search-volume graph with the results plotted on a linear scale. As an example, based on Google Trends (with the query string q=Object-Oriented+Design), Figure 5 shows that in the last few years, the adoption of object-oriented design has been on the rise around the world. However, it should be noted that Google Trends is limited by the nature of query formulations expressed in a natural language. For example, searching for ‘UML’ on Google Trends yield several irrelevant results, and for the ‘Model-Driven Software Development’ (for which the data set was low) there were no results at all. It also turns out that this behavior is not unique to aforementioned terms.

A Social Web Perspective of Software Engineering Education

Figure 5. A snapshot of the adoption of object-oriented design

guidelines for Adoption of social web technologies/Applications in software engineering education It is known that teachers can face various obstacles in enabling an environment for elearning at Universities (Mahdizadeh, Biemans, & Mulder, 2008). The following are a set of guidelines that may help prospective teachers in making an informed decision towards integrating Social Web technologies/applications in software engineering-related courses and hopefully build an amenable ‘culture’ that embraces them:

Guidelines Related to Teachers It is needless to say that the teachers must be aware of the current policies of the institution pertaining to (1) legal and privacy issues regarding students, and (2) security issues regarding information and computing infrastructure. For the sake of increasing potential long-term support, it can also be useful to keep the administration (such as the department Chair or the Dean) abreast of any new endeavors, and periodically inform them of successes and failures.

The educational institutions of the future need the teachers of the future. It is evident that a teacher’s understanding of the relevant Social Web technologies/applications for realizing successful collaborations in an SEE context is essential. To improve self-efficacy, the teachers may need to avail themselves of training sessions pertaining to the technical aspects of the Social Web. The Social Web experience of other teachers in the past, at the same institution or otherwise, including that is anecdotal (Cole, 2009), may be useful.

Guidelines Related to Students The students should be considered as ‘first-class’ participants in any integration efforts. In particular, they should be (1) informed of any ‘social experiments’ being pursued as part of the course, (2) made aware of their rights and responsibilities that comes with flexibility of Social Web technologies/applications, and (3) introduced to the ethical issues in software engineering (Kamthan, 2008a) before they embark on a software project. Then, creative uses of the Social Web (Lazzari, 2009) that require engagement and active participation of students is a possible direction.

487

A Social Web Perspective of Software Engineering Education

Guidelines Related to Infrastructure The selection and adoption of Social Web technologies/applications should be based on objectively verifiable needs centered on software engineering knowledge and students, rather than technological determinism. This introduction of Social Web technologies/applications does not have to be a matter of ‘all or nothing’; indeed, they could be introduced progressively. This also lowers the burden on teacher training of appropriate technologies/applications. It may be useful to initially (1) introduce technologies/applications in activities that can benefit most, (2) select technologies/applications that originate from authoritative sources, are relatively stable, and in which the teacher does not have to relinquish control completely, and (3) the presence of technologies/applications is not apparent to students (that is, technologies/applications are transparent to and do not interfere with the learning goals). Then, based on a retrospective, the use of technologies/applications can be scaled appropriately.

fUtUre reseArCh direCtions It is still early to predict the outcome of the Social Web phenomenon in general and its impact on SEE in particular. The work presented in this chapter can be extended in a few different directions, which are briefly discussed next.

evaluating the effectiveness of sw4se2

technologies/applications. It could also be useful to elicit extensions of pedagogically-inclined patterns specific to the use of IT in education in general and SEE in particular. The introduction of a new technology/application in education is susceptible to indirections by virtue of dependencies. In particular, there are issues of affordances. If the normative copy of an artifact is stored remotely, then unavailability at the time of need or, in the worst case, decommissioning, of the Social Web application can be a major issue. A sustained integration of Social Web technologies/applications in SEE needs to address quality-related concerns that can arise. The maxim ‘users add value’ (or ‘user generated content’) commonly associated with the Social Web has its negative side-effects. In particular, an assessment of the impact on the credibility of information emanating from relaxation of control (from teacher’s viewpoint) and emergence of privacy issues (from student’s viewpoint) is of research interest. The hardware and software requirements on both server-side and client-side of ‘Rich Internet Applications’ that form a large subset of Social Web applications can be resource-intensive. For example, the mashups in which aggregation of information takes place on the client-side expect hardware and software capabilities that a consumer may not have or the file sizes of podcasts that are not streamed but are available only as download could be prohibitive to those on low bandwidth. To that regard, an investigation into associated cost estimation is also of interest.

extensions of sw4se2 Although certain benefits of the Social Web are evident, further evaluation based on long-term teaching experience followed by surveys is indispensable. It would be useful to distill this experience and subsequently present it in form of ‘best practices.’ For example, it could be useful to construct a mapping between pedagogicallyinclined patterns and corresponding Social Web

488

It is evident that SW4SE2 will change as both software engineering and Social Web evolve. For example, Web Services are important to a number of Social Web applications such as mashups, and their impact on the SEE curricula in general and SW4SE2 in particular could also be worth investigating. Similarly, the intersection of the

A Social Web Perspective of Software Engineering Education

Semantic Web and the Social Web (Lassila & Hendler, 2007; Shadbolt, Hall, & Berners-Lee, 2006) brings new dynamics. A study of the impact of such a confluence on SEE in the areas such as domain modeling and requirements elicitation is also of interest. Finally, devising a mapping between software engineering knowledge areas, pedagogical patterns realizable in SEE, and Social Web technologies/applications applicable to SEE is of interest. It is evident that such a mapping would be based upon real-world experiences and lessons learnt from the practice of SW4SE2. It is likely that this mapping would initially change as the Social Web takes shape. However, once established, such a mapping could be a valuable contribution to the successful practices of SEE.

the human Aspect of sw4se2 It is evident that age, background, and preferences of students vary with respect to learning software engineering and the use of Social Web technologies/applications. For example, the students who have grown up with the use of Web may be more inclined and receptive to the use of Social Web in SEE. In the context of global software development, it has been experienced (Gotel et al., 2007) that not all students may have the same exposure to Social Web technologies/applications and as a consequence or otherwise may perceive such technologies/applications to be peripheral rather than essential. In such cases, the significance of communication may need to be reinforced, perhaps with examples of software project successes and failures due to communication. Therefore, an investigation into the connection between students and technologies/applications would be of interest. Similarly, for an optimal delivery of software engineering courses, the issue of teacher training also needs to be addressed.

ConClUsion The challenges facing the practice of software engineering today are as much technical as they are social in nature. The social and organic aspects of software engineering not only need to be acknowledged by teachers but also made explicit to the students in a feasible manner. The Social Web has the potential to revitalize SEE. It can open new vistas for teachers as well as students in a number of ways. In particular, it provides an avenue for both teachers and students to communicate, interact, and experiment, both inside and outside the classroom. In particular, the Social Web lends a unique opportunity to computer science and software engineering students that are accustomed to using the Web (for non-necessarily academically related activities) in their daily lives. By participating in the Social Web, the students become co-producers of information and ‘first-class’ participants in-charge of their own learning. Moreover, with appropriate software project, they can even become innovators of Social Web applications. Thus, students can not only benefit from the Social Web by being a participant but can also help create applications/ technologies that can benefit others in the future, and an educational setting provides a starting point towards that goal. The experience could stimulate learning, and may even be fun. In conclusion, an adoption of the Social Web can be rewarding but may require a re-examination of the current software engineering culture at an institution, both at administrative and at the educational level. An introduction of Social Web technologies/applications in SEE can be disruptive as it requires changes at both logistical and pedagogical levels and, in some cases, radical departure from conventional approaches to which not all may be a willing participant (Collis & Moonen, 2008). Furthermore, as with any other investment, for a long-term sustainability of the integrating Social Web technologies/applications in SEE, the benefits must be kept in perspective

489

A Social Web Perspective of Software Engineering Education

alongside the associated costs to both teachers and students.

referenCes Aiken, J. (2004). Technical and human perspectives on pair programming. ACM SIGSOFT Software Engineering Notes, 29(5). Anderson, T. (2008). The theory and practice of online learning (second edition). Athabasca University Press. Bennedsen, J., & Eriksen, O. (2006). Categorizing pedagogical patterns by teaching activities and pedagogical values. Computer Science Education, 16(2), 157–172. doi:10.1080/08993400600768091 Bergin, J. (2003, June 25-29). Two pedagogical patterns for course design. The Eighth European Conference on Pattern Languages of Programs (EuroPloP 2003), Irsee, Germany. Bergin, J. (2006, October 21-23). Active learning and feedback patterns. The Thirteenth Conference on Pattern Languages of Programs (PLoP 2006), Portland, OR. Bernsteiner, R., Ostermann, H., & Staudinger, R. (2008). Facilitating e-learning with social software: Attitudes and usage from the student’s point of view. International Journal of WebBased Learning and Teaching Technologies, 3(3), 16–33. Bompani, L., Ciancarini, P., & Vitali, F. (2002). XML-based hypertext functionalities for software engineering. Annals of Software Engineering, 13(14), 231–247. doi:10.1023/A:1016553628389 Booch, G., Jacobson, I., & Rumbaugh, J. (2005). The unified modeling language reference manual (second edition). Addison-Wesley.

490

Buschmann, F., Henney, K., & Schmidt, D. C. (2007). Pattern-oriented software architecture, volume 5: On patterns and pattern languages. John Wiley and Sons. Campbell, J. D. (2004). Interaction in collaborative computer supported diagram development. Computers in Human Behavior, 20(2), 289–310. doi:10.1016/j.chb.2003.10.019 Chatti, M. A., Klamma, R., Jarke, M., & Naeve, A. (2007, July 18-20). The Web 2.0 driven SECI model based learning process. The Seventh IEEE International Conference on Advanced Learning Technologies (ICALT 2007) (pp. 780-782), Niigata, Japan. Chong, J., & Hurlbutt, T. (2007, May 19-27). The social dynamics of pair programming. The Twenty Ninth International Conference on Software Engineering (ICSE 2007), Minneapolis. Cole, M. (2009). Using wiki technology to support student engagement: Lessons from the trenches. Computers & Education, 52(1), 141–146. doi:10.1016/j.compedu.2008.07.003 Collis, B., & Moonen, J. (2008). Web 2.0 tools and processes in higher education: Quality perspectives. Educational Media International, 45(2), 93–106. doi:10.1080/09523980802107179 Coplien, J. O., & Harrison, N. B. (2005). Organizational patterns of agile software development. Prentice-Hall. Cormode, G., & Krishnamurthy, B. (2008). Key differences between Web 1.0 and Web 2.0. First Monday, 13(6). Cowling, A. J. (2005). The role of modelling in the software engineering curriculum. Journal of Systems and Software, 75(1-2), 41–53. doi:10.1016/j. jss.2004.02.021

A Social Web Perspective of Software Engineering Education

Cronjé, J. (2006). Paradigms regained: Toward integrating objectivism and constructivism in instructional design and the learning sciences. Journal Educational Technology Research and Development, 54(4), 387–416. doi:10.1007/ s11423-006-9605-1 Damian, D. E. H., Eberlein, A., Shaw, M. L. G., & Gaines, B. R. (2000). Using different communication media in requirements negotiation. IEEE Software, 17(3), 28–36. doi:10.1109/52.896247 Decker, B., Ras, E., Rech, J., Jaubert, P., & Rieth, M. (2007). Wiki-based stakeholder participation in requirements engineering. IEEE Software, 24(2), 28–35. doi:10.1109/MS.2007.60 Decker, B., Ras, E., Rech, J., Klein, B., & Hoecht, C. (2006, March 27-28). Using wikis to manage use cases: Experience and outlook. Workshop on Learning Software Organizations and Requirements Engineering (LSO+RE 2006), Hannover, Germany. Ebersbach, A., Glaser, M., & Heigl, R. (2006). Wiki: Web collaboration. Springer-Verlag. Eckstein, J., Bergin, J., & Sharp, H. (2002a, July 3-7). Feedback patterns. The Seventh European Conference on Pattern Languages of Programs (EuroPLoP 2002), Irsee, Germany. Eckstein, J., Bergin, J., & Sharp, H. (2002b, September 8-12). Patterns for active learning. The Ninth Conference on Pattern Language of Programs (PLoP 2002), Monticello. Eckstein, J., Manns, M. L., Sharp, H., & Sipos, M. (2003, June 25-29). Patterns for teaching effectively. The Eighth European Conference on Pattern Languages of Programs (EuroPloP 2003), Irsee, Germany. Elssamadisy, A., & West, D. (2006, October 21-23). Patterns of agile practice adoption: A beginning. The Thirteenth Conference on Pattern Languages of Programs (PLoP 2006), Portland, OR.

Fincher, S. (1999). Analysis of design: An exploration of patterns and pattern languages for pedagogy. Journal of Computers in Mathematics and Science Teaching, 18(3), 331–348. Firesmith, D. G. (2003). Modern requirements specification. Journal of Object Technology, 2(2), 53–64. Frailey, D. J. (1998). Opportunities for software engineering education. Annals of Software Engineering, 6(1-4), 131–144. doi:10.1023/A:1018957315109 Fuller, U., Johnson, C. G., Ahoniemi, T., Cukierman, D., Hernán-Losada, I., & Jackova, J. (2007). Developing a computer science-specific learning taxonomy. ACM SIGCSE Bulletin, 39(4), 152–170. doi:10.1145/1345375.1345438 Gagné, R. M., Wager, W. W., Golas, K., & Keller, J. M. (2005). Principles of instructional design (fifth edition). Wadsworth. Gotel, O., Kulkarni, V., Neak, L. C., & Scharff, C. (2007, October 21). The role of wiki technology in student global software development: Are all students ready? Wikis for Software Engineering Workshop (Wikis4SE 2007), Montreal, Canada. Graya, K., Thompson, C., Clerehan, R., Sheard, J., & Hamilton, M. (2008). Web 2.0 authorship: issues of referencing and citation for academic integrity. The Internet and Higher Education, 11(2), 112–118. doi:10.1016/j.iheduc.2008.03.001 Guadagno, R. E., Okdie, B. M., & Eno, C. A. (2008). Who blogs? Personality predictors of blogging. Computers in Human Behavior, 24(5), 1993–2004. doi:10.1016/j.chb.2007.09.001 Hadjerrouit, S. (2005). Constructivism as guiding philosophy for software engineering education. ACM SIGCSE Bulletin, 37(4), 45–49. doi:10.1145/1113847.1113875

491

A Social Web Perspective of Software Engineering Education

Hayes, D., Hill, J., Mannette-Wright, A., & Wong, H. (2006, October 21-23). Team project patterns for college students. The Thirteenth Conference on Pattern Languages of Programs (PLoP 2006), Portland, OR. Hsu, J. (2008). Innovative technologies for education and learning: Education and knowledgeoriented applications of blogs, wikis, podcasts, and more. International Journal of Web-Based Learning and Teaching Technologies, 3(3), 62–81. Kamthan, P. (1999, March 7). Java applets in education. Internet Related Technologies. Kamthan, P. (2008a). Ethics in software engineering. In M. Quigley (Ed.), Encyclopedia of information ethics and security (pp. 266-272). Hershey, PA: IGI Global. Kamthan, P. (2008b). A methodology for integrating information technology in software engineering education. In R. Donnelly & F. McSweeney (Eds.), Applied E-Learning and E-Teaching in Higher Education (pp. 225-243). Hershey, PA: IGI Global. Kamthan, P. (2008c). Pair modeling. In G. D. Putnik & M. M. Cunha (Eds.), Encyclopedia of networked and virtual organizations (pp. 11711178). Hershey, PA: IGI Global. Kamthan, P. (2009). A Methodology for Integrating the Social Web Environment in Software Engineering Education. International Journal of Information and Communication Technology Education, 5(2), 21–35. Keller, J. M., & Suzuki, K. (1988). Use of the ARCS motivation model in courseware design. In D. H. Jonassen (Ed.), Instructional designs for microcomputer courseware. Lawrence Erlbaum. Kim, H. N. (2008). The phenomenon of blogs and theoretical model of blog use in educational contexts. Computers & Education, 51(3), 1342–1352. doi:10.1016/j.compedu.2007.12.005

492

Kleinberg, J. (2008). The Convergence of Social and Technological Networks. Communications of the ACM, 51(11), 66–72. doi:10.1145/1400214.1400232 Lassila, O., & Hendler, J. (2007). Embracing Web 3.0. IEEE Internet Computing, 11(3), 90–93. doi:10.1109/MIC.2007.52 Layman, L., Williams, L., Osborne, J., Berenson, S., Slaten, K., & Vouk, M. (2005, October 19-22). How and why collaborative software development impacts the software engineering course. The Thirty Fifth Annual Conference on Frontiers in Education (FIE 2005), Indianapolis, IN. Lazzari, M. (2009). Creative use of podcasting in higher education and its effect on competitive agency. Computers & Education, 52(1), 27–34. doi:10.1016/j.compedu.2008.06.002 Leslie, P., & Murphy, E. (2008). Post-secondary students’ purposes for blogging. International Review of Research in Open and Distance Learning, 9(3). Lethbridge, T. C. (2000). What knowledge is important to a software engineer? Computer, 33(5), 44–50. doi:10.1109/2.841783 Lieberman, B. A. (2007). The art of software modeling. Auerbach Publications. Macaulay, L. (1993, January 4-6). Requirements capture as a cooperative activity. The First IEEE International Symposium on Requirements Engineering, San Diego, CA. Mahdizadeh, H., Biemans, H., & Mulder, M. (2008). Determining factors of the use of e-learning environments by university teachers. Computers & Education, 51(1), 142–154. doi:10.1016/j. compedu.2007.04.004 Naruse, M., Takada, Y., Yumura, Y., Wakamatsu, K., & Iba, T. (2008, October 18-20). Project patterns: A pattern language for promoting project. The Fifteenth Conference on Pattern Languages of Programs (PLoP 2008), Nashville, TN.

A Social Web Perspective of Software Engineering Education

Negash, S., & Wilcox, M. V. (2008). E-learning classifications: Differences and similarities. In S. Negash, M. E. Whitman, A. B. Woszczynski, K. Hoganson & H. Mattord (Eds.), Handbook of distance learning for real-time and asynchronous information technology education. Hershey, PA: IGI Global. O’Reilly, T. (2005, September 30). What is Web 2.0: Design patterns and business models for the next generation of software. O’Reilly Network. Palfrey, J., & Gasser, U. (2008). Born digital: Understanding the first generation of digital natives. Basic Books. Parker, K. R., & Chao, J. T. (2007). Wiki as a teaching tool. Interdisciplinary Journal of Knowledge and Learning Objects, 3, 57–72. Ravid, G., Kalman, Y. M., & Rafaeli, S. (2008). Wikibooks in higher education empowerment through online distributed collaboration. Computers in Human Behavior, 24(5), 1913–1928. doi:10.1016/j.chb.2008.02.010 Rech, J. (2007a). Discovering trends in software engineering with Google trend. ACM SIGSOFT Software Engineering Notes, 32(2). Reichel, M., & Kohlhase, A. (2008). Embodied conceptualizations: Social tagging and e-learning. International Journal of Web-Based Learning and Teaching Technologies, 3(1), 58–67. Reigeluth, C. M., & Stein, F. S. (1983). The elaboration theory of instruction. In C. M. Reigeluth (Ed.), Instructional design theories and models: An overview of their current status (pp. 335-382). Erlbaum Associates. Rezaei, S. (2005, May 5-6). Software engineering education in Canada. The Western Canadian Conference on Computing Education (WCCCE 2005), Prince George, Canada.

Schmolitzky, A. (2007, July 4-8). Patterns for teaching software in classroom. The Twelfth European Conference on Pattern Languages of Programs (EuroPLoP 2007), Irsee, Germany. Schümmer, T., & Lukosch, S. (2007). Patterns for computer-mediated interaction. John Wiley and Sons. Shadbolt, N., Hall, W., & Berners-Lee, T. (2006). The Semantic Web revisited. IEEE Intelligent Systems, 21(3), 96–101. doi:10.1109/MIS.2006.62 Shank, P. (2008). Web 2.0 and beyond: The changing needs of learners, new tools, and ways to learn. In S. Carliner & P. Shank (Eds.), The e-learning handbook: Past promises, present challenges. John Wiley and Sons. Shaw, M. (2000, June 4-11). Software engineering education: A roadmap. The Twenty Second International Conference on Software Engineering (ICSE 2000), Limerick, Ireland. Shih, W.-C., Tseng, S.-S., & Yang, C.-T. (2008). Wiki-based rapid prototyping for teachingmaterial design in e-learning grids. Computers & Education, 51(3), 1037–1057. doi:10.1016/j. compedu.2007.10.007 Shuen, A. (2008). Web 2.0: A strategy guide. O’Reilly Media. Sigala, M. (2007). Integrating Web 2.0 in e-learning environments: A socio-technical approach. International Journal of Knowledge and Learning, 3(6), 628–648. doi:10.1504/IJKL.2007.016837 Smith, G. (2008). Tagging: People-powered metadata for the social Web. New Riders. Smith, P., & Ragan, T. J. (1999). Instructional design (second edition). John Wiley and Sons. Staron, M. (2007, May 19-27). Using experiments in software engineering as an auxiliary tool for teaching: A qualitative evaluation from the perspective of students’ learning process. The Twenty Ninth International Conference on Software Engineering (ICSE 2007), Minneapolis.

493

A Social Web Perspective of Software Engineering Education

Surakka, S. (2007). What subjects and skills are important for software developers? Communications of the ACM, 50(1), 73–78. doi:10.1145/1188913.1188920 Tenenberg, J. (2008). An institutional analysis of software teams. International Journal of HumanComputer Studies, 66(7), 484–494. doi:10.1016/j. ijhcs.2007.08.002 Williams, L. (2000). The collaborative software process. Unpublished doctoral dissertation, Department of Computer Science, The University of Utah. Wirfs-Brock, R., & McKean, A. (2003). Object design: Roles, responsibilities, and collaborations. Addison-Wesley.

The following publications discuss the issues arising in the practice of collaborative software engineering education for global software development:Favela, J., & Peña-Mora, F. (2001). An Experience in Collaborative Software Engineering Education. IEEE Software, 18(2), 47–53. doi:10.1109/52.914742 Gotel, O., Kulkarni, V., Neak, L. C., & Scharff, C. (2007). The Role of Wiki Technology in Student Global Software Development: Are All Students Ready? Wikis for Software Engineering Workshop (Wikis4SE 2007). Montreal, Canada. October 21, 2007. The following publications provide a perspective on the prospects and concerns of Social Web applications from the viewpoint of end-user software engineering:

AdditionAl reAding The following publications introduce the notion of a pattern in the domain of urban architecture and planning, and paved the way to the introduction of patterns in other areas including pedagogy and in software engineering: Alexander, C. (1979). The Timeless Way of Building. Oxford University Press. Alexander, C., Ishikawa, S., & Silverstein, M. (1977). A Pattern Language: Towns, Buildings, Construction. Oxford University Press. Anslow, C., & Riehle, D. (2008). Towards EndUser Programming with Wikis. The Fourth Workshop in End-User Software Engineering (WEUSE IV), Leipzig, Germany, May 12, 2008. Costabile, M. F., Mussio, P., Provenza, L. P., & Piccinno, A. (2008). End Users as Unwitting Software Developers. The Fourth Workshop in End-User Software Engineering (WEUSE IV), Leipzig, Germany, May 12, 2008.

494

Key terMs And definitions Constructivism: A theory of learning that views learning as a process in which the learner actively constructs or builds new ideas or concepts based upon current and past knowledge. It is based on the premise that learning involves constructing one’s own knowledge from one’s own experiences. Information Technology: A technology for activities related to information, such as acquisition, creation, communication, dissemination, processing, archival, retrieval, transformation, and so on, within the context of the Internet and the Web. Objectivism: A theory of learning that views knowledge as some entity existing independent of the mind of individuals. The goal of instruction is to communicate or transfer knowledge to learners in the most effective manner possible. Open Source Software: A single encompassing term for software that satisfies the following conditions: (1) non-time-delimited, complete

A Social Web Perspective of Software Engineering Education

software whose source is publicly available for (re) distribution without cost to the user, (2) imposes minimal, non-restrictive licensing conditions, and (3) is itself either based on non-proprietary technologies or based on proprietary technologies that conform to (1) and (2). Scaffolding: A teaching strategy in which the teacher takes upon a passive role and provides only the basic transient support towards the learning techniques deployed with the goal that the students take responsibility of their own learning. Software Engineering: A discipline that advocates a systematic approach of developing

high-quality software on a large-scale while taking into account the factors of sustainability and longevity, as well as, organizational constraints of resources. Software Process: A set of activities, methods, and transformations that are used to develop and maintain software and its associated products. Web 2.0: A set of economic, social, and technological trends that collectively form the basis for the future Web as a medium characterized by user participation, openness, and network effects.

495

496

Chapter 27

University 2.0: Embracing Social Networking to Better Engage the FacebookGeneration in University Life David Grifin Leeds Metropolitan University, UK

AbstrACt The social networking Web site is one type of Web 2.0 innovation that has been embraced by universityaged young people. The success of Facebook and similar Web sites has prompted universities to explore how they might use social networking Web sites to engage with their students. In this chapter, I argue that universities are misguided in their attempts to use social networking groups to attempt to engage with students registered with the Web sites. I present empirical evidence from a case study university to substantiate this claim. A framework is developed to categorise the university-related Facebook groups and competing theoretical perspectives on diffusion of innovation are employed to analyse the participation in these groups by students. Recommendations are made for universities, and other organisations, intending to use social networking Web sites to engage with students.

introdUCtion “Others now question whether the idea of a Virtual Learning Environment (VLE) … makes sense in the Web 2.0 world. One Humanities lecturer is reported as having said: “I found all my students were looking at the material in the VLE but going straight to Facebook to use the discussion tools and discuss the material and the lectures. I thought I might as well join them and ask them DOI: 10.4018/978-1-60566-384-5.ch027

questions in their preferred space.” (Anderson, 2007, p33).

The social networking site is one type of recent Web 2.0 innovation that has been embraced by universityaged young people. Facebook, for example, has only been in existence since 2004. During this brief period of time its diffusion amongst the young has been rapid. It achieved one million early adopters within its first year of operation. By the end of its second year this had grown to five million users.

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

University 2.0

Within four years, participation in the site had exceeded 50 million active users (Facebook, 2007). The original purpose of the site was to facilitate social networking between classmates and former classmates. The success of Facebook and similar sites has prompted universities (and many other types of organisation) to explore how they might use social networking sites to engage with the millions of members of university age. Will this ‘expansionary’ innovation (Osborne, 1998), using the social networking artefact for different purposes, be successful? In this chapter, I argue that universities are misguided in their attempts to use social networking groups to attempt to engage with students registered with the sites. I present empirical evidence from a case study university to substantiate this claim. The majority of students are active participants in Web 2.0 in general and social networking sites in particular. Universities appear to have adopted a technological-deterministic approach towards social networking sites, assuming that diffusion among their student body will follow the path identified by Rogers (1995). However, here it is suggested that this innovation is socially shaped and its diffusion is better explained using a ‘technology complex’ comprising of hard characteristics, such as the artefact, plus softer aspects, such as the culture of the user group (Fleck and Howells, 2001). Four categories of universityrelated groups are identified on Facebook and the technology complex is utilised to explain the varying success of the diffusion of the innovation in each of the four categories. Based on this analysis, and the results of a survey of student attitudes, I conclude that the softer aspects of the technology complex are likely to inhibit the diffusion of most university-initiated groups on social networking sites. The chapter is organised as follows. First, several perspectives on the diffusion of innovation are introduced. These theoretical frameworks will form the basis of the subsequent discussion of

adoption of social networks in the chapter. Next, the methodology used in the empirical research is presented. Following this, the findings of the case study research are presented and discussed and conclusions drawn. Recommendations are made to university administrators considering using social networking websites and questions are raised concerning the applicability of current diffusion of innovation theory to emerging Web 2.0 channels in which peer production is the predominant economic model.

PersPeCtives on the diffUsion of innovAtion There are two prime theoretical approaches for exploring the diffusion of a technological innovation through a population of social actors: diffusion of innovation (DOI) theory and social shaping theory (Webster, 2007). Rogers, an early proponent of DOI theory, defines innovation as “an idea, practice, or object that is perceived as new by an individual or unit of adoption (1995, p.11).” This definition limits the innovation to the technological artefact. Diffusion then takes place when an innovator introduces this technology to a social group. DOI theory is a technologically-deterministic approach. It is the characteristics of the technology, or to be more precise the artefact itself, that make it useful to its users and these characteristics will determine its ability to be accepted by a population. The diffusion of the innovation through a community takes the form of an S-shaped curve. In the early stages, the innovators and early adopters use the technology, then, at the peak of the S-curve, the majority are using it and, finally, the laggards within the community are persuaded to join in. Eventually, the technology is replaced when a superior technology becomes available. Rogers (1995) does mention sociological aspects that might impede the diffusion of an innovation. Diffusion is likely to be less effective

497

University 2.0

when the change agent and other actors have differing ‘beliefs, education, social status and the like’ (p.19). The norms of the group may affect its rate of adoption. However, the technology is represented as independent from society, an external variable that impacts upon its unit of adoption. A number of criticisms of this approach have been raised in previous studies. Firstly, DOI theory exhibits a pro-innovation bias (McMaster and Wastell, 2005). Secondly, it suggests that diffusion ought to take place, any failure by social actors to adopt the technology is perceived as resistance (McMaster and Wastell, 2005). This is borne out by the classifications used to describe those who are introduced to the technology. The term laggard implies curmudgeonly behaviour by those who are last to make use of the innovation. Thirdly, the theory has a rational bias, assuming that adopters will make rational decisions (Jeyaraj et al, 2006). Previous studies have mainly concentrated on exploring DOI theory in organisational settings in which the adopters have an extrinsic motivation to embrace a new technology. The party introducing the innovation usually has economic power over the passive adopters of the technology. This is not the situation with Web 2.0. The adopters have a range of extrinsic and intrinsic motivations for participating in the websites (Benkler, 2006). Consequently, a different ontological perspective is required for explaining Web 2.0 innovations. Social Shaping theory provides one possible alternative perspective. Social shaping theory rejects the proposition of a cause-and-effect relationship between technology and society. Technology not only impacts upon society, but is itself shaped by the social group which embraces it (MacKenzie and Wajcman, 1985). The relationship between technology and society is much more intricate than that presented by diffusion theorists. Fleck and Howells (2001) have attempted to map the interplay between the two domains by means of a ‘technology complex’. This framework employs a broader definition of

498

technology. It includes the harder elements, such as the artefact, but also takes account of other softer elements such as knowledge, skills and culture. According to Fleck and Howells (2001), failure in the diffusion of specific technologies is a consequence of the tendency to treat the artefact as though it is the sole element of the technology.

reseArCh Methodology This chapter explores the following questions: Q1: How actively are students engaged with Web 2.0 sites in general and social networking sites in particular? Q2: What types of university-related groups have been set up on social networking sites? Q3: What are students’ attitudes to universities using social networking sites to engage with them? These questions will be examined using the case study method. In this chapter, I shall examine the diffusion of Facebook within the groups associated with Leeds Metropolitan University, a large regional university in the North of England. As I shall repeatedly refer to this university, I shall shorten its name to LeedsMet in the following discussion. In this study, the stance is taken that technology is socially-situated, it is bounded not just by the artefact itself, but by many contextual elements. There is not a simple causal relationship between the technology being employed and the organization applying it. Organizational characteristics help to shape the technology (Fleck and Howells, 2001). This stance is consistent with the interpretive tradition (Walsham, 1993). Interpretive methods are useful for understanding the social context in which computer-based systems are exploited and for examining the two-way interaction between this context and the system. The most appropriate method of

University 2.0

Table 1. Issues of validity and reliability of the study (adapted from Yin, 2003) Test

Definition

Tactics employed

Construct validity

Establishing appropriate measures for issues being studied

Several sources of data were used to provide data triangulation at appropriate points; attention has been given in this chapter to making the steps in the development of the argument as transparent as possible.

External validity

Establishing how to generalize from the case study findings

The diffusion of Facebook groups in a single university was studied. No attempt was made to suggest that these findings were representative of other universities or other social networking sites. External validity may be achieved through four types of generalization: theory, implications, concepts and rich insight (Walsham, 1993). Here the external validity was achieved through the first two of these types of generalization. This case study identifies implications that might be relevant for practitioners elsewhere who are considering taking similar action to that of LeedsMet. The chapter also discusses how applicable various diffusion of innovation theories are to social networking sites and other Web 2.0 sites.

conducting interpretive research is the case study (Walsham, 1993). In addition, the case study method is particularly appropriate for dealing with contemporary technological innovations, such as social networking systems, in which the intervention being applied is difficult to distinguish from the context (Yin, 2003). The case study method uses a variety of sources of data to investigate a situation (Keddie, 2006). These sources may include: documentation, surveys, physical artefacts, archive material and observation (Yin, 2003). This study of participation in Facebook groups at LeedsMet features three of these data sources: surveys, artefacts and observation. The case study method is sometimes criticized for providing subjective results (Yin, 2003). In this study, Yin’s suggestions for assuring the validity and reliability have been followed and the tactics described in Table 1 have been used.

MAin findings students engagement with web 2.0 sites A survey was conducted with first-year computing students at the case study university to de-

termine their level of engagement with selected, well-known Web 2.0 sites and applications. A convenience sampling method was used and the questionnaire was issued to all students attending the first class of their introductory module at university. (The survey questions are contained in Appendix A). Ninety-four replied to the questionnaire, giving a response rate of 97%. Participation in the selected sites was measured in two ways: • •

Having accessed the site in the past seven days to view its contents Having contributed something to the site in the past seven days (e.g a discussion posting, an image or a video)

Ninety-three of the 94 students had recently participated in one or more of the selected Web 2.0 sites. Participation in individual sites is shown in Tables 2 and 3. Table 2 presents a ‘league table’ with sites being rated according to the percentage of students accessing them during the seven-day period. The media-sharing site, You Tube, was the most popular site for viewing (of those selected) amongst the students sampled with 90% of them having recently viewed its content on one or more occasions. The two social networking sites, My Space and Facebook, were placed third and fourth by this measure.

499

University 2.0

Table 2. Percentage student access to selected Web 2.0 sites Web 2.0 Application

Percentage of students accessing the site over a seven-day period You Tube

90%

Wikipedia

80%

My Space

80%

Facebook

67%

Blog

45%

Flickr

12%

Second Life

4%

Table 3 presents a further ‘league table’ with sites being rated according to the percentage of students who contributed content to them during the same period of time. It shows that the social networking sites, Facebook and My Space, were the most popular sites on which to contribute digital content amongst the sample.

types of University-related groups set up on facebook

There are 61 groups associated with LeedsMet in Facebook. The groups were analysed according to the focus of the group (academic-related or social) and the target user group (university community or wider community) and placed in the grid shown in Figure 1. The level of group activity was measured for each group using the following characteristics which are recorded in Facebook: •

The Facebook site was selected to explore the set up of university-related sites on social networking sites. There are two significant reasons for this choice of social networking site. Firstly, it is the most popular social networking site in Britain (Blakely, 2007). Secondly, the empirical evidence presented in this chapter suggests that it is the site in which students at LeedsMet are most active.

• •

The number of members of this group (which might include past and present members of the university community plus some outsiders) The number of photographs posted The number of postings made to the ‘wall’

Table 3. Percentage student contributing content to selected Web 2.0 sites Web 2.0 Application

Percentage of students contributing content to the site over a seven-day period

Facebook

39%

My Space

34%

You Tube

15%

Wikipedia

9%

Blog

9%

Flickr

2%

Second Life

1%

500

University 2.0

Figure 1. A categorisation of university-related Facebook groups

event group Eleven event groups were observed. As implied by the category name, these groups are usually set up to advertise an event which has some relevance to the university community. Examples found include a group dedicated to the latest tour by a band; the announcement of a fun run and a group which brought together people interested in a locallyrenowned pub crawl. Web 2.0 sites can provide an inexpensive channel for advertisers to communicate with potential buyers. Private individuals wishing to sell large items such as cars or boats sometimes use eBay for just this purpose. This category of group has the weakest link with the academic community. In the main, as might be expected given the nature of the groups, very few people became members. The exception was a group set up to advertise the new album by a band. This had more than one hundred members. However, these members were probably drawn from a wide community and were not necessarily from LeedsMet itself. Those events that were geographically in the proximity of the university, and might mostly attract students from this university, had a less significant membership. For example, the fun run had six members; the ‘pub crawl’ had 20 members.

On average, there were 36 members, 0.5 photographs and 2.3 postings on the ‘wall’ associated with this group.

Academic network The focus of these groups is on academic-related subjects, but they are targeted at the wider community rather than just people associated with LeedsMet. One academic network was discovered. This group was established to continue the debate that had been started at a one-day workshop hosted at the University. There were 2 members, 0 photographs and 1 posting on the ‘wall’ associated with this single instance of this group.

Curricular group These groups focus on academic courses studied at the university plus its services associated with the academic provision. These groups are targeted at the current and former members of the University community. They are mainly initiated by university staff. Twenty-three curricular groups were found. Most of these related to university schools or courses. On average, there were 21 members, 11 photographs and 1.9 postings on the ‘wall’ associated with this group. The curricular group

501

University 2.0

has, on average, slightly fewer members that the event group, but these are most likely drawn from the student body of the University.

extra-Curricular group These groups exist to fulfil a social purpose and the target participants are members of the University community. Most groups in this category relate to student societies and are mainly set up by students. Twenty-six groups were found. On average, there were 114 members,11.3 photographs and 40 postings on the ‘wall’ associated with this group.

students’ Attitudes to Universities Using social networking sites to engage with them The first year Computing and IT Systems students at LeedsMet were surveyed to determine their attitudes to social networking sites being used for university-related purposes. As part of their studies, these students all attended an elementary systems analysis module that featured a multi-media case study and assignment. At this university it is customary for all assignments to be made available to students online using its virtual learning environment. In addition to this, Web 2.0 versions of the case study were also released in the form of a wiki (http://www.thestudentwiki.org/wiki-dave/

index.php/Main_Page), a social networking site (http://www.myspace.com/systems_modelling) and associated podcasts. The attitude survey was administered after the students had completed the case study assignment. (The full list of survey questions can be found in Appendix B). At this juncture, they had direct, personal experience of the different formats of the case and the multiple channels of Web 1.0 and Web 2.0 communication. It is argued that this was a valuable contextual precursor to the survey, equipping the students with some real-life experience on which to base their attitudes and opinions. Waiting until this point in time to conduct the survey, however, meant that a reduced sample size of 51 students was available. Block delivery is employed on first year courses, so some students were still studying their systems analysis block at the time of writing. The key results are presented in Table 4.

disCUssion of findings Diffusion of Web 2.0 innovations is taking place throughout the student body. The empirical research presented here indicates that almost all students are actively engaged in one or more Web 2.0 applications. More than two-thirds of these students had accessed the Facebook social

Table 4. Students’ attitudes to using social networking sites for academic purposes Student Attitude

Results

In favour of academic use

• 59% of the sample felt that it is a good idea to use social networking sites for university assignments. • 46% would use social networking sites to discuss course matters with other students. • 45% felt they could be freer expressing views about university life on a social networking site, 39% had no view on this and 16% disagreed with this view.

Against academic use

• 84% stated a preference to access the assignment from the University’s VLE and 87% intended to use the VLE. • 45% would prefer to use the VLE for course-related discussions.

In the balance

• 34% felt that social computing should be kept for social purposes only, 28% had no view on this and 38% disagreed with this view.

502

University 2.0

networking sites in the week prior to the survey, with forty percent of these having contributed content to it during that period of time. With student participation levels such as this, it is understandable that universities might attempt to set up groups within the social networking artefact for the purpose of engagement with their students. The LeedsMet case study illustrates this. Within Facebook, there are 61 groups associated the university. Students set up some of these; others established by academics or administrators, and yet others have been placed there by outside individuals or organisations. This chapter has attempted to assess the degree of diffusion of these groups. The level of activity summarised in the Facebook metrics is taken as an approximation of diffusion. By analysing the groups according to the focus of the group and the target population, it was possible to arrange them into four categories: curricular, extra-curricular, event and academic network. It was observed that the extra-curricular group had the most significant level of diffusion. This group had the highest averages for membership, photographs placed on the sites and postings on the wall. The extra-curricular groups appear to be most closely aligned to the technology complex for the Facebook innovation. These groups facilitate social networking among students and past students from the University, using the technology for its original purpose. Whilst it is not unusual for an innovation to be used for purposes other that that which was initially intended (Tuomi, 2002), this characteristic alone does not explain the difference in take up between the extra-curricular group and the other three quadrants presented in Figure 1. It is possible to view the rollout of Facebook into university life, setting up university-related groups, from a technologically-deterministic standpoint. Thus a proponent of DOI theory might suggest: “This software is popular with young people, so surely it will attract their involvement

in university-related groups.” However, this oversimplifies the relationship between the various constituent parts of the Facebook innovation. The artefact is not an external variable applied to a social group. The social networking site is the union of the artefact and its adoption in a social setting. This is noticeably the case with Web 2.0 applications. Sites such as Facebook are shaped, developed, extended and enhanced by their participants. An approach which identifies some of the rich characteristics of the innovation within Facebook is the technology complex (Table 5). This highlights the softer, social attributes that are the essence of Facebook, whilst also accounting for the harder aspects of the technology. Attempting to diffuse this social networking site under conditions which do not match these characteristics is likely to be unsuccessful. The curricular and academic network types of university-related group are examples of situations which do not correspond fully to the various aspects of the technology complex. Consider, for instance, the organisational structure and cultural characteristics of this innovation. These do not reflect the expectations and norms of university life.

organisational structure The new mode of organisation of productive resources that is found on Web 2.0 sites has been called commons-based, peer production (Benkler, 2006). But is this structure not found in academic research communities? After all, the progress of most scientific knowledge follows this approach. Nevertheless, this organisational structure is not the one learners experience during undergraduate studies. Most of the resources they use will have either been produced by the hierarchy mode of production (developed by members of the faculty) or sourced from the market (textbooks, academic articles and so on). Consequently, students experience the peer production mode during their

503

University 2.0

Table 5. Exploring the Facebook innovation using the technology complex (adopted from Fleck and Howells, 2001) Technology Characteristic

Applicability to the Facebook innovation

Artefact

This characteristic is frequently considered to be the complete innovation. It is the most obvious and visible component of the technology. The artefact comprises of hardware and software elements. The key software involved is the Facebook suite of programs supplemented by the operating system; the hardware is the physical computing technologies needed for the Facebook participant to connect up to the internet.

Knowledge and Skills

All users need to acquire some skills to successfully adopt ICT-enabled change. In Web 2.0 innovations, this user knowledge takes on a different dimension or it actually helps to shape the innovation itself. A preferable term for a user is participant. On Facebook and similar sites, participants variously take on the role of consumer and producer. Their explicit and tacit knowledge (Davenport and Prusak, 1998) have helped to develop the site, building its growing network of members, adding social messages, photographs, posting discussion items and the like.

Organisational structure

Until recently, there were two modes of economic production: in-firm production, with a hierarchical structure used to co-ordinate and monitor activity and external sourcing in the market when the transaction cost of this mode cost less than producing in-house (Benkler, 2006). The voluntary, mass collaboration found on Facebook and other Web 2.0 sites represents a newly-emerged mode of production (Tapscott and Williams, 2007). Benkler (2006) has coined the term ‘commons-based, peer production’ for this activity. Is this a transitionary mode of production or will it survive longer term? We await the outcome of the Benkler-Carr Wager (Carr, 2006) to test the proposition that mass collaboration will survive as a distinct mode of production.

Culture

There are three aspects to the cultural characteristic of an innovation: roles, values and norms (Checkland and Howells, 2001). How do these apply to Facebook? As in any community, there are differing roles. Some lead by setting up new groups and networks, others contribute postings. On Web 2.0 sites such as Wikipedia, more experienced participants take on moderation roles. The values shared by the Facebook community include freedom for all to participate in it; everyone has the equal right to share their thoughts with others. If we take the norms to mean the standards applied by the community, we can observe a preference for using conversational language. This is in contrast to the more formal language that would be expected in academia or business.

social activity and are used to the hierarchy and market modes in their role as a learner (Stiles and Yorke, 2006).

Culture A core Facebook value is that of it being a community of peers, with all possessing the right to contribute and steer the direction of content (Tapscott and Williams, 2007). In contrast, the principle behind the curricular group, for example, is that of teacher-learner. In learner mode, students prefer to use sites based upon the teacher-learner premise. Accordingly, 87% of the students surveyed in this study intended to access learning materials via the University’s VLE even though these materials were available from a social networking site and a wiki.

504

If this is the case, why did 59% of the sample feel it was a good idea to use social networking sites for university assignments? Whilst they may have considered this to be beneficial in principle, their actual behaviour opposed this. The vast majority intended to go to the VLE instead. Furthermore, the research instrument itself may have played a part in the delivery of this contradictory result. The students’ lecturer administered this survey in class. A different response may have been forthcoming if an independent party had conducted the survey outside of class. A recent study conducted for the Joint Information Systems Committee in the UK found that 65% of students preparing to go to university use social networking sites and most of them “resented the idea that [the social networking sites] might be invaded by academics” (Swain, 2007, p1). More than one third of the students in the sample

University 2.0

surveyed at LeedsMet agreed with this viewpoint. A similar proportion was found in a previous study at a US college (Hewitt and Forte, 2006). Facebook, and other Web 2.0 sites, have norms that are quite different that those of academia (Educause, 2006; BBC, 2007, Goss and Acquisti, 2005). It is usual to adopt a conversational style of language on Facebook; a more rigorouslydeveloped argument is expected in university work. Carr (2005), a celebrated opponent of Web 2.0, adds that this kind of site represents the cult of amateur, exhibiting distrust for the professional, letting free content take over from quality content. Even if we disagree with Carr’s views, it should be clear that the norms and values associated with social networking sites are quite different from those that universities would expect to appear in the curricular group shown in Figure 1.

ConClUsion The external validity of case study research is sometimes called into question. Readers who are more familiar with positivist research might be concerned to ask how representative the findings from this case study are of all the members of the population from which it is drawn. Two types of generalisation are claimed by this in this chapter (see Table 1): ‘theory’and ‘implications’(Walsham, 1993).

implications for theory Previous studies have identified the need for further interpretive case studies to evaluate DOI theory (Wainwright and Waring, 2007). The empirical evidence presented in this chapter is a response to this call. Most previous DOI studies were based on the assumption that the economic activity being explored is arranged into a hierarchy or market (Yetton et al, 1999; Kautz and Larsen, 2000). This assumption is not applicable to Web 2.0 innovations in which peer production is the prevalent economic model.

Furthermore, DOI theory assumes that the artefact alone defines the innovation and its likelihood of adoption. The adopters are merely passive recipients of this technology (McMaster and Wastell, 2005). This assumption does not correspond to the practice found on Facebook and other Web 2.0 sites either. Recipients of the Facebook innovation are better described as prosumers (Tapcott and Williams, 2007), being actively involved in the shaping of the innovation itself. The technology complex (Fleck and Howells, 2001) is an analytical framework, based on the assumption that technology is socially shaped, which enables the softer, social elements of Facebook to be explored.

implications for Practice The empirical evidence presented here illustrates the extent to which young people of university age are participating in Facebook and other Web 2.0 sites. It is understandable that universities (and other organisations) might then wish to exploit Facebook as a means of communicating with current and prospective students and alumni. Nevertheless, universities should not rush to set up Facebook groups. This study has identified four types of group associated with the sector in Facebook. Of these, the extra-curricular group appears to be most successful in gaining student participation. The technology complex provides insight into the harder and softer aspects of the social networking innovation and pinpoints the significance of softer characteristics, such as organisational structure and culture, in the adoption of Facebook for new purposes. There has been recent discussion within the learning technology community (Atwell, 2007) about personal learning environments (PLEs). The PLE is a mixture of VLE and social networking space. This development appears once again to take a technologically-deterministic approach to the diffusion of innovation. To quote Atwell (2007, p1) “We have to look at the new opportunities for learning afforded by emerging technologies”. The 505

University 2.0

example of one university’s attempt to exploit Facebook presented in this chapter suggests that a number of softer and social characteristics of the PLE should be evaluated before attempting to diffuse this new Web 2.0-based learning environment throughout the sector.

Areas for future research Prevailing diffusion of innovation theory assumes that there are distinct producer and consumer groups involved with any innovation. With Web 2.0 applications, this delineation is much less clear. In fact, there is a sizable sub-group that occupies a place in both groups, consuming other people’s product and uploading their own. Further research is required to examine this aspect of Web 2.0 and its implications for diffusion theory. The current study evaluated a single social networking site in the context of a case study university. Future researchers might wish to explore other sites (e.g. Myspace) and a larger sample of universities to determine whether the behaviour identified in this chapter is replicated elsewhere. Finally, it is interesting to note that universities are now beginning to experiment with various Web 2.0 channels of communication with their students (e.g. Youtube, and Twitter). Will the universities experience similar results to the current study when using these technologies? This will be an interesting development from the current research.

referenCes Anderson. (2007). What is Web 2.0? Ideas, technologies, and implications for education. JISC Technology and Standards Watch. Retrieved on March 27, 2008, from http://www.jisc.ac.uk/media/documents/techwatch/tsw0701b.pdf

506

Attwell, G. (2007). Personal learning environments–the future of e-learning? eLearning papers, 2(1), 1-7. BBC. (2007, July 17). Unruly students’ Facebook search. BBC News. Benkler, T. (2006). The wealth of networks. New Haven: Yale University Press. Berners-Lee, T. (2006). Web 2.0: Nobody even knows what it means. Ars Technica. Retrieved on March 27, 2008, from http://arstechnica.com/ news.ars/post/20060901-7650.html Blakely, R. (2007, September 26). Facebook networking talks with Microsoft may value site at $10bn. The Times. Carr, N. (2005). The amorality of Web 2.0. Retrieved on March 27, 2008, from http://www.uic. edu/htbin/cgiwrap/bin/ojs/index.php/fm/issue/ view/263/showToc Carr, N. (2006). Calacanis’s wallet and the Web 2.0 dream. Retrieved on April 7, 2008, from http://www.roughtype.com/archives/2006/07/ jason_calacanis.php Checkland, P., & Howells, S. (2001). Information, systems, and information systems. Chichester, UK: Wiley. Davenport, T., & Prusak, L. (1998). Working knowledge. Boston, MA: Harvard University Press. Educause. (2006, September). 7 things you should know about Facebook. Educause Learning Initiative. Facebook. (2007). Company timeline. Retrieved on December 12, 2007, from http://www.facebook.com/press/info.php?timeline First Monday. (2008). Critical perspectives on Web 2.0. First Monday, 13(3). Retrieved on March 26, 2008, from http://www.uic.edu/htbin/cgiwrap/bin/ ojs/index.php/fm/issue/view/263/showToc

University 2.0

Fleck, J., & Howells, J. (2001). Technology, the technology complex, and the paradox of technological determinism. Technology Analysis and Strategic Management, 13(4), 523–531. doi:10.1080/09537320120095428 Gross, R., & Acquisti, A. (2005, November 7). Information revelation and privacy in online social networks. WPES’05, Virginia, VA.

Rogers, E. M. (1995). Diffusion of innovations. New York: The Free Press. Stiles, M., & Yorke, J. (2006). Technology supported learning–tensions between innovation and control, and organisational and professional cultures. Journal of Organisational Transformation and Social Change, 3(3), 251–267. doi:10.1386/ jots.3.3.251_1

Hewitt, A., & Forte, A. (2006, November 4-8). Crossing boundaries: Identity management and student/faculty relationships on Facebook. CSCW’06, Alberta, Canada.

Swain, H. (2007, October 18). Networking sites: Professors–keep out. The Independent.

Jeyaraj, A., Rottman, J. W., & Lacity, M. C. (2006). A review of the predictors, linkages, and biases in IT innovation adoption research. Journal of Information Technology, 21, 1–23. doi:10.1057/ palgrave.jit.2000056

Tuomi, I. (2002). Networks of innovation. Oxford: Oxford University Press.

Kautz, K., & Larsen, E. (2000). Diffusion theory and practice: Disseminating quality management and software process improvement innovations. Information Technology & People, 13(1), 11–26. doi:10.1108/09593840010312726 Keddie, V. (2006). Case study research. In V. Jupp (Ed.), The sage dictionary of social research methods. London: Sage. MacKenzie, D., & Wajcman, J. (Eds.). (1985). The social shaping of technology. Buckingham: Open University Press. McMaster, T., & Wastell, D. (2005). Diffusion–or delusion? Challenging an IS research tradition. Information Technology & People, 18(4), 383–404. doi:10.1108/09593840510633851 O’Reilly, T. (2005). What is Web 2.0? Retrieved on March 26, 2008, from http://www.oreillynet. com/pub/a/oreilly/tim/news/2005/09/30/what-isweb-20.html?page=1 Osborne, S. (1998). Voluntary organizations and innovation in public services. London: Routledge.

Tapscott, D., & Williams, A. (2007). Wikinomics. London: Atlantic Books.

Wainwright, D., & Waring, T. (2007). The application and adaptation of a diffusion of innovation framework for information systems research in NHS general medical practice. Journal of Information Technology, 22, 44–58. doi:10.1057/ palgrave.jit.2000093 Walsham, G. (1993). Interpreting information systems in organizations. Chichester: Wiley. Webopedia. (2008). Retrieved on March 26, 2008, from http://www.webopedia.com/TERM/W/ Web_2_point_0.html Webster, D. (2007). Myths, rhetoric, and policy in the information age: The case of closed circuit television. In D. Griffin, P. Trevorrow & E. Halpin (Eds.), Developments in e-government: A critical reader. Amsterdam: IOS Press. Yetton, P., Sharma, R., & Southon, G. (1999). Successful IS innovation: The contingent contributions of innovation characteristics and implementation process. Journal of Information Technology, 14, 53–68. doi:10.1080/026839699344746 Yin. (2003). Case study design: Research and methods, 3rd ed. Beverley Hills: Sage.

507

University 2.0

AdditionAl reAding Benkler, T. (2006) The Wealth of Networks, New Haven:Yale University Press Berners-Lee T (2006) Web 2.0: Nobody even knows what it means, Ars Technica, http://arstechnica.com/ news.ars/post/20060901-7650.html Retrieved 27/03/2008 Carr, N. (2005) The Amorality of Web 2.0 http:// www.uic.edu/htbin/cgiwrap/bin/ojs/index. php/fm/issue/view/263/showToc, Retrieved 27/03/2008 McMaster, T., & Wastell, D. (2005). Diffusion – or Delusion? Challenging an IS Research Tradition . Information Technology & People, 18(4), 383–404. doi:10.1108/09593840510633851 O’Reilly, T. (2005) What is Web 2.0? http:// www.oreillynet.com/pub/a/oreilly/tim/ news/2005/09/30/what-is-web-20.html?page=1 Retrieved 26/03/2008 Rogers, E. M. (1995) Diffusion of Innovations, New York: The Free Press Tapscott, D., & Williams, A. (2007) Wikinomics, London: Atlantic Books Walsham, G. (1993) Interpreting Information Systems in Organizations, Chichester: Wiley Yin (2003) Case Study Design: Research and Methods, 3rd Edn., Beverley Hills: Sage.

Key terMs And definitions Diffusion of Innovation: Rogers (1995, p5) defines diffusion as “the process by which innovation is communicated through certain channels over time among members of a social system.” In this chapter, the term has been defined as the process of adoption of the innovation by actors within a social system. Crucial to the argument presented here has been the notion that these ac-

508

tors may adopt different behaviours in the various social systems to which they belong. Innovation: A process, product or object that is perceived as new by the social group adopting it. The process of innovating consists essentially of two stages: creativity, during which the new idea is formed or adapted from elsewhere, and implementation, during which the innovation is successfully introduced to the social group and adopted by them over a period of time. Interpretivism: This research approach asserts that there is no objective knowledge waiting to be discovered. Reality and knowledge are socially constructed by human beings (Walsham, 1993). This epistemological position contrasts with positivist science in which hypotheses concerning an objective reality are tested and may be replicated by others. Social Networking Website: A website that facilitates online relationships between participants. Common facilities include messaging, photograph sharing and announcements. Web 2.0: Web 2.0 is a recently-coined term. O’Reilly (2005) claims to have been the first to use it at a conference in 2004. In this chapter, the following definition of Web 2.0 is employed: “Web 2.0 is the term given to describe a second generation of the World Wide Web that is focused on the ability for people to collaborate and share information online.”(Webopedia, 2008) This term has not received universal acceptance. BernersLee (2006) suggests that there is nothing new about this range of services. In his opinion, the World Wide Web has always been about peerto-peer communication. Web 2.0 adds nothing significant to the original design of the Web, it is merely a marketing hype. Carr (2005) discusses the amorality of Web 2.0, giving rise to the cult of amateur, in which professional quality is being sacrificed for wider democratic participation in the provision of digital content. Further critical perspectives regarding Web 2.0 can be found in a recent edition of the online journal First Monday (2008).

University 2.0

APPendiX A - stUdent web 2.0 engAgeMent sUrvey QUestions

A. social networking Participation 1. 2. 3. 4.

Which of the following sites have you used in the last week? (Wikipedia, Myspace, Youtube, Facebook, Perfspot, Second life, Flickr Have you visited any blog websites in the last week? How many podcast feeds are you currently subscribed to? How many podcasts have you produced since starting university?

b. social networking infrastructure 1. 2. 3. 4.

Do you have an iPod or similar player? If Yes, can it play video podcasts? Do you have an internet-ready phone? Have you accessed a wireless network in the last week?

C. About you 1. 2. 3.

Your gender Your date of birth Your ethnicity

APPendiX b - stUdents’ AttitUde sUrvey QUestions All statements require the respondent to choose the response that most closely matches their opinion from a Lickert scale.

statements: 1. 2. 3. 4. 5. 6. 7.

I think it is a good idea to use social networking sites for university assignments I intend to watch the case study interviews on my iPod I prefer to use the VLE for my university coursework If I need to discuss the course with other students, I will use social networking sites for this purpose I intend to watch the case study interviews in the VLE The wiki presents the case study materials in a more accessible format than the other formats I prefer to access the case study materials using the wiki

509

University 2.0

8. Social computing is just that. These sites should not be used for academic purposes 9. I prefer using my iPod for songs. I am not interested in storing case study interviews there 10. I would like more course materials made available for students on the move to play on their iPods 11. If I want to discuss the course online I will use the VLE 12. I resent the fact that my social networking site is being invaded by academics 13. I feel that I can be freer when expressing my views about university life on a social networking site rather than on the VLE

510

511

Chapter 28

On Using Wiki as a Tool for Collaborative Online Blended Learning Steve Wheeler University of Plymouth, UK

AbstrACt This chapter explores the use of the wiki and its role as a cognitive tool to promote interaction and collaborative learning in higher education. The importance of the software to enable student created content, storage, and sharing of knowledge is reviewed. This chapter provides an evaluation of some of the affordances and constraints of wikis to promote critical thinking within a blended learning context. It assesses their potential to facilitate collaborative learning through community focused enquiry for geographically separated students and nomadic learners. One particular focus of the chapter is the development of new digital literacies and how students present their written work in wikis. The chapter also examines group dynamics within collaborative learning environments drawing on the data from a study conducted at the University of Plymouth in 2007, using wikis in teacher education. Finally, the chapter highlights some recent key contributions to the developing discourse on social software in what has been termed ‘the architecture of participation.’

the iMPortAnCe of interACtion in online leArning Interactive digital media are assuming an increasingly important role in all sectors of education, with many universities developing e-learning strategies. The importance of interaction in distance education has been strongly emphasised (Moore, 1989; DOI: 10.4018/978-1-60566-384-5.ch028

Swan, 2002) and the use of technology to mediate communication between separated individuals is well documented (Shin, 2003; Gunawardena, 1990). Technology supported distance education can encourage and enhance collaborative learning processes (Jonassen, Peck & Wilson, 1999) where students actively seek out engagement with others because it is both useful and satisfying (Horizon Report, 2007). There is evidence that purposeful interaction can increase learner knowledge (Ritchie

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

On Using Wiki as a Tool for Collaborative Online Blended Learning

& Hoffman, 1997) but may be intensely personal and welcomed more by some students than others (Godwin, Thorpe & Richardson, 2008). The use of technology to support and facilitate interaction, if applied appropriately, tends to produce good learning outcomes, and new web based tools are increasingly available to the distance educator. The advent of Web 2.0 for example, has provided teachers with unprecedented opportunities. Web 2.0 based technologies are replete with rich social opportunities. For a growing number of teachers and students, social networking and social software have become fertile environments within which communities of learning can flourish and learn from each other (Wheeler, Yeomans & Wheeler, 2008; Ebersbach, Glaser & Heigl, 2006). There is also evidence that the practice of enabling students to generate their own content can encourage deeper levels of engagement with course content through the act of authoring, simply because the awareness of an audience, no matter how virtual or tentative, encourages more thoughtful sentence construction (Jacobs, 2003) and deeper critical engagement (Wheeler, Yeomans & Wheeler, 2008). Writing in blogs and wikis for example, compel students to carefully manage their impression (Goffman, 1959) encouraging them to think more clearly and critically about their arguments, and to articulate their ideas coherently and persuasively on a publicly accessible web space for an undetermined and invisible audience. Furthermore, there is a need to incorporate collaborative learning practices more deeply within all forms of education (Jonassen et al, 1999). Coupled with this need is a growing awareness that teacher roles need to be refined in a new knowledge economy. There is an established trend toward a form of learning where teachers abdicate their roles as instructors, and adopt a more supportive role (Harden & Crosby, 2000; O’Neill & McMahon, 2005). There is a tension here. Teachers fulfil a particularly important role as without teacher support, students can flounder, lose motivation,

512

or even drop out of the course. At the same time, the reduction in tutor-led instructional methods encourages students to take more responsibility for their learning. A fine balancing act is thus required where teachers facilitate and support learner participation, intervening where necessary, rather than providing sustained instruction. Students are increasingly adopting new roles as producers, commentators and classifiers (Horizon Report, 2007) within Web 2.0 based learning environments. They are participating more in the construction and organisation of their own knowledge rather than merely reproducing content as exemplified in instructional practices (Jonassen, et al, 1999) and this occurs increasingly outside the boundaries of contiguous education. This shift in emphasis, although grounded in social constructivist theory, also has drivers in new technologies (Richardson, 2006), and a post-modernist belief that knowledge should be discursively constructed across a multiplicity of sites (Gale, 2003). Such an approach to pedagogy, although arguably no longer radical, none the less constitutes an important part of the essence of blended learning, and has implications for a growing population of younger learners who appear to have a natural affinity to digital technologies (Prensky, 2006). It is also apparent that younger learners are more often on the move than earlier generations, and tend to engage in a ‘patchwork’ or portfolio of careers, job hopping as the need or interest dictates. Students are also more physically mobile than their forbears, and use cell phones and handheld devices to connect to their network of peers. Such nomadic wandering demands a new range of flexible learning skills and consequently a new culture of educational provision.

leArning As A noMAd Nomadic learning has been defined as ‘a form of learning in which a learner has continuity of service across different sessions and, possibly, dif-

On Using Wiki as a Tool for Collaborative Online Blended Learning

ferent locations’ (IEEE WG1, 2003). As a reworking of distributed learning, in which the learner’s study space has the same appearance regardless of location in space and time, nomadic learning must rely heavily upon ubiquitous information and communication technologies, pervasive digital services and adaptable, personalised software for its success. Nomadic learning in a more purist form could simply rely upon the student’s willingness to wander through unknown territory and to explore it with purpose. Baudelaire’s notion of the ‘Flâneur’ (Baudelaire, 1863) who wanders around a city to experience it. Baudelaire developed his concept of the Flâneur as someone who played an important role in understanding, and ultimately portraying the city. In a pedagogical context learners might stroll throughout a knowledge landscape, observing and enquiring where they may as they actually define it. This concept resonates within the new participatory and hypertextual territories constantly under construction across the social web. In contrast to early distance learning experiences, nomadic wanderings in digital worlds are rarely solitary activities and the opportunities for social contact and participation increase as one encounters blogs, wikis, social networking sites and massively multi-player online games. Students can freely develop their own knowledge content using a number of freely available social software applications, yet ostensibly they will seldom be alone within the architecture of participation known as ‘Web 2.0’(O’Reilly, 2004; Barsky & Purdon, 2006; Kamel Boulos, Maramba & Wheeler, 2006). Social software such as weblogging, social tagging, picture and file sharing, and of course the increasingly popular freely editable wiki, are providing students with unprecedented collaborative and interactive learning opportunities. It also offers students the chance to personalise their own routes to learning and trajectories of study. There is a further connection to be made between social software and personalised learning.

When mediated through the use of hypermedia, formalised education delivery begins to mirror the connective matrix of the human brain, assuming an infinite number of rhizomatic, non-linear forms (Deleuze & Guattari, 1987), that branch out into new territories, as the student finds new directions for study. Rhizomes, by their very nature, have no centre and their boundaries are dictated purely by environmental constraints. They will continue to grow exponentially in any and all directions as conditions allow. In the same manner, learning through social software enables students to decentralise from institutional constraints and follow their curiosity in any direction they see fit to pursue. The idealism of ‘personalisation’, previously so elusive, may actually be realised for many students who are empowered to wander down their own pathways within a strange yet captivating digital terrain. Indeed there now appears to be an increasing demand for flexible and independent learning that can be accessed without the barriers of time and place (Koper & Manderveld, 2004).

the soCiAl web Interaction and collaboration are increasingly being mediated through the social affordance of web based environments. Social networking spaces such as FaceBook, Myspace, Bebo and photo sharing sites such as Flickr and Picasa proffer unprecedented opportunities for students to share their ideas, celebrate and showcase their creativity, and receive immediate feedback from fellow networkers (Kamel Boulos et al, 2006; Richardson, 2006). Students around the globe are able to ‘swarm together’ (Rheingold, 2003), coalescing and co-ordinating their activities within rich and dynamic social environments, rather than wandering aimlessly through a socially ‘cold’ digital wasteland (Wallace, 1999). Whether through the sharing of useful bookmarks, holding real time conversations over Skype or voting in a

513

On Using Wiki as a Tool for Collaborative Online Blended Learning

poll to decide which idea wins the day, students can participate in a varied online environment that is rich in immediacy and social presence. Social networking thus encourages learners to participate in this digital milieu and brings them back regularly to repeat enjoyable and productive experiences. These are very desirable attributes for educational materials to exhibit (Horizon Report, 2007) and essential features for distance educators to exploit.

wiKis And groUP writing Knowledge creation is assuming increasing relevance and importance for students living in a world of constant change. For researchers and scholars too, there is a need for quick and easy access to the latest knowledge surrounding any given subject. One social software tool - the wiki - is developing quickly as a popular means to create, store and share knowledge in many areas of teaching and learning (Horizon Report, 2007). The word ‘wiki’ (from the Hawaiian wiki wiki) is translated as ‘to hurry’, which is particularly apt, as wikis can enable rapid and easy authoring and publishing direct to the web. More recently WIKI has been interpreted as the backronym ‘What I Know Is’ – a reference to the facility to contribute, store and share knowledge freely (Answers.com, 2007). Wiki pages can be used by any or all to publish new content direct to the web, including text, images and hyperlinks, to edit existing content, and also, because the wiki is fluid and open to all, it can be used openly to publish text, images and hyperlinks directly to the web without the need to learn HTML or other programming languages. Students seldom need to study alone due to facilitation of participation in a technologically mediated social space. Wikis are conducive to the formation of communities of practice (Kamel Boulos, Maramba & Wheeler, 2006). Moreover, wikis enable students to collaboratively gener-

514

ate, mix, edit and synthesise subject specific knowledge in an aggregative manner that often surpasses similar processes witnessed in traditional classrooms (Wheeler & Wheeler, 2009). The combined knowledge of the group – dubbed ‘the wisdom of crowds’ – is assumed to be greater than that of the individual, and content management is regulated by group members. Whilst this ‘architecture of participation’ (O’Reilly, 2004) has obvious attractions to the digital generation, what is contentious is the extent to which lay-generation of digital artefacts is accurate and appropriate to professional education. Moreover, the problem some teachers wrestle with is whether student generated content can be legitimised or should remain as ‘lay knowledge’. Clearly, wiki activities are afloat upon a sea of issues. Such new writing spaces challenge our conceptions of the manner in which knowledge is created, used and shared (Kimber & Wyatt-Smith, 2006). Perhaps the most important issue for educators to address centres upon the potential problems student created content can engender. There are no guarantees for accuracy and veracity on a wiki, as has been highlighted in a recent critique by Keen (2007). However, a recent survey conducted through the journal Nature found that Wikipedia, one of the most popular wiki knowledge repositories, is at least as accurate as Encyclopaedia Britannica (Terdiman, 2006). Wikis are susceptible to vandalism and malware (virus) attacks (Terdiman, 2006) so those moderating their use must be vigilant. Although the open nature of wikis creates opportunities for the deliberate sabotage, Owen, Grant, Sayers and Facer (2006) point out that there is often a critical mass of users who have sufficient ownership of the wiki to quickly intervene and clean up unwanted postings and recover the site. Really Simple Syndication (RSS) feeds alert community members to any changes that have been made to content, so that validation of the entries can be undertaken quickly and effectively. Ultimately, the ‘roll-back’ correction facilities built into most

On Using Wiki as a Tool for Collaborative Online Blended Learning

wikis can be used to quickly restore a page to its previous condition if required.

soMe PedAgogiCAl APPliCAtions Teachers need to be convinced of the usefulness of any new technology before they feel justified in adopting it. This section details some of the useful applications of wikis in education that have already been documented in previous studies. A range of high level thinking skills and rich social activities could result from the use and management of wikis. A few teachers are already exploiting the potential of wikis to transform the learning experience into one in which student centred learning can be facilitated (Wheeler et al, 2008). In classroom learning, teachers will need to encourage all members to contribute, thereby fostering a sense of community, but it is inevitable that some students will generate more content than others (Trentin, 2008). Wikis offer appropriate environments within which students who are geographically separated from one another can develop and maintain social ties. Teachers can encourage distributed students to ‘draw together’ by allocating each physically dislocated member a specific section or ‘stub’ on the wiki to manage. Other students are then able to add to it over the lifetime of a programme of study. Individual students could be assigned the task of finding relevant and reliable websites they can hyperlink back to the main wiki. Each student could also be assigned a specific time period during which they have responsibility to ‘patrol’ the wiki to ensure it has not been sabotaged or defaced in some way.

tAgging And folKsonoMies Students can tag web sites they find particularly useful so that they are visible to search engines.

Students can also make them available to others within one of the many free and highly visible social spaces such as Delicious or Digg. The practice of ‘social tagging’ replaces traditional, externally imposed hierarchies of categorisation (the taxonomy) with a knowledge organisation method that reflects the interest of the group, known as a ‘folksonomy’. Arguably, such folksonomies, because they are artefacts of the combined efforts of a community of interest, provide a more democratic and accurate representation of the current needs and aspirations of the user group (Owen et al, 2006) which can change in response to shifting interests (Kamel Boulos et al, 2006). Ultimately, the tagging of web pages makes their content more visible to a larger audience through an investment of semantic sifting by consensual decision making. Previous research has shown that larger audiences often encourage students to be more fastidious in their construction of wiki pages (Wheeler et al, 2008; Wheeler & Wheeler, 2009). One thorny issue emerging from the use of wikis in the classroom in one study was the problem of ownership and intellectual property (Wheeler et al, 2008). A large proposition of students strongly resisted the possibility that their contributions could be either altered or deleted by other group members. This seems to be less of a problem in user generated content sites such as Wikipedia, where contributors are relatively anonymous. In classroom contexts however, ownership appears to be an issue. For distance learners, where anonymity from the group is normally an issue, content generation within a wiki may prove to be less problematic. A recent social experiment by the publisher Penguin saw the creation of an online wiki novel which anyone was able to read, write and edit. More accurately, the community authored text actually resembled 150 or so short stories which had tenuous or almost non-existent connections to each other. It was the evolving product of the imagination of thousands of predominantly ama-

515

On Using Wiki as a Tool for Collaborative Online Blended Learning

teur writers. A warning at the foot of each editing section declared: ‘Please note that all contributions to PenguinWiki may be edited, altered, or removed by other contributors. If you don’t want your writing to be edited mercilessly, then don’t submit it here’(www. amillionpenguins.com) Although the experiment has now finished, it has raised important questions. It remains to be seen whether such a diverse and disparate group of authors can manage to coalesce their ideas into a coherent, unified and convincing fictional voice. As with any new technology or approach, the use of wikis in formalised education will accrue benefits for, and impose limitations upon all users. The aim of the current study was to evaluate the wiki as a tool to promote collaborative learning and to encourage independent forms of learning leading to critical thinking.

week to attend face to face lectures, and then for the majority of their time studied away from the group. The groups used the wiki for one entire term as an integral part of their undergraduate teacher training studies (n=35). There were 26 females and 9 males in the sample, with ages ranging from 18 to 45 years. All were learning how to use Information and Communication Technology (ICT) as teaching tools, but none had previously used a wiki.

resUlts And disCUssion There was an enthusiastic response from the participants, with all contributing to the wiki from the outset during classroom sessions. Initially students experienced some confusion and several expressed doubt over whether they could use the wiki effectively. Eventually, working through their confusion, students started to organise their activities by dividing tasks up and began to use it as a collaborative tool to support the learning of the entire group:

Method In this study, a wiki was created using free software at wikispaces.com. Thirty-five undergraduate and postgraduate students enrolled in a teacher training programme were recruited as participants for the study. As time went by and the wiki content grew, the activities of students within the space were evaluated through an examination of the discussion group contributions by the four groups. Toward the end of the program of study, the researcher posted strategically placed questions to the wiki site and then collated responses and coded them. A qualitative approach to the analysis of the discussion group postings was chosen in the expectation that students could elaborate on their experiences by writing their reflections onto the wiki discussion pages in response to questions from the researcher. All participation in the study was voluntary and all responses were anonymized. All the groups studied in a blended learning format, meeting once each

516

‘I was very skeptical about how the wiki could be used within this context. Immediately, we were confronted with utter confusion. However, we dealt with this well. Once we had delegated tasks to each other, I was much more happy to continue. […] Also, unless one person is in charge of the structure of the links, an unorganized set of pages may occur. Overall, a useful tool to develop one’s own and others learning’. (Second year male undergraduate) ‘To begin with, this session was quite stressful and I feel we weren’t voicing our ideas very well therefore it took a while to get started. However having organized what everyone was doing I think it worked really well- it is simple to use and easy to find your way around’. (First year female undergraduate)

On Using Wiki as a Tool for Collaborative Online Blended Learning

During their programme of study, students not only self-organised their roles and responsibilities, but some also collaboratively organised their content generation, posting ideas about their joint projects, highlighting items of mutual interest, and soliciting help from each other. One student for example, offered advice on the creation of hyperlinks within the wiki: ‘Just a humble suggestion when creating links. It is better to use a normal English explanation in the title of the resource link, then the actual URL itself does not have to be (or need to be) visible to the User. It makes for much easier visual filtering of sites that may be of interest to the browser. Remember there is a box specifically for the hyperlink itself as opposed to the link name on the ‘insert link’ wiki editing page. Hope this makes sense’. (Male postgraduate) Students also participated in specially devised group activities which required them to contribute to the wiki content so that over the course of the module a repository of knowledge was established. Some discovered that apportioning specific tasks to group members enabled the whole group to more effectively collaborate in the construction and maintenance of their knowledge repository. Although this strategy avoided conflict, it also had the deleterious effect of compartmentalising individual or small group contributions, so that students tended to read very little of the content created by their peers. During knowledge creation, issues of another kind arose. Students discovered that the limited capacity of the free wiki software had a detrimental effect on multiple-user editing. When two or more students attempted to edit or add to shared pages at the same time, conflict sometimes occurred, because the software was confounded by simultaneous postings. Students voiced their frustration about this on the discussion board. When asked the question: Now you have had the opportunity

to work on a shared space for the first time, what are your feelings? They answered: ‘Using a shared area is fun but can be annoying when the work you have just typed gets wiped because someone else is editing the page at the same time!!!! We worked in two teams and there was definitely competition’. (First year female undergraduate) ‘It is very difficult to share a page, as sometimes it can get deleted which can be very annoying!!!’ (First year female undergraduate) ‘…anger, as every time you type something and someone else is on the page, it will just delete everything you do. I believe that if you decide what to put on the page before you start typing, this will solve a lot of arguments’. (First year male undergraduate) This problem, it was observed, only occurred when students were studying face to face in the classroom, and ceased to be a problem when they were using the wiki off campus. Other questions were posted to the discussion group over the course of the 10 week module, including: To what extent has writing on the wiki helped you to improve your writing skills in general? There was a more positive response to this question from several students: ‘I think my writing has become more thought provoking. I’m now thinking at a higher level than I normally would’. (First year male undergraduate) ‘It has made me consider how i need to write so that it is suitable for other people to read it and understand it whereas normally it is only

517

On Using Wiki as a Tool for Collaborative Online Blended Learning

me that will be reading it’. (First year female undergraduate) ‘Writing on the wiki is a challenging activity which involves much thought about the length and structure of sentences as it is able to be read by anyone. The exclusion of a spellcheck also provides more challenges of careful thinking.’ (First year male undergraduate) One student recognised a change in his approach in which he adopted a more analytical style, and claimed that this was a direct result of reading the opinions of others on the wiki: ‘I think I am now developing a healthy critical and analytical writing style thanks to the wiki. Looking at other people’s opinions and findings has helped me to question what’s in front of me and I have found myself researching certain areas further to see if all opinions are the same’. (Second year male undergraduate) Such comments indicate that the students became more critically engaged as the module progressed, and began to think more carefully and analytically about the structure and content of their writing. Some also recognised their dependency on digital tools such as spellcheckers and agreed that this was a weakness in their academic armoury. Some admitted that they composed their contributions in a word-processor first, and then spell-checked before copying and pasting straight to the wiki. When asked: Has using the wiki limited any of your writing skills? Does it constrain you to do certain things in certain ways? One student felt that the wiki actually had constrained her writing: ‘I feel that it has limited my input, if I am unsure about something I won’t include it on the page. It has also made me more cautious with spelling etc’. (First year female undergraduate)

518

Another thought that she was less free in her writing because she was aware of people reading her contributions. She became less willing to take risks: ‘I feel that I’m limited on what I write because I know other people that I don’t even know will be reading it. I will only write something that I’m sure about, and not things that I think might be wrong or questioned by other people’. (First year female undergraduate) Other issues emerged during the first few weeks of implementation including an observation that most students contributed toward the wiki only during face to face lessons. Those few who contributed outside the classroom usually did so in the late evenings, or over weekends, when they were not required to attend other classes. This is an issue relating closely to that identified by Ebersbach, Glaser and Heigl (2006) who suggest that if such tools are not integrated into a regular pattern of learning activity, the result is that one or two people usually do the writing and others merely read. Further, it was observed that students tended to read only pages in which they had ownership over the content, which tended to nullify the original objective of collaborative learning through content generation. In situations where content was jointly developed by small groups of students, more reading was undertaken across several pages. Initially, students tended to copy and paste items from sites such as Wikipedia directly into the wiki pages, instead of creating hyperlinks to those sites. After several sessions, students learned to write their own annotations and commentaries and were encouraged by teaching staff to include images alongside their text, and other media such as movies and sounds. At times however, ‘design’ issues tended to obfuscate the original aims of personal learning through research, due to the awareness of a potential, hidden audience.

On Using Wiki as a Tool for Collaborative Online Blended Learning

reCoMMendAtions And ConClUsion Students will need to develop new skills to enable them to participate in the knowledge based global economy of the 21st Century. These skills will include knowing how to evaluate information critically, how to work independently without close supervision and being creative (DfES 2006). Wikis provide the perfect tool for teachers to extend those skills for the students in their care. Initially students may feel naturally daunted by the prospect of ‘writing to the web’, and about the potential to receive criticism from their peers of from an unseen web audience may provoke some anxiety. Such anxiety could be assuaged through confidence building exercises using simulated shared writing spaces which are open only to the peer group, prior to live wiki use (Wheeler, et al, 2008). Students should also be fully apprised of the probability of their work being edited or extended by others, or even deleted if considered inaccurate, irrelevant or inappropriate. All contributors should be aware that content editing is a natural and discursive feature of the wiki, and that collaborative learning requires negotiation of meaning and frank exchange of ideas (Kamel Boulos & Wheeler, 2007). Students should understand that once the ‘send’ button has been pressed, the idea no longer belongs exclusively to the originator, but now becomes the property of the whole learning community. Wikis are always a ‘work in progress’ so the untidy and chaotic nature of the pages should not be considered a limiting feature. Although design issues encourage readers to explore pages, content accuracy and relevancy should be prime considerations. Time should be provided for students to discuss their feelings and perceptions about participation and the social and pedagogical implications of user created content. Collaboration, rather than competition, should be emphasised as a key aim of any wiki based activity. Students should also be encouraged to

contribute to the wiki outside of classroom contact hours, and to share their thoughts, useful resources and discoveries as they generate them. When in class, wiki content creation should be an activity integrated into the fabric of lessons. Teachers should act as moderators rather than instructors, and may need to restrain themselves from direct action, in order to promote free and democratic development of content according to the principles embodied in the ‘wisdom of crowds’. Clearly there are opportunities to further investigate the use of wikis as a collaborative tool for learning. There are several key areas in which work can be done, including study of how to structure wiki activities to encourage better learner engagement, better integration of wikis and other social software tools into the classroom, and an examination of the many ways in which social software tools can be used in a variety of blended and nomadic learning contexts. As with many of the emerging social software applications, wikis have the potential to transform the learning experiences of students worldwide. The benefits appear to outweigh the limitations and there is clear evidence that when used appropriately, they encourage a culture of sharing and collaboration. For many students, wikis will be particularly appealing, providing instant, anytimeanyplace access to a dynamic and ever building digital repository of user-specific knowledge and a voice in a live community of practice.

referenCes Answers.com. Wiki Definition and History. (n.d.). Retrieved on March 2, 2007, from http://www. answers.com/wiki&r=67 Barsky, E., & Purdon, M. (2006). Introducing Web 2.0: Social networking and social bookmarking for health librarians. Journal of the Canadian Health Libraries Association, 27, 65–67.

519

On Using Wiki as a Tool for Collaborative Online Blended Learning

Baudelaire, C. (1863). La modernité, le peintre de la vie moderne, IV. Publisher unknown. Deleuze, G., & Guattari, F. (1987). A thousand plateaus: Capitalism and schizophrenia. Minneapolis: University of Minnesota Press. DfES. (2006). 2020 vision-report of the teaching and learning in 2020 review group. Nottingham: DfES. Ebersbach, A., Glaser, M., & Heigl, R. (2006). Wiki: Web collaboration. Berlin: SpringerVerlag. Gale, K. (2003). Creative pedagogies of resistance in post compulsory (teacher) education. In J. Satterthwaite, E. Atkinson & K. Gale (Eds.), Discourse, power resistance: Challenging the rhetoric of contemporary education. Stoke: Trentham Books. Godwin-Jones, R. (2003). Emerging technologies: Blogs and wikis: Environments for online collaboration. Language Learning & Technology, 7, 12–16. Goffman, E. (1959). The presentation of self in everyday life. New York: Doubleday. Gunawardena, C. N. (1990). Integrating telecommunications systems to reach distance learners. American Journal of Distance Education, 4(3), 38–46. doi:10.1080/08923649009526715 Harden, R. M., & Crosby, J. R. (2000). The good teacher is more than a lecturer: The twelve roles of the teacher. Medical Teacher, 22, 334–347. doi:10.1080/014215900409429 Jacobs, J. (2003). Communication over exposure: The rise of blogs as a product of cybervoyeurism. Cited in J. B. Williams & J. Jacobs (2004), Exploring the use of blogs as learning spaces in the higher education sector. Australian Journal of Educational Technology, 20, 232–247.

520

Jonassen, D. H., Peck, K. L., & Wilson, B. G. (1999). Learning with technology: A constructivist perspective. Upper Saddle River, NJ: Merrill. Kamel Boulos, M. N., Maramba, I., & Wheeler, S. (2006). Wikis, blogs, and podcasts: A new generation of Web-based tools for virtual collaborative clinical practice and education. BMC Medical Education, 6, 41. Retrieved on July 18, 2007, from http://www.biomedcentral.com/14726920/6/41 Kamel Boulos, M. N., & Wheeler, S. (2007). The emerging Web 2.0 social software: An enabling suite of sociable technologies in health and healthcare education. Health Information and Libraries Journal, 24(1), 2–23. doi:10.1111/j.14711842.2007.00701.x Keen, A. (2007). The cult of the amateur: How today’s Internet is killing our culture and assaulting our economy. London: Nicholas Brealey Publishing. Kimber, K., & Wyatt-Smith, C. (2006). Using and creating knowledge with new technologies: A case for students as designers. Learning, Media and Technology, 31(1), 19–34. doi:10.1080/17439880500515440 Koper, R., & Manderveld, J. (2004). Educational modelling language: Modelling reusable, interoperable, rich, and personalised units of learning. British Journal of Educational Technology, 35(5), 537–551. doi:10.1111/j.00071013.2004.00412.x Moore, M. G. (1989). Editorial: Three types of interaction. American Journal of Distance Education, 3(2), 1–6. doi:10.1080/08923648909526659 O’Neill, G., & McMahon, T. (2005). Student centred learning: What does it mean for students and lecturers? In G. O’Neill, S. Moore & B. McMullin (Eds.), Emerging issues in the practice of university learning and teaching. Dublin: AISHE.

On Using Wiki as a Tool for Collaborative Online Blended Learning

O’Reilly, T. (2004). Open source paradigm shift. Retrieved on July 18, 2007, from http://tim.oreilly. com/articles/paradigmshift_0504.html Owen, M., Grant, L., Sayers, S., & Facer, K. (2006). Opening education: Social software and learning. Bristol: Futurelab. Retrieved on November 20, 2006, from http://www.futurelab. org.uk/research Prensky, M. (2006). Listen to the natives. Educational Leadership, 63(4), 8–13. Report, H. (2007). Retrieved on March 2, 2007, from http://www.nmc.org/pdf/2007_Horizon_Report.pdf Rheingold, H. (2003). Smart mobs: The next social revolution. Cambridge, MA: Perseus Books. Rheingold, H. (2003). Smart mobs: The next social revolution. Cambridge, MA: Perseus Books. Richardson, W. (2006). Blogs, wikis, podcasts, and other powerful Web tools for classrooms. Thousand Oaks, CA: Corwin Press. Ritchie, D., & Hoffman, B. (1997). Incorporating instructional design principles with the World Wide Web. In B. H. Khan (Ed.), Web-based instruction (pp. 135–138). Englewood Cliffs, NJ: Educational Technology Publications. Shin, N. (2003). Transactional presence as a critical predictor of success in distance learning. Distance Education, 24(1), 87–104. doi:10.1080/01587910303048 Swan, K. (2002). Building learning communities in online courses: The importance of interaction. Education Communication and Information, 2(1), 23–49. doi:10.1080/1463631022000005016 Terdiman, D. (2006). Study: Wikipedia as accurate as Britannica. CNET News. Retrieved on January 5, 2007, from http://news.com.com/2100-1038_35997332.html

Trentin, G. (2008). Using a wiki to evaluate individual contribution to a collaborative learning project. Journal of Computer Assisted Learning. doi:.doi:10.1111/j.1365-2729.2008.00276.x Wallace, P. (1999). The psychology of the Internet. Cambridge: Cambridge University Press. Wallace, P. (1999). The psychology of the Internet. Cambridge: Cambridge University Press. Wheeler, S., & Wheeler, D. (2009). Using wikis to promote quality learning outcomes in teacher training. Learning, Media and Technology, 34(1), 1–10. doi:10.1080/17439880902759851 Wheeler, S., Yeomans, P., & Wheeler, D. (2008). The good, the bad, and the wiki: Evaluating student generated content as a collaborative learning tool. British Journal of Educational Technology, 39(6), 987–995. doi:10.1111/j.14678535.2007.00799.x

Key terMs And definitions Blended Learning: Learning which takes place both on and off campus, usually mediated via technology Collaborative Learning: Learning which encourages teams and small groups to work together Nomadic Learning: Learning on the move Social Software: software that encourages participation and collaboration, e.g. wikis and blogs Web 2.0: A term used to describe the participative and social elements of the World Wide Web Wiki: A website that can be edited by all those who have access

521

522

Chapter 29

Integration of Web 2.0 Collaboration Tools into Education: Lessons Learned Phillip Olla Madonna University, USA Elena Qureshi Madonna University, USA

AbstrACt Web 2.0 is opening new capabilities for human interaction. It also broadens the way technology is used to collaborate more effectively. This chapter discusses instructional strategies and techniques used to successfully utilize Web 2.0 tools for classroom collaboration. It will also shed light on pedagogical issues that arise with the implementation of Web 2.0 into the educational setting. The chapter will present case studies describing how various Web 2.0 applications can be incorporated into a variety of courses in the areas of nursing, education, and computer information systems. Finally, recommendations for teachers and students on how to effectively use Web 2.0 tools to improve collaboration will be outlined.

introdUCtion The Internet has had a phenomenal impact in the educational setting creating opportunities in e-learning, information access, publishing and research. Some University officials were concerned that the Internet would destroy the traditional campus life (Ryan, 2001). This is far from the case. The Internet has presented new opportunities along with some significant challenges to the DOI: 10.4018/978-1-60566-384-5.ch029

educational setting. The emergence of Web 2.0 into the education setting is having the same impact as Web 1.0 but is much more pervasive and powerful. The idea that students can collaborate in real-time to create digital contents such as words, programs, images or theories is a compelling notion. In addition to content creation, students can now access a vast amount of information from a variety of excellent and dubious sources. The challenge for educators is to comprehend how to utilize the tools and applications to improve the teaching and learning process.

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Integration of Web 2.0 Collaboration Tools into Education

One of the main objectives of this chapter is to demonstrate how Web 2.0 is changing the landscape of higher education and the application of Web 2.0 learning technologies. Further, this book chapter will present a number of case studies describing how various Web 2.0 applications can be incorporated in variety of courses in the areas of nursing, education, and computer information systems. The chapter will discuss instructional strategies and techniques used to successfully utilize Web 2.0 tools for classroom collaboration. It will also shed light on pedagogical issues that arise with the implementation of Web 2.0 into the educational setting. Specifically, the chapter will be broken down into four sections as follows: The first section will focus on the concept of collaboration and the benefits of collaboration in the classroom environments, and theories of learning in collaboration. The second section discusses Web 2.0 concepts and terminology and describes Web 2.0 services. The third section focuses on four educational cases utilizing Web 2.0 applications in graduate and undergraduate courses in the areas of nursing, education, and computer information systems. Prior to the conclusion, the fourth section will outline recommendations for teachers and students to effectively use Web 2.0 tools to improve collaboration.

seCtion 1: CollAborAtion This section will focus on the concept of collaboration and its benefits as related to Web 2.0 phenomenon.

the importance of Collaboration in education The importance of collaboration in education has long been acknowledged. It is a critical aspect of successful teaching and learning practices and achieving better outcomes. Collaboration is an

intricate concept with multiple attributes. The term is hard to define because it encompasses a variety of activities such as peer response groups, peer tutoring, peer workshops, group research projects, classroom group discussions, and learning communities. The focus of this chapter is on what is most useful about collaboration within the context of education: how we can usefully share ideas and information among the members of a project and what it is about this team project that creates benefits for the collected individuals involved in this project. According to Linden (2002), collaboration can provide the following benefits: Better use of scarce resources; cost savings; ability to create something that you cannot create on your own; higher quality, more integrated product for the end users; potential for organizational and individual learning; and better ability to achieve important outcomes. When creating an effective collaborative environment, instructors always consider using technology and the Internet. In the last few years Web 2.0 – more collaborative Internet – has created a buzz in education. Web 2.0 programs are rapidly becoming tools of choice for a growing body of classroom educators. University instructors are discovering that Web 2.0 tools provide compelling teaching and learning opportunities. It is obvious that the Web 2.0 is changing the very nature of student work. The fact that a student’s work can be seen, commented on, and collaboratively improved by a larger participative group of people has a very favorable effect on students’ engagement with course content. Students become more involved in educational discussions and debates. They come to realize that they work collaboratively with their peers and not just their instructors in the discovery, exploration, and clarification of knowledge. A very proactive learning environment is the result of effective use of the Web 2.0 tools. To sum up, Web 2.0 is opening new capabilities for human interaction. It also broadens the ways

523

Integration of Web 2.0 Collaboration Tools into Education

we can use technology to help us collaborate more effectively. The next section of this chapter will deal with the Web2.0 concept and terminology.

seCtion 2: web 2.0 ConCePts And terMinology This section will categorize Web 2.0 services.

web 2.0 Concept and terminology Web 2.0 is a difficult concept to understand or define, the problem stems from the diverse views on the topic by industry experts. There are two contradictory views on the Web 2.0 phenomenon. The first perspective synonymous with Tim O’Reilly is that Web 2.0 is a trend in the use and design of internet based technology that aims to promote concepts such as creativity, knowledge generation, information sharing, and collaboration. These important concepts have facilitated the development and evolution of virtual communities and online services, such as social-networking sites, social bookmarking, blogs, and wikis. The Web 2.0 term was coined at the first O’Reilly Media Web 2.0 conference (O’Reilly, 2005). Tim O’Reilly is the Founder of O’Reilly Media and defines Web 2.0 as”the business revolution in the computer industry caused by the move to the Internet as a platform, and an attempt to understand the rules for success on that new platform” (O’Reilly, 2006). The main problem with the Web 2.0 concept is that some people believe that it insinuates new version of the World Wide Web, when in fact it does not introduce any new technical specification updates to the original www. It just transforms the way software developers and end-users use Web. The most famous adversary to the Web 2.0 concept is Sir Tim Berners-Lee, the inventor of the Web. Tim Berners-Lee once described the term “Web 2.0” as a “piece of jargon” in a podcast. His argument was that Web 2.0 is all about blogs and wikis,

524

which are people to people. The notion of people to people is the original premise of the Internet (Anderson, 2006). Another criticism of Web2.0 is the lack of sustainable business models from the companies operating in this arena. There is a déjà vu of the Dot-com bubble between 1995–2001, and Web 2.0 has also been dubbed “Bubble 2.0” by the economist. It is difficult to differentiate Web 1.0 and Web 2.0 sites but typically Web 1.0 sites adopt an hierarchical structure, with a front page or home page leading to various subpages, augmented by linking and search capabilities, while Web 2.0 sites resemble real-world social networks exhibiting different structures. Another difference is the fact that Web 2.0 sites aim to display a user-centric view of the site; this means that each individual will only view details applicable to them, such as friends, documents, and pictures. Another difference is that rate of content updates. In Web 2.0, the content of a site can change frequently due to the fact that user generated content can be incorporated into the site. Web 2.0 Websites can incorporate a varieties of technologies and programming language, however, one approach that has fuelled the Web 2.0 development is the use of AJAX and Flash. Ajax is short for asynchronous JavaScript and XML, and is considered an important building box in popular Web 2.0 technologies. Ajax is a combination of a variety of programming languages that integrate data presentation, interactive data exchange between the database and Web page, client side scripts, and asynchronous update of server response. Ajax acts as an intermediary and resides on the client and sends requests to a server while updating the Web pages asynchronously. Flash objects provide similar functionality to Ajax because they also communicate asynchronously with a server. Flash applications require a widely available Adobe plug-in to be installed. Developers use a variety of software toolkits to create internet applications that allow the developed application to be rendered either as Flash objects or Ajax components (Cormode & Krishnamurthy, 2008).

Integration of Web 2.0 Collaboration Tools into Education

Figure 1. Educational Web 2.0 services

Categorization of web 2.0 services by functionality Due to the constantly evolving technology and the lack of a clear consensus on what Web 2.0 really is or whether Web 2.0 actually exists, it is difficult to categorize applications and services that are deemed Web 2.0 compliant. To resolve this issue, this article uses the concept of service functionality to categorize the applications. Most of the applications mentioned below are already being incorporated into education to various degrees. These following applications are not real technologies, but a collection of interconnected services or user processes that utilize the Internets open standards and building blocks (Anderson, 2007). Important applications used in the case study in the following section are online collaboration sites. In addition to the functionality, simplicity and user-friendly access are the most important attributes to consider (Kaplan, 2002) for collaboration sites. The benefits of Web 2.0 sites is the ease of use, and the lack of software installations required. This supports the notion by Kaplan

(2002) that suggests users should spend little time leaning the application or the technology that runs the collaboration site and spend more time performing the tasks and learning about the content. The technology should be transparent to the instructor as well as the learner; no prior technical expertise should be required to customize or manage the environment. Some of the collaboration sites considered for class exercises are discussed in Table 1 below. Another Web 2.0 concept that is being incorporated into education is a wiki. A wiki is Website that allows users with access to collaboratively create, edit, link, and categorize the content of a Website in real time covering a variety of reference material. Wikis have evolved from being purely a reference site into collaborative tools to run community Websites, corporate intranets, knowledge management systems, and educational sites. Although, Wikipedia is the most common wiki, it was not the first. The first ever Wiki was developed by Ward Cunningham and it was called WikiWikiWeb, originally described as “the simplest online database that could possibly work” (Cunninghan, 2005). Advocates of Wikis are en-

525

Integration of Web 2.0 Collaboration Tools into Education

Table 1. Example of collaboration Websites Collaboration Suites

Address

Purpose

Comments

Vyew

www.vyew.com

Online Collaboration suite to share documents, whiteboard, talk, video conference shared workspace.

This Website is has so many features that it is heavy on resources and response and load time may be delayed. Also very heavy on the network traffic. Basic Version is free.

Talk and Write

TalkAndWrite.com

Real time interaction software. It simulates the interaction of up to 10 partners, working side by side on a common document. It supports handwrite, draw, erase, highlight, insert text on a document while discussing data over Skype. Both can see their own and the partner’s mouse pointers, which can be used to point out items on the document.

This application is a Skype plug-in. A download is required, but it works well with no delays. This is a free plug-in.

Twiddla

http://www.twiddla.com/

Real collaboration, in real time. Mark up Websites, graphics, and photos, or start brainstorming on a blank canvas. Browse the Web with your friends or make that conference call more productive than ever.

No plug-ins, downloads, browser-agnostic, user-friendly, with one-click audio chats. There can be wireless network delays if the whole class attempts to access the application at the same time.

Bump In

http://site.bumpin.com/

BumpIn is a browser add on that allows you to chat with people browsing the same page as you. You can have private chats, or shout to everyone visiting that particular page. You can also use the Web-based version, by clicking on the homepage, and start chatting with people without installing any software on your machine.

This is a social browsing application that will allow students to discuss a Website in real-time from the comfort of their home.

thusiastic about the ease of use, flexibility, open access, however, there are considerable problems with using wikis in the educational setting (Ebersbach et al., 2006; Lamb, 2004). A new concept that is becoming popular with education research is Social Bookmarking. Social bookmarking is a technique that users use to organize, categorize, search, share, and manage bookmarks of Web pages they are interested in on the Internet using metadata. Users save links to Web pages and can either make the links public or keep them private and share with a specified user group. Users with access rights can view the bookmarks chronologically, by category or tags, or via an Internet search engine. A tag is a keyword that is added to a digital object (e.g., a Website, picture, or video clip) to describe it,

526

but not as part of a formal classification system (Anderson, 2007). With the growing popularity of sites such as delicious, CiteULike, etc., this concept has had phenomenal growth with over million Websites tagged. Most of the social bookmark sites support the use of informal tags to book mark sites as an alternative to the traditional browser-based system of folders. This approach has the benefit of allowing users to view bookmarks associated with a chosen tag along with information, such as number of users who have bookmarked the same site. The new generations of social bookmarking Websites also draw inferences from the relationship of tags to create clusters of tags or bookmarks. The concept of tagging is not just a social bookmarking phenomenon.

Integration of Web 2.0 Collaboration Tools into Education

Table 2. Examples of wiki’s used in education http://wikisineducation .wetpaint.com/

Wetpaint Wikis in Education, a place where educators come together to share tips about using wikis to enhance the learning experience.

This is a very important resource for educators. It provides a wealth of knowledge on how to use a wiki in education.

http://pbwiki.com/education.wiki

This Wiki creation Website is free to educators and offers videos and presentation materials especially for educators.

The interface it intuitive and easy to use. Once setup correctly this is a safe and easy way for teachers to get students collaborating.

http://sleducation.wikispaces.com/

The Second Life in Education Wiki space is designed to provide an overview of the educational possibilities of virtual worlds, in particular Second Life.

This is a very good example of how a wiki can be used in education.

It is now common with other digital artifacts using services such as Flickr photos and videos and YouTube videos and podcasts that allow a variety of digital artifacts to be socially tagged.A Web 2.0 phenomenon that has received a lot of attention is the social networking concept. A social networking Website is an online resource for building virtual social network communities of individuals with common interests or who are interested in exploring the interests and activities of others. The technical features that allow these sites to be so successful include the functionality to chat, send private and public messages, email, video + voice chat, file sharing, blogging, discussion groups, and application sharing. The concept of Social networking has revolutionized how people interact, communicate and share information with one another in today’s society. There is no question that the online social networking phenomenon is now entrenched in the lifestyles of the X generation. Recent research rivals this phenomenon with television. The study by Grunwald Associates LLC that was conducted in cooperation with the National School Boards Association demonstrated that 9- to 17-year-olds report spending almost as much time using social networking services and Web sites as they spend watching television. Among teens, that amounts to about 9 hours a week on social networking activities, compared to about 10 hours a week watching TV (NSBA, 2007). In total, 96 percent

of students with online access reported that they had used social networking technologies, such as chatting, text messaging, blogging and visiting online communities, such as Facebook, MySpace and services designed specifically for younger children, such as Webkins and the chat sections of Nick.com. Eighty-one percent said they had visited a social networking Web site within the past three months and 71 percent said they had used social networking tools at least once weekly. Further, students reported that one of the most common topics of conversation on the social networking scene was education. The study has also demonstrated that almost 60 percent of students who used social networking talked about education topics online and more than 50 percent talked specifically about schoolwork. What made this study so compelling is the enormous effort by educational institution to keep this type of technology out of the educational systems during school time. There is more research to investigate the positive benefits of social networking. There has been explosive growth in creative and authoring activities by students on social networking sites in recent years. With words, music, photos and videos, students are expressing themselves by creating, manipulating and sharing content online. Section 3: Case Studies.This section will focus on a few specific examples of using Web 2.0 applications in graduate and undergraduate

527

Integration of Web 2.0 Collaboration Tools into Education

courses in the areas of nursing, education, and computer information systems. The section will also present applications that can be used to create multimedia applications. All the participants of the described below case studies were graduate students enrolled on a full and part time basis at the medium size University in South Eastern Michigan. Over sixty students taking graduate courses in education, nursing, and computer information systems were asked to participate in this study during Fall and Winter semesters of 2007-2008 academic year. The main purpose of these case studies was to evaluate students’ attitudes towards using Web 2.0 resources and to assess the effects of using Web 2.0 tools on student learning.

CAse stUdy 1: Using web 2.0 resoUrCes to CreAte An ePortfolio Seven graduate students in the College of Education Technology program were participants of this case study. They were all mature females (age 30-45) currently teaching in K12 system and working towards their Masters degree in Educational Technology. One of the course objectives was to learn how to document personal growth and development using Web 2.0 tools. Electronic portfolios are becoming a popular alternative to traditional paper-based portfolios because they offer practitioners and peers the opportunity to review, communicate and assess portfolios in an asynchronous manner. An ePortfolio is a collection of work developed across varied contexts over time. The portfolio can advance learning by providing students and/or faculty with a way to organize, archive and display pieces of work. Students in EDU 6260 Instructional Design and Multimedia class were required to create a Flash Website to be used as an ePortfolio for this course. Students were required to update their Website after each class by adding a new

528

page or comment with material that pertained to their course work. There is a consensus that selfreflection is an important component of electronic portfolio development. Instructors should encourage students to self-reflect on the artifacts they add to the portfolio, only then students will gain from the rich learning experience that ePortfolio development can provide. Each ePortfolio page was consisted of a short reflection from the days course, and in addition, some kind of resource, reference, Website, document, piece of multimedia, that adds on to what students were doing in the course (for example, students could include a Website that was a useful link for anyone interested in learning more about storyboarding). In addition, students were asked to view the other students’ ePortfolios and comment on their ideas and reflections. The Website chosen to assist in completion of this project was http://www.cabanova.com. It is online software for creating free Flash Web sites. Overall students’ comments about their experiences working on their ePortfolios were very positive. Many students agreed that ePortfolios helped them track their progress over time. It was “a great way to communicate with one another.” Other Web 2.0 tools that were evaluated as a part of their ePortfolio “helped them bookmark and edit resources that they needed;” Web 2.0 applications that students came across with let them “share information and learn from one another.” Graduate students repeated over and over again that the resources that they had used in this class were very easy to use and could be incorporated into lessons within the K-12 classroom. One of the students pointed out the essence of an ePortfolio stating that “it creates a platform for taking control of student’s personal knowledge management.” Further, all the students seemed to agree that their “ideas/opinions have changed since they were able to evaluate various online resources in class.” Prior to this course they were “unaware of just how much “stuff” is out there, both free and

Integration of Web 2.0 Collaboration Tools into Education

Figure 2. Example of e-portfolio

not free.” Each student was able to compile a list of valuable free resources that they would be able to use in their classrooms. One of the concerns that was brought up related to the lack of students’knowledge with respect to the social aspect of Web 2.0 and its effective application in the classroom. In spite of this concern, students were optimistic about the usage of Web 2.0 tools stating that “technology is the present and the future and although it is hard to keep up with all of the new things as educators we must be somewhat up-to-date on new things because our students are.”

CAse stUdy 2: web 2.0 resoUrCe evAlUAtion Graduate students in EDU6260 Instructional Design and Multimedia were asked to review at least seven different types of Web 2.0 teacher re-

sources such as start pages, quizzes, concept maps, multimedia production software, etc. Students were required to participate in the class discussion about the resources. They were also asked to integrate one of the resources into a classroom lesson/project (PowerPoint assignment). Students were required to use the following criteria to help them evaluate the resources: • • • • • • •

How could I use this as a tool with students? Parents? Other teachers? Community? Ease of Use Cost Flexibility of Design Security Ability to correct mistakes Add media?

Students commented on coming to the realization how ignorant they were about the concept

529

Integration of Web 2.0 Collaboration Tools into Education

Figure 3. Example of Web resources

of Web 2.0: “there is so much in cyberspace that I don’t use or know about. The free sites are amazing! “ “After evaluating all the different Web-sites, I never realize how much you can get off the internet. You can find anything and everything, free and not free. You can meet people from all over the world. I always found it over whelming when trying to find things. I never knew where to look. There is so much out there that you could use in your classroom that if free of charge or with a small fee that is affordable. “ Graduate students pointed out that a large number of Web 2.0 sites they looked at “allowed teachers and students to share and discuss information and learn from each other from different perspectives.” One of the students stated, “What could be a better forum for the students who already spend a great deal of time on the computer? When students (and teachers) can communicate

530

with text and graphics and then be able to discuss the information and viewpoints of others - everybody wins.” In the process of completing the Web 2.0 resource evaluation, students expressed the following concerns: “I have explored the sites recommended in class and as much as I think they are great tools, my concern is having the time to keep up with them. Integrating the sites into my routine, whether in the classroom or at home, takes time, dedication and the technological tools to do so.” Other concerns that were voiced out were: “the amount of information available about anything and anyone is astounding. All this info at the tip of our fingertips. However, as teachers we need to be careful and teach the students that there are sites out there that have false information and not all that is on the net is true. Also, copyright laws and plagiarism is another issue that needs to be addressed with students as they work on projects.

Integration of Web 2.0 Collaboration Tools into Education

“ According to the students’ comments, the Web 2.0 resource evaluations helped them “orient themselves in the overwhelming pool of Web 2.0 applications.” Students improved their online resource assessment and evaluation skills. They also mentioned that they developed a strategy of dealing with new online applications and felt more competent while making a decision about the applicability of a specific Web 2.0 resource in their classroom teaching. The majority of students in this class felt very strongly about the fact that teachers need to become very familiar with all that Web 2.0 has to offer. “It is our responsibility to help the students navigate through all of the social networking that is out there so that they are use it in a safe and effective manner.” Overall, the resource evaluation exercise has clearly demonstrated that with respect to Web 2.0 tools the positives far out-weigh the negatives. In the word of one of the students, “With science, and this “new wave of innovation”, the sky’s the limit!! “

CAse stUdy 3: CollAborAtion APPliCAtions - siMUlAting A teleMediCine ConsUltAtion. The Nursing Informatics (MIS 5230) course was designed to present applications of informatics systems to nursing and healthcare practitioners. The course addressed healthcare informatics issues covering hardware, software, databases, communications applications, and computer developments and associated legal and ethical issues. All the participants in this class where nurses who where managers or managers in their respective fields ranging from home care manager to Incentive Care Unit (ICU) supervisor. The students were all in the MBA program and had a wealth of

field experience in the nursing domain but limited Information Technology experience. The assignment for the class involved separating the students into two groups. Group A: Rural Hospital. The objective of this group was to present a patient exhibiting a variety of conditions using the telemedicine infrastructure to Group B. Group B: Urban Hospital. The objective of this group was to diagnose the condition as quickly as possible and suggest a treatment plan. The students used a Web collaboration platform called Vyew. This Website provides real-time interaction between people and content. Vyew is flexible and allows users to import documents in a variety of formats such as MS Office documents, pdf, Flash, MP3, video, graphics, and screen captures. It also provides support for saving, tracking, and logging all meeting activities. The medical conditions used in the telemedicine scenario were researched by the students who were all qualified nurses. The information was presented in a manner that would make the diagnosis difficult as this was a time challenge. The students made use of the following functionalities: • •



Document Sharing: The patient notes were uploaded to the workspace for both groups to discuss. Video conferencing: The students used the Web camera to show images of a rash that the patient had on his arm. Chat features: At one point the Group A students disabled the video / audio connection and were only accessible via chat. This was done as a ploy to slow down the process.

531

Integration of Web 2.0 Collaboration Tools into Education

Figure 4. Example of collaboration site, source www.yvey.com

Outcomes: 1.

2.

3.

532

Both groups were very successful in completing the task relativity quickly - seven minutes and ten minutes. The students were very impressed at how easy they could set-up exercise. They all agreed that this was a valuable exercise that could be replicated in the real world if security and privacy could be guaranteed. There was no prior warning about this activity and no training on using the application, but the students completed the activity with minimum supervision.

CAse stUdy 4: CollAborAtion design of heAlth toolbAr. This assignment was designed to show how projects can be completed online with minimal contact. The students were tasked with developing a health information application that would help people with low literacy levels to validate health Websites. The Internet has been acknowledged by a variety of studies as an essential source for people to access health information. However, it is widely agreed that health information on the Internet is of inconsistent quality. Numerous approaches for evaluating health Web sites have been proposed to address this issue such as: quality labels, user guidance systems, health Website

Integration of Web 2.0 Collaboration Tools into Education

evaluation guidelines and third party certification process. There are still groups of users, however, who encounter barriers during health information seeking due to their low health literacy, poor information literacy skills, or difficulty with the medical vocabulary. This assignment involved the students collaborating to design the rules for identifying good and bad sources. Initially, the students were presented with Web space to present their finding, but this quickly became chaotic and they were informed to collaborate on a diagrammatic Web space to present the design rules. The students evaluated three Websites and agreed to use gliffy.com. An example screenshot is presented below. Gliffy has a very user friendly interface with ‘drag and drop’ feature. It supports text editing, different colors, sizing, connector tools, and contains over 100 objects (symbols). The feature that the students used the most was collaborate and publish feature. Outcomes: Initially, the students were introduced to Microsoft Visio that provided a baseline for the functionality. They were then asked to evaluate some online applications. They chose Gliffy because of the functionality and ease of use. Other Web 2.0 applications reviewed included Cumulate Draw - http://draw.labs.autodesk.com/ADDraw/ draw.html and zcubes.com. Although zcubes.com seemed to have a lot of functionality, the interface was difficult to use and the students were presented with too much information. The learning outcome from this exercise was immense. Once the Web 2.0 application was selected, the students’ main task was to agree and draw a diagram. Over a period of two days, the diagram seemed to change on an hourly basis, as the group could not decide on the shapes, direction, or themes to research. Once these problems were discussed via chat, the situation improved. The following week was dedicated to building one master diagram with five lower level supporting diagrams. Students presented the work in class. They all agreed that if they had access

to a desktop application it would have been very challenging to coordinate the tasks and complete them in the timeframe. Section 4: Recommendations and Future Trends.This section will outline recommendations for teachers and students to effectively use Web 2.0 tools to improve collaboration. Basic definitions of terms will be provided at the end of the chapter. Collaborative learning provides an environment to enrich the learning process. Introducing Web 2.0 tools into an educational system creates more realistic social contexts, thereby increasing the effectiveness of the system. Such an environment would help sustain the student’s interests and would provide a more natural learning habitat. It is apparent that Web 2.0 collaboration is becoming one of the promising learning paradigms at the higher education level. Despite the complexity involved in the design of effective collaborative learning, more research efforts should be spent to explore this paradigm to provide better learning environments. Lessons that were learned in the process of utilizing Web 2.0 can be divided into two categories: (a) Benefits and (b) Challenges. Benefits: •



Sharing documents and task lists online. One of the beneits cited by the students of using the Web2.0 resources was the capability to share and store information in a central location. Students described in the review sessions how they are more likely to utilize these types of technologies for future group project work to make the working process easier when working on online projects. Access to ePortfolios. Students were very excited at the possibility of having an electronic portfolio that could be accessed online and would not be tethered to the institutions Learning Management System (LMS). One of the main concerns was the

533

Integration of Web 2.0 Collaboration Tools into Education

Figure 5. Example of drawing application, source www.gliffy.com



534

fact that students spend a lot of time working on their electronic portfolio, but once they leave the institution, they no longer have access because the ePortfolio typically is accessed from within the LMS. This is not a problem with a Web2.0 ePortfolio. Using Web 2.0 tools to design and develop an ePortfolio might be a great solution for educational institutions that do not have this feature built-in to their LMS (such as Blackboard for example). Understanding the concept of telemedicine. Although most of the students had heard about telemedicine, they had a view that this technological concept had to involve very expensive and sophisticated

technology. After they completed the exercise using standard Web based technology they commented on how the technology was actually just an enabler that the facilities connecting people, the most important aspect of telemedicine is still the medical personal. They also discussed the need for more training in the social elements of telemedicine as opposed to the technology aspects. It was clear from the subsequent discussion of the nurses that they now understood all the components involved in holding a live telemedicine session both social and technological. The fact that the Webcam was turned off during the session, was seen as an opportunity to use more

Integration of Web 2.0 Collaboration Tools into Education



verbal description along with supporting documents to complete the session. Learning how to evaluate online resources. Many student- teachers hear a lot of buzz words but do not know where to start as far as their own classroom implementation goes or how to correctly assess the appropriateness of various Web 2.0 tools. As a result, courses such as described in this chapter that teach students how to ind, evaluate, and use Web 2.0 tools become critical. Challenges:

1.

2.

3.

4.

5.

6.

Finding great Websites that were 2.0 proved to be a larger challenge than it had been anticipated by the course instructors especially for EDU classes. A large number of 2.0 tools had to be carefully assessed before selected for course work. Slow Internet connection can disrupt the class, it is important to test the Website in the classroom with more than one computer connected. Software updates: Web based applications such as Java and Flash must be up-to-date, valuable class time will be wasted if students have to download and install new versions of flash or java to support the Web 2.0 site. Always have a plan B. There is a chance that during an in-class session the Web2.0 site could be down for maintenance, it is vital that the instructor has an alternative exercise for the class. It is very important the students and faculty take regular back-ups of data that is stored on the Web 2.0 site. All the applications discussed in this paper provide the option for downloading the data in a variety of formats. If the Web 2.0 site is critical for the class, the instructor must register for news updates and newsletter from the site, this

will prevent surprises like takeovers or the service becoming unavailable. 7. Most of the functionality used in the case studies was free. There are always limitations of the basic or free model, it is important that the instructor / faculty understands the business model of the Website to ensure that the limitations do not interrupt the class exercise. 8. Most of the free sites have adverts, this was an acceptable inconvenience us, but this may not work for everyone. Check to make sure that the adverts being presented to the students are in line with the mission of your institution. 9. Using wireless networks to access Web2.0 sites can cause network issues, make sure this is tested properly before the class. 10. Sometimes the free version does not allow more than 5 users to connect to the same session. This problem was avoided by allowing the students to work in groups; this also helped with the network load. 11. All students must have an active email address to sign up for services. This should not be a problem, but some students did refuse to register because they did not have an active personal email and they refused to use their work emails for fear of spam.

ConClUsion The chapter presented instructional strategies and techniques used to successfully utilize Web 2.0 tools for classroom collaboration. Using four case studies it provided example of how Web2.0 could be used in university courses, and presented pedagogical issues that arise with the implementation of Web 2.0 into the educational setting. This book chapter was an attempt to consolidate some of the helpful online products and services that can help students, teachers, and administrators

535

Integration of Web 2.0 Collaboration Tools into Education

utilize valuable Web 2.0 resources into their teaching and learning. It is obvious that the new Web will be a platform that joins up all the functions and data with a new network of connections below the application level, enabling a new generation of applications that will make teaching and learning easier, more productive, and more fun. Given that we live in a society that is heavily dependent on computer technology, and there is strong evidence that students can benefit from using various Web 2.0 tools in their collaborative efforts.

Cormode, G., & Krishnamurthy, B. (2008). Key differences between Web 1.0 and Web 2.0. AT&T Labs–Research. Retrieved on September 1, 2008, from http://www.research.att.com/~bala/papers/ Web1v2.pdf Cunningham, W. (2006). What is a Wiki? Retrieved on May 13, 2008, from http://www.wiki. org/wiki.cgi? Cych, L. (2006). Social networks. In Emerging technologies for education, BECTA. Coventry, UK: Becta ICT Research.

Anderson, N. (2006). Tim Berners-Lee on Web 2.0: Nobody even knows what it means. Retrieved on May 12, 2008, from http://arstechnica.com/news. ars/post/20060901-7650.html

Dillenbourg, P., Baker, M., Blaye, A., & O’Malley, C. (1996). The evolution of research on collaborative learning. In E. Spada & P. Reiman (Eds.), Learning in humans and machine: Towards an interdisciplinary learning science (pp.189-211). Oxford: Elsevier.

Anderson, P. (2007). What is Web 2.0? Ideas, technologies, and implications for education. Technology & Standards Watch Report. Retrieved on May 13, 2008, from http://www.jisc.ac.uk/ media/documents/techwatch/tsw0701b.pdf

Doise, W. (1990). The development of individual competencies through social interaction. In H. C. Foot, M. J. Morgan, & R. H. Shute (Eds.), Children helping children (pp.43-64). Chichester: J. Wiley & Sons.

Benkler, Y. (2006). The wealth of networks: How social production transforms markets and freedom. USA: Yale University Press.

Ebersbach, A., Glaser, M., & Heigl, R. (2006). Wiki: Web collaboration. Germany: SpringerVerlag.

referenCes

Brittain, S., Glowacki, P., Van Ittersum, J., & Johnson, L. (2006). Podcasting lectures. Educause Quarterly, 29(3). Boulder, CO: EDUCAUSE. Retrieved on May 1, 2008, from http://www. educause.edu/apps/eq/eqm06/eqm0634.asp

Felix, L., & Stolarz, D. (2006). Hands-on guide to video blogging and podcasting: Emerging media tools for business communication. Massachusetts: Focal Press.

Bruffee, K. (1984). Collaborative learning and the conversation of mankind. College English, 46(7), 635–652. doi:10.2307/376924

Henneman, E., Lee, J., & Cohen, J. (1995). Collaboration: A concept analysis. Journal of Advanced Nursing, 21, 103–109. doi:10.1046/j.13652648.1995.21010103.x

Bruffee, K. (2003). Collaborative learning and the Conversation of Mankind. In V. Villanueva (Ed.), Cross-talk in comp theory (pp.415-436). Urbana, IL: NCTE.

Kaplan, S. (2002). Building communities-strategies for collaborative learning. Retrieved on May 1, 2008, from http://www.learningcircuits. org/2002/aug2002/kaplan.html%20

536

Integration of Web 2.0 Collaboration Tools into Education

Lamb, B. (2004). Wide open spaces: Wikis, ready or not. Educause Review, 39(5), 36–48. Retrieved on May 1, 2008, from http://www.educause.edu/ pub/er/erm04/erm0452.asp Linden, R. (2002). Working across boundaries: Making collaboration work in government and nonprofit organizations. Jossey Bass Nonprofit & Public Management Series. Loan-Clarke, J., & Preston, D. (2002). Tensions and benefits in collaborative research involving a university and another organization. Studies in Higher Education, 27(2), 169–185. doi:10.1080/03075070220120001 Loucks-Horsley, S., Hewson, P., Love, N., & Stiles, K. E. (1998). Designing professional development for teachers of science and mathematics. Thousand Oaks, CA: Corwin Press. Millen, D., Feinberg, J., & Kerr, B. (2005). Social bookmarking in the enterprise. ACM Queue. Retrieved on May 10, 2008, from http://www. acmqueue.com/modules.php?name=Content&p a=showpage&pid=344 NDBA. (2007). Creating & connecting research and guidelines on online social—and educational—networking. NATIONAL SCHOOL BOARDS ASSOCIATION. Retrieved on May 12, 2008, from http://www.nsba.org/SecondaryMenu/ TLN/CreatingandConnecting.aspx O’Reilly, T. (2005). What is Web 2.0. O’Reilly Network. Retrieved on May 2, 2008, from http://www.oreillynet.com/pub/a/oreilly/tim/ news/2005/09/30/what-is-Web-20.html O’Reily, T. (2006). Web 2.0 compact definition: Trying again. Retrieved on May 10, 2008, from http://radar.oreilly.com/archives/2006/12/Web20-compact-definition-tryi.html

Ractham, P., & Zhang, X. (2006, April 13-15). Podcasting in academia: A new knowledge management paradigm within academic settings. In Proceedings of the 2006 ACM SIGMIS CPR Conference (SIGMIS CPR ‘06) on Computer Personnel Research (pp. 314-317), Claremont, CA. New York: ACM Press. Stewart, D. (1988). Collaborative learning and composition: Boon or bane? Rhetoric Review, 7(1), 58–83. Stvilla, B., Twidale, M., Gasser, L., & Smith, L. (2005). Information quality discussions in Wikipedia. (Tech. Rep.). Florida State University. Retrieved on May 14, 2008, from http://mailer. fsu.edu/~bstvilia/ Yamane, D. (1996). Collaboration and its discontents: Steps toward overcoming barriers to successful group projects. Teaching Sociology, 24(4), 378–383. doi:10.2307/1318875

Key terMs And definitions Blog: A Blog is a site maintained by an individual, organization or group or people, which contains recurrent entries of commentary, view points, descriptions of events, or multimedia material, such as images, pictures or videos. The entries are typically displayed in reverse chronological order with the most recent post being the current focus. ePortfolio: An ePortfolio is a collection of work developed across varied contexts over time. The portfolio can advance learning by providing students and/or faculty with a way to organize, archive and display pieces of work. Podcast: A podcast is a series of digital files distributed over the Internet using RSS for playback on portable media players, such as IPods, PDA, smartphones, or computers. RSS: RSS is short for “Really Simple Syndication.” This is a technique to easily distribute

537

Integration of Web 2.0 Collaboration Tools into Education

content such as news headlines, Websites update notices, and sometimes movies and applications to a wide audience. An RSS document can be referred to as a “feed”, “Web feed,” or “channel.” The feed will contain either a summary of content being distributed from an associated Web site or the full text of the article. Telemedicine: Telemedicine is an application of clinical medicine where medical information is transferred via telephone, the Internet or any other telecommunication networks for the purpose of consulting, diagnosing or performing remote medical procedures or examinations.

538

Wiki: A wiki is Website that allows users with access to collaboratively create, edit, link, and categorize the content of a Website in real time covering a variety of reference material. Wikis have evolved from being purely a reference site into collaborative tools to run community Websites, corporate intranets, knowledge management systems and educational sites.

539

Chapter 30

ECHO: A Layered Model for the Design of a Context-Aware Learning Experience Hadas Weinberger HIT – Holon Institute of Technology, Israel

AbstrACt In this chapter, we suggest Echo, a model for utilizing Web technologies for the design of Web-based context-aware learning. Web technologies are continuously evolving to enhance information retrieval, semantic annotation, social interactions, and interactive experiences. However, these technologies do not offer a methodological approach to learning. In this chapter, we offer a new approach to Web-based learning, which considers the role of the user in shaping the learning experience. The key feature in Echo is the analysis and modeling of content for the design of a Web-based learning experience in context. There are three elements in Echo: 1) a methodology to guide the learning process, 2) techniques to support content analysis and modeling activities, and 3) a three-layered framework of social-semantic software. Incorporating this framework facilitates knowledge organization and representation. We describe our model, the methodology, and the three-layered framework. We then present preliminary results from on-going empirical research that demonstrates the feasibility of Echo and its usefulness for the design of a context-aware learning experience. Finally, we discuss the usefulness of Echo and its contribution to further research in the ield of Web technologies.

introdUCtion Web-based learning is a multifaceted phenomenon informed by a spectrum of theories. Theories of communication (Alavi & Leidner, 2001; Rafaeli & Raban, 2005; Te’eni, 2001) eLearning (Al-Kahlifa DOI: 10.4018/978-1-60566-384-5.ch030

& Davies, 2007; Paavola et al., 2004; Parameswaran & Whinston, 2007; Schmidt, 2005; Schmidt, 2008; Tzitzikas et al., 2006) and eLearning 2.0 (Downes, 2005; Ebner, 2007; O`Hear, 2006) guide the design of the learning processes and media integration. Theories of knowledge management (Grace & Butler, 2005; Nonaka & Tekeuchi, 1995), information science (Hjorland, 1997; Latham, 2002; Muresan

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

ECHO

& Harper, 2004), information retrieval (Feng et al., 2005), organizational memory (Weinberger et al., 2008b) and organizational learning (Argrys & Scon, 1978; Paavola et al., 2004; Weiling, 2006) inform the management of content related aspects. Other research contributions have taken the technology perspective (Ebner et al., 2007; Schmidt, 2008) or focused on specific media (Abel et al., 2004; Bao et al., 2007; Hotho et al., 2006b, Javanovic et al., 2007) to inform the design of learning activities. What is yet lacking is a comprehensive and systematic model of systems and practices for the design of Web-based learning. In this chapter we define Web-based learning as the manipulation of a set of content analysis techniques aiming to establish a conceptual model of a task specific domain. This definition indicates that the learner is responsible for constructing the learning process in context. Context awareness and the design of a conceptual model are essential to this process since without context the learning experience would be meaningless. We identified three challenges to be considered as part of the design of context-aware learning. First is the need for a framework of tasks and deliverables required to guide Web-based learning. Such a framework would advise the user about the how (tasks, activities and techniques) and what (deliverables to be developed and socialsemantic applications to be used) of this experience. Principles of Organizational Memory life cycle (Weinberger et al., 2008a) could be adapted for the definition of a dedicated methodology, describing the processes and the tasks to be performed. Second is the need to compile a collection of adequate techniques and advise the user about the specific methods for using these techniques to support the advised systematic methodology. Example techniques could follow classification methods and knowledge organization systems, such as taxonomy, thesaurus and ontology (Abel et al., 2004; Christiaens, 2007; Feng et al., 2005; Latham, 2002). However, while these techniques are clearly vital to this end, their association with

540

Web-based learning is not obvious. The third challenge identified for this research is the need to model the learning process in an integrated way towards context-aware learning incorporating social-semantic applications and the adequate techniques. In this chapter we suggest a balance between technical feasibility (i.e., Web technologies) and human action (i.e., learning). Of the many aspects of Web technologies, we focus on the role of the human agent at the interface between the technology perspective and the content perspective. Specifically, we challenge the way individuals retrieve information from the Web to enhance learning. This chapter takes a stance towards these challenges to suggest the Echo model for the design of a context-aware learning experience. In Echo we follow principles of Web software development such as Express Programming and Information Science techniques (Hjorlad, 1997; Xia et al., 2006). For the design of our initial framework we follow Paavola, Lipponnen & Hakkarainen (2006) theory of Trialogical learning (TL). The TL theory suggests that successful learning occurs when learners collaboratively develop shared objects of activity in a systematic way. In their theory, Paavola et al. (2006) attribute collaborative knowledge creation processes to three learning modes: knowledge acquisition, social participation and collaborative knowledge creation. We adopt this theory and incorporate its elements as a basis for the design of our model. Echo provides a needed balance between technology and human action and supports peers collaboration. Such a balance has been lacking in previous work, and we believe it reflects a more realistic picture of information retrieval in general and Web-based learning in particular. There are three contributions in this chapter corresponding to the three elements in Echo. First is a methodology to guide the learning process – advising the tasks, activities and deliverables that are required for a context-aware learning

ECHO

experience. Second is a collection of techniques to support content analysis and modeling activities – facilitating the development of context. Third is a three-layered framework providing the collaboration mechanisms that facilitate the procession of the advised methodology using social-semantic applications (e.g., Folksonomy, Ontology and Mashup). Based also on preliminary findings from on-going empirical research we believe this framework can contribute to the advancement of Web-based context-aware learning in education as well as in business settings. In the next section we review previous research on Web-based learning and social-semantic software. In the following section we present Echo and discuss the research methodology. We continue with a discussion of the feasibility of Echo based on preliminary findings from empirical research. We then give an outlook on future research prospects and conclude with a summary and discussion of the suggested model contribution to the field.

bACKgroUnd web-based learning Web technologies and learning are two intertwined topics that have grown in importance for individuals, academia and organizations (Abel et al., 2004; Al-Khalifa et al., 2007; Weiling, 2006). Web technologies are designed to affect not only what we seek – by advising us about categories relevant to the subject of our quest, but also what we find – the results of the information retrieval process. Consequently, these technologies shape our learning experience. Web-based learning takes place as part of an educational or an organizational setting conducted independently of a specific domain. In this chapter we use the term Web-based learning rather than, for instance, eLearning 2.0 (Downes, 2005). This is for several reasons.

eLearning 2.0 refers to the application of social software for learning, suggesting increased user involvement, for instance by the design of learning objects (Ebner, 2007; Gasevic, 2005; Sa’ncehzAlonso & Vovides, 2007). However, beyond the discussion of the necessity of the ‘e’ in eLearning, this concept is also used in the context of Learning Management Systems (LMS), while our focus is set on Web content and Web technologies. We incorporate the attributes of Downes’ definition as part of the term Web-based learning that we believe to be more accurate from an ontological perspective. While there is a spectrum of innovative Web 2.0 social-semantic applications available for integration into educational and business settings, there are still difficulties that hinder a Web-based learning experience from being effective (Bao et al., 2007; Schmidt, 2008). One reason for this is that in order to achieve meaningful results a learning experience should be set in context, which the prevailing technologies do not support in a satisfactory manner. One way of understanding the difficulties that are part of Web-based learning would be using as lens the concept and classification of Web generations. Web 3.0 (i.e. the Semantic Web) emerged from understanding that context builds meaning otherwise hidden in text. Web 2.0 and the increased user involvement (Ramakrishnan & Tomkins, 2007), challenge designers towards innovative perception of the user perspective and user content. Recently the emergence of socialsemantic applications (e.g., blogs, folksonomies and Wiki) motivated a change in the perception of Web generations bringing together the user perspective and the content perspective in a unified perception of context, for instance by using (user-generated) metadata. We believe that the key to resolving the difficulties is in understanding the challenge to be at the intersection between content, context and learning. Current research focuses on practical issues of Web IR such as ranking of previous search results, aiming to include also textual-semantic

541

ECHO

results based on user-feedback (Hotho et al., 2006; Schmidt 2008; Veres, 2006). What is still missing is a thorough inquiry of the users’ role. While indeed social-semantic applications highly involve user’s content, this is not the case for Web IR, which has not yet matured towards incorporating the user’s perspective. One way of doing so is through context. Context is both the anchor of learning and basis for evaluation of effective learning. Context is part of learning because without context the association between a learning goal and a learning outcome will be meaningless. The process of communication can be useful as a metaphor to the requirements embedded in Web-based learning. The two main forces challenging the communication process are how we communicate, and what we communicate (Te’eni, 2001). Successful Web-based learning is subject to an effective utilization of the how for the purpose of meeting the goals defined for the what. In this case the how is enabled by the composition of media – social-semantic software and methods – the adequate techniques, and the what is the content we wish to obtain. The problems acknowledged by this metaphor are the need to guide the how and what in context. Effective learning, much like useful communication, relies on bridging between media and message, as means to establish context, since without context the content would be meaningless. Brezillon & Brezillon (2006) cite Wittgenstein, who from philosophical-linguistic perspective defines context as: “the meaning of a word is its use in a language”. Taking this observation one step further, Zimmerman, Lorenz & Oppermann (2006) – motivated by the Object Oriented paradigm, suggest that the definition of context will be extended in a faceted approach referencing categories that are of relevancy to the subject entity describing formal and operational attributes. To paraphrase these observations we consider the definition of context in the context of its evolution as part of a learning process. Web-based

542

learning evolves from a generative dance between the three content types possessed by an individual prior to the Web interaction (i.e., data, information and knowledge; Brashnik, 2007) and what data information, knowledge and meta-knowledge (i.e., or metadata) is acquired and used by the individual as a result of the interaction. The latter – metadata, is considered an agent of learning (Christiaens, 2007). In the process of learning we build our model based on the acquired knowledge and consequently establish our context. The investigation of context and its modeling procedures, i.e., the processes of assigning meaning to concepts by establishing the relationships between concepts (Brezillon & Brezillon, 2007), follows three perspectives: 1) the task perspective, 2) the content perspective and 3) the user perspective. The task perspective concerns the objects of interest of the information retrieval, such as the learning goal (Mullaholand et al., 2001). The content perspective is focused on the many aspects of Information Retrieval (IR) required to meet the task, such as the objects of inquiry and the relevant communication means (Nonaka & Takeuchi, 1995; Rafaeli & Raban, 2005). The user perspective is focused on achieving meaningful results (Feng et al., 2005; Kwan & Balasubramanian, 2003) using adequate techniques. Considering these three perspectives as part of design coincides with the principles suggested by the TL learning paradigm – emphasizing the user rule, content management and structured learning. The three aspects of context that should be considered as part of context design are: syntax, semantics and pragmatics (Frank, 2007). Syntax concerns the data level, for instance search terms, which are the building blocks of the inquiry. Semantics concerns the information level represented by the objects of inquiry. Pragmatics concerns the integration of data information and knowledge as a result of the context-aware learning experience. From a philosophical perspective understanding context requires adhering to the principle of the hermeneutic circle. This principle suggests:

ECHO

“that all human understanding is achieved by iterating between considering the interdependent meaning between parts and the whole that they form” (Klein & Myers, 1999), which is the approach taken here.

social-semantic Applications Web technologies and the emergence of socialsemantic software and its applications support the feasibility of rich content and user-centric design (Brezillon & Brezillon, 2006; Kraines et al., 2006; Kwan & Balasubramanian, 2003). Web 2.0 and its associated social applications, which are known for their usability (Akavipat et a., 2006; Studer, 2006; Kraines et al., 2006) and for their intuitiveness, offer the user an opportunity for taking an active role in the process of content generation, annotation and representation, for instance as means to encourage collaborative learning. Examples are knowledge management and sharing systems, aka, Bookmarks (e.g., Bibsonomy.org), Syndication applications, adapted search engines, e.g., Google customized search; SWICKI (Google.com; www. eurekster.com, respectively) and on demand ontologies (e.g., http://www.wikimindmap.org/). The semantic Web builds on mark-up language and standards using XML syntax that is the basis for semantic annotation and for the resource description framework (RDF), which in turn supports conceptual modeling using RDF Schema, knowledge representation and visualization and the OWL ontology modeling and representation language (Studer, 2006). Taking the user perspective, these technologies support several tasks and applications which can be integrated as part of Web-based learning. We use several examples to illustrate this. XML, for example, is the format for dynamic content update carrying automated updates from Web resources using syndication technologies (e.g., RSS). Web applications and Application Programming Interface (API) are a basis for Mashup. Mashup – Web application hybrid, is an architecture using AJAX

(Asynchronous Java Script and XML) allowing the integration of different content types of various digital genres (Askehave & Nielsen, 2005); Open API and a user-focused approach provide also for services such as for the definition of a personalized (i.e., customized and adapted) search engine, for which iGoogle (http://www.google. com/ig) is an example. Based on these standards, social-semantic applications allow the construction of collaborative classification and knowledge management systems such as folksonomy, a lightweight conceptual structure (Hotho, 2006). Folksonomy is a social-semantic network for the management of context. One example is the use of Folksonomy in a knowledge management and sharing system (i.e., bookmarks). Folksonomies are not only part of a host of methods and techniques intended for knowledge organization, presentation and collaboration (Hotho et al., 2006a; Kings et al., 2007); they are also an integral part of these applications. Previous experience with social Web applications in organizational settings (Wang et al., 2003) demonstrates that folksonomy can be used in more than one way. For example, folksonomy can be formulated and negotiated as basis for enterprise knowledge management, a part of ontology development (Sicilia et al., 2006; Veras, 2006). The semantic-collaboration mechanism, behind Folksonomy and other social-semantic software, is metadata that is constructed, embedded and represented as tags and labels. These semantic annotation mechanisms are shaping cognition and innovations, for instance in eLearning (Al-Khalifa & Davies 2007) reflecting personal knowledge and encouraging knowledge creation and knowledge management (Falconer, 2006; Maamar et al., 2004; Veres, 2006). Social annotation is but an example of a technique used by naïve users as well as by Web developers, who utilize it for ranking techniques (Bao et al., 2007). The most formalized knowledge organization system that can be developed based on Semantic

543

ECHO

Web technologies is ontology. Ontologies can affect learning (Abel et al., 2004; Mika, 2005) and are used for eLearning as means for knowledge organization, presentation and reuse (Jovanovic et al., 2007; Brazhnik, 2006; Christiaens, 2006; Tzitzikas et al., 2006). The development of an ontology as domain model could contribute to the success of learning processes in several ways (Jovanovic et al., 2007). One example is by connecting one or more ontologies to constitute a learning object (Gasevic et al., 2005; Weinberger, 2008c). However, technological feasibility does not guarantee assimilation. In order for these standards and their applications to be more relevant to practice there is a need to provide users with a unified framework that is prescriptive enough to direct action and descriptive enough to allow its adaptation in various settings. The user view is sought after also from the developers’ perspective. One example is search engines that are incorporating data obtained based on user-feedback. This example belongs in the context of several problems that prevailed in Information retrieval (IR) research. One such instance is concept selection and flexibility in query expansion. Noteworthy, of the two common query expansion automated techniques being either global (e.g., term clustering, latent semantic indexing, similarity thesauri, phrase finder) or local (e.g., feedback mechanism), the latter have shown to be more effective than the former in general. However, there are also serious drawbacks to local techniques (XU and Croft, 2000) for which the analysis technique based on concurrences with the query terms within the top-ranked documents was proved to be more effective. Currently, researchers are experimenting with user input, in utilizing search engines as well as social-semantic software, for the development of context. One example is integrating user input into ranking algorithms and ontologies development, experimented in several semantic and socialsemantic search engines. Yet another example is Genre-based navigation on the Web (Chandler,

544

1997) that is a powerful tool for the definition of users’ need (Askehave & Nielsen, 2005Chandler, 1997; Meyer & Stein, 2004) and personalization of search results (Maamar et al., 2004). Two complementary notions direct the course of thinking that is governing the suggestion made in this chapter. First is the understanding that learners, as end-users, need guidance as to how and when to best use social-semantic software. Yet another reason is aiming at improving the user experience, which could be comprehended as complementary to the course of research aiming at optimization of search engine (Bao et al., 2007), query automation (Xu & Croft, 2000), automation of metadata association (Hotho et al., 2006; Wu, 2006), and evaluation of alternative resources (Feng et al., 2005). The discussion of social-semantic applications would not be complete without addressing its pitfalls and its limitations. While the contribution of collaborative tagging for improving search results is widely acknowledged, “semantic locality” (Akavipat et al., 2006) and social-semantics are yet a source for serendipitous discoveries that are often of low quality (Wu et al., 2006). Indeed, the call for re-examining the use and integration of social-semantic software in favor of human judgment (Paynter et al., 2005) originate from several perspectives. The challenge for developers is continuously evolving, introducing questions such as: how can search engines use the query input to automatically enhance the next iteration, how to incorporate the metadata associated with social tagging to make search results more effective (Hotho, 2006) – as part of user-centered design as well as for the design of eLearning systems (Kritikou et al., 2007). Between Web 2.0 and the semantic Web, one possible solution could be found with Web X.0 as a result of advancement in Artificial Intelligence and interactive design. Yet another direction, which is the approach taken here (i.e., that does not contradict the former), would be enhancing the user experience using a prescriptive

ECHO

Figure 1. The ECHO model

framework as a basis for exploiting the semantic potential of the Web.

the eCho Model Echo suggests a four-stage methodology, which instructs the development of a three-layered framework for utilizing social-semantic software towards a systematic and effective integration of Web technologies as part of learning (Figure 1). Figure 1 demonstrates how a) the three perspectives on context and its modeling procedures inform b) the methodology that guides the development of three aspects of context as part of c) a three-layered framework. The key steps in the Echo methodology are: 1) analysis, 2) conceptual design, 3) structural design and 4) documentation and evaluation. In applying Echo we follow the three perspectives on context. First, we identify the task and consider the content elements to be included in the inquiry framework. Next we commence an iterative activity of analysis and knowledge acquisition focusing on the various aspects of the media and our task. We then continue with conceptual design, modeling the various content types based on our understanding of the domain. This, in turn, leads towards structural design. Documentation and evaluation occur iteratively throughout.

Analysis Analysis involves the manipulation of content analysis techniques for the purpose of determining the focus of the investigation (i.e., in order to obtain meaningful information from the Web). Of the three aspects of context (Fig. 1), this stage is focused on search terms, links and their annotation that are associated with the syntax layer. Analysis, as an iterative activity, is a composition of several tasks such as: 1) knowledge acquisition – the identification of relevant terms and appropriate resources; 2) classification – the aggregation of terms into categories to create clusters. These two activities are basis not only for knowledge acquisition, but also for the development of a concept map and a corresponding collection of tags (i.e., metadata) to be used for the annotation of relevant resources; 3) specification – of media- and content-based genres to be considered as part of a query formulation; and 4) knowledge organization and representation – using socialsemantic applications. Several classification techniques are used for content analysis. These are: thesaurus, faceted analysis, WH questions and genre-based classification. Thesaurus construction guidelines and the notions of wider term, narrower term and related terms support preliminary content analysis and

545

ECHO

classification. Knowledge acquisition for this purpose can benefit from using search engines that support semantic clustering, categorization and visualization (e.g., clusty.com; freebase. com. Kartoo.com, respectively). Faceted analysis technique is used for determining the aspects of a specific domain in context. Beyond semantic search engines, initiatives such as WikiMindMap. org that presents Wiki-based ontologies can help determine subject facets once employed parallel to the investigation. WH questions are recommended for analyzing relevant aspects of the subject domain. Last but not least, genre-based classification guides a three faceted classification distinguishing content from media (e.g., blog, Wiki) and genre-based media (e.g., video) from content-based media (e.g., news). There is in each of these techniques to contribute to the enhancement of the search process. The three deliverables of this stage are: 1) a folksonomy facilitated by a bookmark management system – holding links and the associated metadata; 2) RSS reader – referencing dynamic updates on selected content; 3) a preliminary concept map – a roadmap for knowledge representation. Bookmark management systems vary by their semantic attributes. Example advanced semantic features are: the management of association between tags, indicating relations between selected tags of the folksonomy (http://www. Bibsonomy.org) and recently contextualized folksonomies enabled using RDF, as in (http://www. Twine.com). There are also several options for the conceptualization and representation of a concept map (e.g., freemind.sourceforge.net).

Conceptual design Conceptual design involves the modeling of the information acquired through the analysis stage and its structuring in a more formal approach. Of the three aspects of context (Fig. 1), this stage is focused on the information and knowledge that are the objects of the inquiry. Conceptual design

546

is a composition of several tasks such as: 1) ontology development – based on a descriptive and prescriptive representation of the knowledge acquired through the analysis stage and included in the concept map; 2) the definition of an adapted search engine; and 3) modification of previously developed deliverables. Developing ontology means taking concepts of the previously defined concept map one step further from descriptive to prescriptive representation to include, for instance, semantic relationships between concepts and prescriptive representation. An adapted search engine is a dynamically evolving application that builds on the information (resources as links and metadata as refinements) acquired through the previous stage. The techniques that guide these activities are classification techniques aiming to define and categorize resources and assign tags and labels as part of the annotation of metadata. Specifically, genre-based classification is helpful at this stage as a basis for knowledge acquisition and for knowledge representation as well as for the refinement of previous queries. There are two deliverables to this stage: ontology and an adapted search-engine. In this context we also mention the knowledge of the learner now enriched with knowledge of the domain and proficiency in using classification techniques, search technology and social-media (i.e., Web 2.0 tools). In this sense conceptual design is about the modeling of a knowledgebase as part of shaping cognition, i.e., learning. As in the case of the previously mentioned deliverables, the development of adapted search engine and ontology contributes to knowledge evolution and context-aware learning.

structural design Structural design is aiming at two procedures. First, the various applications developed so far are integrated for representation using Mashup. Next, the information obtained so far is structured

ECHO

as part of the Mashup interface (i.e., using the tab option) to represent facets (i.e., upper-level concepts) of the domain. Of the three aspects of context (Fig. 1), this stage is about the pragmatic layer, focusing on the integration of social-semantic media and on the integrated representation of knowledge, information and metadata. Assigning tabs as part of the Mashup interface (i.e., using iGoogle) follows and represents core concepts identified for each subject domain (i.e., learning project). Finally, other relevant applications (i.e., gadgets) that may serve the learner goals are included as well. Most significantly, the benefit of this stage is in facilitating a kind of an organizational memory (OM; Weinberger et al., 2008b) that forms the basis for future knowledge evolution. Hence, this OM is acting also as a collaboration mechanism (Ford, 1996) also as basis for developing Learning Objects (Weinberger, 2008c). In the context of integration we mention a required proficiency that is recognizing best practices of using search engines. This means identifying search engines types and expertise. One way of doing so is classifying search engines either by Web generations (i.e., Web 2.0, Web 3.0 etc.), representation (i.e., visual), genre (i.e., scholar), focal point (i.e., geospatial, social, 3D) or else by the level of facilitating user contribution (e.g., Freebase.com, Feedmil.com).

documentation and evaluation Documentation is imminent to Echo. Tied to the discussion of Echo is the description of the various systems and techniques used for documentation. Examples are the information incorporated as part of the bookmarks system and the corresponding metadata included in the adapted search engines, Folksonomy, concept map, ontology and Mashup applications. Evaluation in Echo is composed of two procedures. First is the assessment of content and second

is the evaluation of usefulness. Evaluation of content, which aims at the assessment of conceptual coverage, is applied through design following evaluation conventions of Information Systems (Zachman, 1987), conceptual modeling and ontologies (Frank, 2007; Weinberger et al., 2008a), using criteria such as completeness, consistency, coherence and extendibility (Gruber, 1995). Evaluation of usefulness aims at extracting utility of the suggested methodology based on its feasibility. Albeit beyond the scope of this chapter (which is restricted to the introduction of preliminary experiences), evaluation is performed based on user feedback and involves a quantitative and qualitative approach using questionnaires and interviews with participants.

web-bAsed leArning: findings froM eMPiriCAl investigAtion Herein we report on our experience with an ongoing case study designed to test the feasibility of Echo as part of teaching introductory courses on Web technologies for undergraduate students from the department of Instructional Systems Technologies System at our institution. In implementing Echo we followed the constructivist approach to learning by adopting action research methods throughout the development process. Doing so we also adhere to the call made in Information Systems research for researchers to make their research more relevant to practice (Baskerville & Myers, 2004). Several examples are used to demonstrate several manipulations of Echo as part of a students’ project. Each project commenced with assigning a domain specific scenario rooted in real-life experience to a group of students. We follow Baskerville and Myers (2004) for the definition of a case study, and the guidelines for its implementation. To this end, we selected a diverse set of scenarios that could each demonstrate the application of the Echo approach. This research

547

ECHO

involves different groups of students, while for each we uniformly applied the aforementioned methodology. Since we reference an on-going research this chapter is limited to a description and a discussion of the findings regarding the feasibility of Echo based on our experience, without addressing the evaluation of the findings.

Analysis: establishing the syntax layer Analysis involves three parallel courses of 1) studying and applying content analysis techniques, 2) acquiring domain specific knowledge and 3) learning the media. The former two activities evolves parallel to the development of context, aiming at the conceptualization and modeling of domain knowledge for the purpose of establishing the syntax layer,. For instance, the definition of a domain-specific taxonomy (i.e., folksonomy) instructs the allocation of links to include as part of the bookmarks and vice versa. Without the taxonomic commitment for the concept map knowledge acquisition could become arbitrary. Without the concept map, the focus of the analysis process could be distracted. The suggested analysis techniques are complementary. Thesaurus conventions are used for the initial orientation in a new domain. Faceted analysis is implemented for the identification of categories, once initial knowledge is established and WH questions are used for the evaluation of conceptual coverage. Genre-based classification is used for knowledge acquisition and even more for knowledge annotation. One example is the development of a collaborative (i.e., group account) bookmarks management system using Bibsonomy (Web: http://www.bibsonomy.org). There, for the description of each link we used genre-based classification to describe issues such as: language, purpose and origin. Specifically, we used Bibsonomy for the benefit of Echo users (http://www.

548

bibsonomy.org/user/echo). Google Bookmarks was also used as a bookmark system and Google Reader was used for syndication. For either one of the social-semantic applications, participants maintained consistency in their use of metadata as part of knowledge organization and annotation. In this case, consistency serves as an internal validation mechanism, as part of design.

Conceptual design: establishing the semantic layer The focus of conceptual design is making context explicit by attending to knowledge representation, i.e., metadata. One example is a concept map developed for the subject of accessibility. This concept map is developed also as basis for ontology development – an effort that is yet underway (i.e., using Protégé`). There, the higher-level concepts are: assistive technology, disability and community. The first represents a spectrum of aids and methods, the second represents sensorial or mental disabilities and the third is focused on social aspects. There is in this initial classification to denote a comprehensive approach to investigation of conceptual coverage. From the context perspective, the development of prescriptive, rather than a descriptive, domain model supports semantic alignment between applications (i.e., the annotation of tags, refinements, folders and tabs). Considering the cognitive perspective, learning evolves through design, as the learner acquires knowledge and practice. Learning and design are informed also based on embedding the adapted SE (or several of these) and ontology. Both these applications also require continuous semantic modeling effort, which in turn affects also the previously developed applications. Other example projects were developed for emerging domains such as: m-Learning – including facets such as platforms, facilities, activities and learning types, and serious games – including game environment, technology, game type,

ECHO

learning and collaboration, as well as issues such as: Web 2.0, graphical design, do-it-yourself, and social and ecological aspects of the 21st century.

structural design: establishing the Pragmatic layer Considering the task perspective (Fig. 1), Mashup facilitates the integration of content to form the pragmatic layer. From the content perspective we note that Mashup references both the object of the inquiry (i.e., content and context) and the communication means (i.e., social-semantic applications). Considering the user perspective, Mashup facilitates an effective interface providing for content management, maintenance and evolution. Along this lane, the negotiation of content as part of using either the adapted SE or any other application provides new insights which in turn affect content evolution, while the application becomes a means for learning. Considering the previous example, not only can new knowledge instruct content development, it may also inform design as in the case of adding applications. One example is integrating games and simulations as part of the accessibility project that allows users to experience with innovative devices. As in the case of an Organizational Memory, Mashup facilitates the continuous introduction of new content. Specifically, content development is based on using the adapted SE and updating other applications. As advised by the aforementioned iterative lifecycle, the learning experience is instructed by an iterative procession of analysis and conceptual design in which the learner is responsible for the maintenance of the content applications. This includes tasks such as omitting, weeding out less effective information or adding new knowledge.

fUtUre reseArCh direCtions Several directions for further research can be instructed based on the suggestion made in this chapter. We use Web generations and areas of applications as lens to examine these suggestions. The viability of the paradigm of Web-based learning, which motivated Echo, could be further investigated in empirical research, assessing usefulness and utility. Empirical research could also follow attributes of Web 2.0 and the establishment of a virtual campus (i.e., knowledgebase) of Learning Objects. The notion of social-semantic software and applications, as part of implementation issues of the proposed model, could be further developed as part of design, emphasizing learning and interaction processes – beyond collaboration and information retrieval. There are several directions of research and action to be considered. One way could be directing Web 2.0 and Web 3.0 towards a more powerful Web X.0 that includes characteristics of virtual worlds, gaming and simulations. Yet another direction, which does not contradict the former, could be advising the development of the user model as means for contextualization. Last but not least is the development of tools and techniques for enhanced utilization of additional Web technologies, for instance, by advising specific use cases. At the intersection between technological aspects and application areas, we see promise in emerging interdisciplinary design approaches as basis for assimilation. From the technological perspective, this could be considering the user model and the user task as part of design. From the user perspective, this would be challenging cognition and providing the user with a richer, active and viable interface. In order for these technologies and their applications to be more relevant to practice there is a need to provide users with a unified framework that is prescriptive enough to direct action and

549

ECHO

descriptive enough to allow its adaptation in various settings.

ConClUsion In this chapter we suggested Echo, a model and a methodology for the design of a context-aware learning experience. There are three contributions in this chapter, responding to the three challenges presented for this research: 1) a methodology to guide the learning process, 2) techniques to support content analysis and modeling (i.e., conceptual and structural design) and 3) a three-layered framework of social-semantic applications to be incorporated as part of the advised methodology. Echo guides learners in exploiting principles of knowledge organization systems such as thesaurus, concept map and ontology, through processes of content analysis and context modeling in a way which instructs context-aware learning experience. The offer made in this chapter follows an interdisciplinary approach, bringing together theories and methods of Information Systems, eLearning, Information Science and Conceptual Modeling to suggest an integrated approach for the design of Web-based context-aware learning experience. Based on previous work in related areas, the suggested model also attends to the three perspectives of context – the task, the user and the content perspectives and to the three aspects of context – the syntax, the semantic and the pragmatic aspects. Following the Echo model and its methodology, the three perspectives of context inform the design of a three-layered framework that evolves from the syntax layer through the semantic layer to the pragmatic layer. There are four-stages in the suggested methodology. For each stage we have defined the tasks, activities and deliverables.

550

For each activity we have assigned the methods to be applied and the social-semantic application to be used. The feasibility of ECHO and its usefulness are established based on preliminary findings of empirical investigation attending to the various aspects of Web-based learning. Researchers can find this model inspiring for future development by considering this model as part of the design of Web 4.0 intelligent personal agent or Web X.0 interactive experience. Practitioners could integrate Echo into academic and enterprise setting, thus guiding users towards a meaningful experience of Web-based learning.

ACKnowledgMent This paper builds on my experiences in recent years with several classes at the Department of Instructional Systems Technologies at HIT, Israel. I wish to thank all participating students for their motivation and cooperation. The students whose work is specifically mentioned in this chapter are: Bar-Sheshet Ran, Almog Adva, Dagan Assaf, Guzikov Oleg, Raby Keren and Talmor Ran. I am also grateful for the constructive comments of the anonymous reviewers. Special thanks to Ariel J. Frank, Department of Computer Science, Bar- Ilan University, for fruitful discussions, to Shirley Goldrei, Sydney, Australia and to Hagar Weinberger for their careful reading and constructive comments. Last but not least I wish to thank my family for their encouragement and support.

referenCes Abel, M. H., Benayache, A., Lenne, D., Moulin, C., Barry, C., & Chaput, B. (2004). Ontology-based organizational memory for e-learning. Educational Technology & Society, 7(4), 98–111.

ECHO

Akavipat, R., Wu, L.-S., Mencer, F., & Maguitman, A. G. (2006). Emerging semantic communities in peer Web search. P2PITO6, Proceedings of the International Workshop on Information Retrieval in Peer-to-Peer Networks (pp. 1-8). Arlington, VA. Al- Kahlifa. H. S., & Davies, H. C. (2007). Delicious learning resources. In Iskander (Ed.), Innovations in e-learning, instruction technology, assessment, and engineering education (pp.139143). Berlin-Heidelberg: Springer. Alavi, M., & Leidner, D. (2001). Review: Knowledge management and knowledge management systems: Conceptual foundations and research issues. MIS Quarterly, 25(1), 107–136. doi:10.2307/3250961 Argrys, C., & Schon, D. (1978). Organizational learning: A theory of action perspective. Reading, MA: Addison Wesley. Askehave, I., & Nielsen, A. E. (2005). Digital genres: A challenge to traditional genre theory. Information Technology & People, 18(2), 120–141. doi:10.1108/09593840510601504 Bao, S., Wu, X., Fei, B., Xue, G., Su, Z., & Yu, Y. (2007). Optimizing Web search using social annotations. International World Wide Web Conference (IW3C2), WWW 2007, Banf, Alberta Canada. Baskerville, R., & Myers, M. D. (2004). Special issue on action research in information systems: Making IS research relevant to practice-forward. MIS Quarterly, 28(3), 329–335. Brashnik, O. (2007). Database and the geometry of knowledge. Data & Knowledge Engineering, 61(2), 207–227. doi:10.1016/j. datak.2006.05.005 Brezillon, J., & Brezillon, P. (2006). Context modeling: Context as dressing of a focus. B. Kokinov, et al. (Eds.), CONTEXT 2007. (LANI 4635, pp.136-149).

Chandler, D. (1997). An introduction to genre theory. Retrieved on January 2, 2008, from http:// www.aber.ac.uk/media/Documents/intgenre/ intgenre.html Christiaens, S. (2007). Metadata mechanisms: From ontology to folksonomy…and back. In R. Meersman, Z. Tari & P. Herrero (Eds.), On the Move to Meaningful Internet Systems 2006: OTM 2006 Workshops (pp.199-207). Downes, S. (2005) E-learning 2.0. ACM eLearn Magazine, 10. New York. Retrieved on December 26, 2007, from http://elearnmag.org/subpage.cfm ?section=articles&article=29-1 Ebner, M. (2007). E-learning 2.0= e-Learning 1.0+Web 2.0? The Second International Conference on Availability, Reliability, and Security (ARES’07) (pp. 1235-1239), Vienna, Austria. Falconer, L. (2006). Organizational learning, tacit information, and e-learning: A review. The Learning Organization, 13(2/3), 140–152. doi:10.1108/09696470610645476 Feng, L., Jeusfeld, M. A., & Hoppenbrouwers, J. (2005). Beyond information searching and browsing: Acquiring knowledge from digital libraries. Information Processing & Management, 41, 97–120. doi:10.1016/j.ipm.2004.04.005 Ford, C. M. (1996). A theory of individual creative action in multiple social domains. Academy of Management Review, 2(4), 1112–1142. doi:10.2307/259166 Frank, U. (2007). Evaluation of reference models. In P. Fettke & P. Loos (Eds.), Reference modeling for business systems analysis (pp. 118-140). Hershey, PA: IGI Global. Gasevic, D., Jovanovic, J., Devedzie, V., & Boskovie, M. (2005). Ontologies for reusing learning object content. Proc. IEEE International Conference on Advanced Learning Technologies (ICALT’05).

551

ECHO

Grace, A., & Butler, T. (2005). Learning management systems: A new beginning in the management of learning and knowledge. International J. of Knowledge and Learning, 1(1/2), 12–24. doi:10.1504/IJKL.2005.006248 Gruber, T. R. (1995). Toward principles for the design of ontologies used for knowledge sharing. International Journal of Human-Computer Studies, 43(4-5), 907–928. doi:10.1006/ijhc.1995.1081 Hjorland, B. (1997). Information seeking and subject representation: An activity-theoretical approach to information science. London: Greenwood Press. Hotho, A., Jaschke, R., Schmitz, C., & Stumme, G. (2006). Bibsonomy: A social bookmark and publication sharing system. In A. de Moor, S. Polovina & H. Delugach (Eds.), Proceedings of the Conceptual Structures Tool Interoperability Workshop at the 14th International Conference on Conceptual Structures, Aalborg University Press, Aalborg, Denmark. Hotho, A., Jaschke, R., Schmitz, C., & Stumme, G. (2006b) Emergent semantics in bibsonomy. Proceedings of Workshop on Applications of Semantics, Knowledge & Data Engineering Group. Retrieved on April 21, 2006, from kde. cs.uni-kassel.de Jovanovic, J., Gasevic, D., Knight, C., & Richards, G. (2007). Ontologies for effective use of context in e-learning settings. Educational Technology & Society, 10(3), 47–59. Kings, N. J., Gale, C., & Davies, J. (2007). Knowledge sharing on the Semantic Web. In E. Franconi, M. Kifer & W. May (Eds.), 4th European Semantic Web Conference, ESWC. (LNCS 4519, pp. 281295). Berlin Heidelberg: Springer-Verlag. Klein, H. K., & Myers, M. D. (1999). A set of principles for conducting and evaluating interpretive field studies in information systems. MIS Quarterly, 23(1), 67–94. doi:10.2307/249410

552

Kraines, S., Weisen, G., Brian, K., & Yutaka, N. (2006). EKOSS: A knowledge-user centered approach to knowledge sharing, discovery, and integration on the Semantic Web. In the Fifth International Semantic Web Conference, ISWC. (LNCS, pp.42-73). Berlin: Springer. Kritikou, Y., Demestichas, P., Adampoulou, E., Demestichas, K., Theologou, M., & Paradia, M. (2007). User profile modeling in the context of Web-based learning management system. Journal of Network and Computer Applications. doi:. doi:10.1016/j.jnca2007.11.006 Kwan, M., & Balasubramanian, P. R. (2003). KnolwedgeScope: Managing knowledge in context. Decision Support Systems, 35(4), 467–486. doi:10.1016/S0167-9236(02)00126-4 Latham, D. (2002). Information architecture: Notes toward a new curriculum. Journal of the American Society for Information Science and Technology, 53(10), 824–830. doi:10.1002/asi.10097 Maamar, Z. ALKhatib, G., Mostefaoui, S. K., Lahkim, M., & Mansoor, W. (2004). Context-based personalization of Web services composition and provisioning. Proc. EUROMICRO. IEEE Computer Society. Meyer, S., & Stein, B. (2004). Genre classification of Web pages user study and feasibility analysis. In S. Biundo, T. Fruhwirth & G. Palm (Eds.), German Conference on Artificial Intelligence, KI 2004. (LNAI 3238, pp. 256-269). Berlin: SpringerVerlag. Mika, P. (2005). Ontologies are us: A unified model of social networks and semantics. In Y. Gil, et al. (Eds.), International Semantic Web Conference, ISWC. (LANCS 3729, pp.522-536). Mullaholand, P., Zdeahal, Z., Domingue, J., & Hatal, M. (2001). A methodological approach to supporting organizational learning. International Journal of Human-Computer Studies, 55, 337–367. doi:10.1006/ijhc.2001.0494

ECHO

Muresan, G., & Harper, D. J. (2004). Topic modeling for mediated access to very large document collections. Journal of the American Society for Information Science and Technology, 55(10), 892–910. doi:10.1002/asi.20034 Nonaka, H. T., & Takeuchi, H. (1995). The knowledge-creating company, now Japanese companies create the dynamics of innovation. New York: Oxford University Press. O`Hear, S. (2006). ReadWriteWeb. Retrieved on March 24, 2007, from http://www.readwriteweb. com/archives/e-learning_20.php Paavola, S., Lipponen, L., & Hakkarainen, K. (2004). Models of innovative knowledge communities and three metaphors on learning. Review of Educational Research, 74(4), 557–576. doi:10.3102/00346543074004557 Paramaswaran, M., & Whinston, A. B. (2007). Research issues in social computing. Journal of the Association for Information Systems, 8(6), 336–350. Paynter, G. W. (2005). Developing practical automatic metadata assignment and evaluation tools for Internet resources. Joint Conference on Digital Libraries, Jcdl’05, Denver, CO. Phaedra, M., & Mohan, P. (2007). Contextualizing learning objects using ontologies. Computational Intelligence, 23(3). Rafaeli, S., & Raban, D. (2005). Information sharing online: A research challenge. International Journal of Knowledge and Learning, 1(1/2), 62–79. doi:10.1504/IJKL.2005.006251 Ramakrishnan, R., & Tomkins, A. (2007). Toward a people Web. [IEEE.]. Computer, 63–72. doi:10.1109/MC.2007.294 Sa’nchez-Alonso, S., & Vovides, Y. (2007). Integration of metacognitive skills in the design of learning objects. Computers in Human Behavior, 23(6).

Schmidt, A. (2005). Bridging the gap between knowledge management and e-learning with context-aware corporate learning. In WM 2005: Professional Knowledge Management-Experiences and Visions, 3rd Conference Professional Knowledge Management-Experiences and Visions (pp.170-175), Kaiserslautern, Germany. DFKI. Schmidt, A. (2008). Enabling learning on demand in semantic work environments: The learning in process approach. In J. Rech, B. Decker & E. Ras (Eds.), Emerging technologies for semantic work environments: Techniques, methods, and applications (pp.21-28). Hershey, PA: IGI Publishing. Sicilia, M. A., Lytras, M., Rodriguez, E., & GarciaBarriocanal, E. (2006). Integrating descriptions of knowledge management learning activities into large ontological structures: A case study. Data & Knowledge Engineering, 57, 111–121. doi:10.1016/j.datak.2005.04.001 Studer, R. (2006). Semantic Web: Customers and suppliers. Invited talk. The 5th International Semantic Web Conference (ISWC2006), Athens, GA. (LNCS, 4273). Berlin/Heidelberg: Springer. Te`eni, D. (2001). Review: A cognitive-affective model of organizational communication for designing IT. MIS Quarterly, 25(2), 251–312. doi:10.2307/3250931 Tzitzikas, Y., Christophides, V., Flouris, G., Kotzinos, D., Markkanen, H., Plexousakis, D., & Spyratos, N. (2006). Trialogical e-learning and emergent knowledge artifacts. In Innovative Approaches for Learning and Knowledge Sharing, Proc. First European Conference on Technology Enhanced Learning, EC-TEL, Crete, Greece. (LNCS). Berlin/Heidelberg: Springer. Veres, C. (2006). The semantics of folksonomies: The meaning in social tagging. In Proc. 12th American Conference on Information Systems, Mexico.

553

ECHO

Wang, Q., Quing, L., & Schang, Y. (2003). Collaborative knowledge management in the extended enterprise: Supported by an information portal. International Conference on Systems, Man and Cybernetics, IEEE, 516–521. Weiling, K. (2006). Organizational learning processes: Its antecedents and consequences in enterprise system implementation. Journal of Global Information Management, 14(1), 1–22. Weinberger, H. (2008c). WELL: Web-based learner library for e-learning 2.0. In G. Papadopoulos & R. Williams (Eds.), Proc. the 7th European Conference on E-Learning (ECEL), Agia Napa, Cyprus. Reading, UK: Academic Conferences. Weinberger, H., Te`eni, D., & Frank, A. J. (2008a). KnowledgeEco: Ontology for the domain of OM. In P. Rittgen (Ed.), Handbook of ontologies for business interaction. Hershey, PA: IGI Global. Weinberger, H., Te`eni, D., & Frank, A. J. (2008b). Ontology-based evaluation of organizational memory. Journal of the American Society for Information Science and Technology, 59(9), 1454–1469. doi:10.1002/asi.20859 Wu, X., Zhang, L., & Yu, Y. (2006). Exploring social annotations for the Semantic Web. The International World Wide Web Conference Committee, IW3C2, WWW (pp. 417-426), Edinburgh, Scotland. Xia, L., Beaudoin, J. A., Bui, Y., & Desai, K. (2006). Exploring characteristics of social classification. In J. Funer & J. Tennis (Eds.), Advances in classification research, the 17th ASIS&T SIG/ CR Classification Research Workshop. Xu, J., & Croft, B. W. (2000). Improving the effectiveness of information retrieval with local analysis. ACM Transactions on Information Systems, 18(1), 79–112. doi:10.1145/333135.333138 Zachman, J. A. (1987). A framework for information system architecture. IBM Systems Journal, 38(2&3), 454–470.

554

Zimmerman, A., Lorenz, A., & Oppermann, R. (2007). An operational definition of context. In B. Kokinov, et al. (Eds.), CONTEXT 2007. (LANI 4635, pp. 558-571).

Key terMs And definitions Action Research: IS research paradigm encouraging participation between researchers and participants. Conceptual Design: Modeling (i.e., using content analysis methods) of information knowledge in a subject domain (e.g., using folksonomy, concept map or ontology) and its structuring in a more formal approach. Content Analysis: Applying a series of techniques for the identification of core concepts in a subject domain as basis for of domain modeling. This could be done using KOS methods as well as facet analysis and genre-based classification. Context-Aware Learning: Establishing learning based on descriptive or prescriptive representation of a subject domain. ECHO: A model for the design of a threelayered framework that is guiding context-aware learning experience on the Web. Knowledge Organization System (KOS): a means for knowledge management and knowledge representation by specific method, such as: thesaurus, Taxonomy, Folksonomy and Ontology: KOS can be applied independently (logically) or as part of social-media software. Learning Object: Domain-specific or task specific knowledge aggregated using socialsemantic software that is the result of individual or collaborative learning. Mashup: Web application hybrid, is an architecture using AJAX (Asynchronous Java Script and XML) allowing the integration of different content types of various digital genres Metadata: Data assigned for the description of information and knowledge. Social-semantic

ECHO

software uses several types of metadata such as: tags, labels, folders and tags. Ontology: Structured representation of conceptual model. Organizational Memory: The memory of an organization. Social-Semantic Software: Applications that are designed to enable the development, maintenance and evolution of semantically enabled collaborative knowledge management, such as: Bookmark management system, Folksonomy, concept map, ontology and Mashup; also known

as web 2.0 tools or Web 3.0 social-semantic technologies. Web-Based Learning: The manipulation of a set of content analysis techniques aiming to establish a conceptual model of a task specific domain.

endnote 1

Corresponding author: mail to: hadasw@ hit.ac.il

555

556

Chapter 31

Advancing Learning Through Virtual Worlds Steve Mahaley Duke Corporate Education, USA Robin Teigland Stockholm School of Economics, Sweden

AbstrACt Higher education institutions and corporations are increasingly exploring new pedagogical methods to align with learning styles of incoming students and employees, who are amazingly adept at using Web 2.0 applications. This chapter explores the use of virtual worlds, in particular that of Second Life, in educational activities by organizations such as higher education institutions or corporations. We begin by introducing virtual worlds with a particular focus on Second Life. We then provide an overview of the beneits of this environment for learning activities before presenting a set of potential learning activities that can be conducted within Second Life. We then discuss an in-depth example of 3D teaming-one learning activity within Second Life conducted by the authors. After a discussion of implementation challenges, we then present areas for future research.

introdUCtion To learn effectively we need not only to experience but also to be able to share our experience with others. In education institutions this has traditionally meant listening (to a talking head in front of the class), reading assigned texts, and communicating what has been learned by answering some pre-defined questions. A more recent view of learning adds to both the experience and the communicating aspects DOI: 10.4018/978-1-60566-384-5.ch031

of learning. In this view, more emphasis is placed on experiences where students discover, are involved in, and are exposed in different ways to the topic at hand. Communication is redefined so that not only is it recognized as a means for repeating facts and information but also as a means for reflection and “building” wisdom. Learning is recognized as acquired know-how and skills, changes in attitudes, new theories, and/or new ways of thinking. This more recent view of learning, however, leads to a number of new opportunities and challenges faced by both teachers and students.

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Advancing Learning Through Virtual Worlds

Meanwhile, the continuous development of internet-enabled communication technologies has resulted in the rapid growth of online activities, such as individuals making new friends through sharing personal profiles (e.g., Facebook, LinkedIn), tracking one another through microblogging (e.g., Twitter), exchanging multimedia files (e.g., YouTube), co-creating content (e.g., blogs, wikis), and collaborating through virtual worlds (e.g., Second Life). Individuals are increasingly using these web applications, grouped under the umbrella of Web 2.0, in their private lives for creating and maintaining social networks and discussing common hobbies and social interests with others across the globe (Hustad & Teigland 2008). In particular, students at higher education institutions are highly adept at using these communications technologies, with many web applications (e.g., Facebook) even stemming from university students themselves in response to changing communication behaviors. As a result, the borders between work, play, and learning are dissolving as the demands of the “virtual gaming” generation are fundamentally

changing how and where work gets done (Beck & Wade 2006, Johnson 2006). Proserpio&Gioia (2007) argue that these social and technical changes in the wider environment are so profound that educators need to explore their pedagogical implications and account for these changes. If educators ignore these implications, then we will see a growing misalignment between our current education world and the technology-rich world of the younger generations. Thus, the purpose of this chapter is to explore the educational use of one web 2.0 application – that of virtual worlds and in particular the virtual world of Second Life -- as a means to align teaching and learning styles in education (figure 1). This chapter is organized as follows. After an introduction to virtual worlds and in particular Second Life, we provide an overview of the benefits of this environment for learning activities as well as a set of potential learning activities that can be conducted within Second Life. We then discuss an in-depth example of 3D teaming - one learning activity within Second Life conducted by the authors. After a discussion of implementa-

Figure 1. Alignment between teaching and learning styles (Proserpio&Gioia 2007)

557

Advancing Learning Through Virtual Worlds

tion challenges, we then present areas for future research.

An introdUCtion to virtUAl worlds And seCond life Virtual worlds are becoming increasingly sophisticated, enabling organizations and individuals to “step into the internet.” A virtual world is a computer-based, simulated environment where individuals assume a virtual identity called an avatar. Avatars inhabit the virtual worlds and interact with each other via computer-based chat, or more recently, voice. Virtual worlds are common in multiplayer online games (such as Citypixel), virtual environments (such as Second Life), and role-playing games (such as Lineage and World of Warcraft). Due to increasing broadband internet access, virtual worlds are rapidly emerging as an alternative means to the real world for communicating, collaborating, and organizing activity. For example, in the virtual world Second Life, more than 50 multinational organizations, such as Accenture, IBM, and Unilever conduct operationsand Anshe Chung, the avatar for a Chinese-born woman living in Germany with employees in China, became the first USD millionaire resulting from her virtual real estate activities.1 Furthermore, companies such as Forterra Systems, Protonmedia, and Qwaq provide Fortune 500 companies like Johnson & Johnson, Novartis, Motorola, Intel withcompletely secure, private virtual business worlds in which to collaborate and conduct economic activities. Gartner Group further predicts that by 2012, 80% of all active internet users and Fortune 500 enterprises will have an avatar or a presence in a virtual world.2 In this chapter we have chosen to highlight one virtual world, Second Life, since it has enjoyed exponential global growth over the last severalyears both in terms of membership and economic activity. Second Life is a three dimensional platformthat is visualized graphically in

558

which individuals are represented by avatars that interact with each other and their environments (see figure 2). In March 2009,Second Life reported that there were1.4 million residents or avatars logging in during the previous 60 days, and more than 1100 individuals earning more than USD 1,000 in profit monthly through their online activities3. The members have a mean age of 32 years, are 42% female, and 55% come from outside of North America.4 The success of Second Life is due in large part to the fundamental design principle that all user-created content in Second Life remains the intellectual property of the creator and is protected as all other forms of intellectual property. This has generated a dynamic and growing market for goods and services in Second Life, such as clothing and accessories for the avatars, buildings and home furnishings for avatars’ property, and transportation vehicles to name a few (see the website http://xstreetsl.com/ for insights on the types of items being developed and sold). The virtual world design concept is based on geographic space.Avatars travel in Second Life essentially in the same fashion as in the real world (although avatars can fly and teleporting is possible), and users interact with other avatars as an essential element of this virtual world.For example, users are not simply shopping online but are in a store with sales reps and other avatars (Wasko et al 2007). Finally, Second Life has its own currency, the Linden Dollar, which is the primary currency used to purchase Second Life items and is exchangeable to US Dollars (appx. $270L = $1US). Virtual world development is still in its infancy, and we are just beginning to explore and understand how activities in a virtual world enhance or replace real world economic, social, and educational activities. A pressing need to develop an understanding of emerging virtual world dynamics exists, and there are many potential organizational, ethical, legal and other issues involved, In an attempt to develop this understanding, we investigate the benefits that the use of Second Life may bring to higher education below.

Advancing Learning Through Virtual Worlds

Figure 2. North Carolina State University, College of Design5

benefits of virtUAl worlds While the views are mixed regarding the performance of marketing and public relation activities of corporations in virtual worlds, one of the most promising areas for virtual world development is as an arena for learning activities within organizations and educational institutions. A quick look at the Second Life (SL) educational wikiconfirms this as it shows that there are hundreds of educational institutions from across the globe active in SL with numerous institutions even owning their own private island. Pedagogically speaking though, for what purposes can we use virtual worlds such as Second Life? At first glance, virtual worlds do provide an alternative meeting place for groups, and there are many examples of academic and corporate institutions developing campuses in Second Life to provide these immersive meeting places, often as visual replicas of real locations.From a learning standpoint, however, it is important to understand what virtual worlds offer for the learners that is either an extension of existing practices and experiences or completely new experiences that can only be created in the virtual space. Below we discuss some of the unique qualities of virtual worlds, and

how they can be used to create new and effective learning experiences for organizations. The first step is to understand some basics about these environments, and what advantages they provide over conventional technologies.The first key differentiator is that while virtual worlds, like online games such as World of Warcraft or A Tale in the Desert, are persistent environments, they differ in that there is no prescribed script or set of objectives for the participants in these spaces. Additionally, the participants in Second Life are referred to as “residents” rather than “players”, hinting at something deeper than what you might find in game-based environments. We will return to this below. Related to this is that virtual worlds is provide tools for the residents to use in constructing content – often three-dimensional virtual objects such as buildings, landscapes, vehicles, clothing items, skin textures and waterfalls – just about anything that you can imagine. Many of these objects will contain scripts, little snippets of computer code that run animations of the object, play media files (such as sound or video) or otherwise enable the resident to do or experience something new. Finally, there are often tools that allow residents tocapture in video format the actions taking place in the virtual world – a type of media called machinima.6

559

Advancing Learning Through Virtual Worlds

Second, objects created in virtual worlds can be given away to other residents or even sold to other residents for the in-world currency. In Second Life, the currency is Linden Dollars, and what is really fascinating is that Linden dollars can be converted to real US dollars.7 For example, a resident in Second Life can choose to link a credit card to his or her account (there are free accounts, but many residents will opt for a paid account in order to buy virtual land) and with that can purchase in-world currency (in this case, Linden Dollars) to be used to buy anything from land, buildings, and housing to clothing and even stock in virtual companies. A third, and potentially the most important, differentiator is the avatar, or the three-dimensional representation of the resident (or participant), as mentioned above (figure 3). Whereas in gaming environments the choices are fairly limited as to character, appearance, and affiliation with groups, in many virtual worlds an individual can build his or her avatar almost entirely built from scratch. In Second Life residents have an expansive set of choices to make about their avatar, initially based on a male or female human form. From that initial choice, they may make countless adjustments to skin color (all of the spectrum is covered here), body style, clothing, hair length, etc. And once in-world (in Second Life), they may choose to spend their Lindens on additional clothing, accessories, tattoos, eyeglasses, etc. One may even

Figure 3. Avatars can take many forms

560

abandon the human form altogether and assume the appearance of an animal or even of objects such as a cloud. In addition, the choices for an avatar extend beyond body style and clothing. Once in Second Life, residents can choose (or be invited) to join or build clubs or groups. These affinity groups may range from groups providing services in-world to groups dedicated to a particular field of study or cause, both in-world and in real life. The result of all these choices related to the avatar’s presence in-world is that the real-life person becomes more and more deeply invested (personally, emotionally, and sometimes monetarily) in their in-world identity or identities. They are extending their personal selves into a three-dimensional online environment that provides affiliations, interactions, experiences, and relationships. This personal investment reinforces the notion that these participants are, in fact, residents of the virtual world. They very often build or purchase housing, and then rely, depend upon, seek out, and look forward to the interactions with their friends in this online space. The fourth and final differentiator is the beguilingly simple fact that virtual worlds are not constrained by real world physics (figure 4). While this may appear to be self-evident, it is worth consideration when imagining and planning what one can do in these virtual environments. For example, in Second Life, avatars when tired

Advancing Learning Through Virtual Worlds

of walking across the landscape can simply fly up into the virtual sky, soar to their next destination, float to observe the goings-on below, or even teleport to a predetermined location. Moreover, the “landscape” does not even need to be a landscape as we know it in the physical world. It can be just emptiness populated with objects, such as the planets and moons of our solar system, where a resident can ‘stand’ on the rings of a virtual Earth or Saturn. Or, consider the molecular scientist who wishes to demonstrate, in large, manipulative format a molecule otherwise difficult to see and nearly impossible to push around.8 While all of this may make our heads spin, it is important to consider the key differentiators from an educator’s perspective: 1. 2. 3.

4.

Residents / avatars have tools at their disposal to create objects and media Objects or environments created by residents persist for others to experience Individuals spend considerable time investing in their avatars and relationships (with individuals and groups) Real world physics need not apply

develoPing virtUAl world ACtivities Below we provide some thoughts on how educators may approach the design of virtual world learning activities as well as some examples.

initial design Process The design process of any educational activity should always begin with clearly articulated learning outcomes. For academic institutions these will be tied to the course objectives, and for corporate learning programs, these should be tied to the capabilities and skills that the target population needs in order to be successful in their jobs. Moreover, those capabilities and skills should be part of a larger plan to achieve the overall organizational strategy. Learning outcomes typically fall into three categories: 1) Knowledge – what we want the learners to know that is new and different, 2) Behaviors – what we want them to be able to do differently as a result of the experience, 3) Attitudes – what we want the participants to believe differently about themselves, their organization and the current environment.9

Figure 4. Virtual worlds are not constrained by real world physics

561

Advancing Learning Through Virtual Worlds

Once the outcomes are established, then we must understand the participant profile (including demographics, learning styles and motivation) and context (where they are geographically, their current work realities, technologies available to them, and how the learning fits into their day-today life). This is where we will begin to see how a virtual world experience can provide maximum benefit. As the design process continues, we aim to achieve a match between the learning outcomes, participant profile and context, and a virtual world environment and what it offers. At this point, the designer must know in detail what capabilities the particular platform has to offer. We have discussed Second Life above, but there are numerous other worlds available, including Olive (Forterra, Inc.), Protosphere (ProtonMedia, Inc.) and Wonderland (Sun Microsystems, Inc.). We will speak below of Second Life, as it is publicly available, and it provides a wide range of communication, group organization, and building tools.

A Continuum of learning experiences

2.

In either case of scripted or open access, the design of the in-world activities should be built upon the known competency frameworks and integrated with existing learning programs; the link between the virtual world activity and the knowledge, behaviors (skills), and attitudes needed at a given role in the organization or in the academic course needs to be made explicit, such that transfer of the learning ‘finds a home’ and is supported and rewarded. Below we provide some examples of scripted and open access.

examples of scripted Access •

Learning events created using Second Life can range from relatively passive roles for your audience to complete immersion. This range of tightly controlled and ‘scripted’ events to the less controlled and fully immersive ‘open access’ events gives educators a large palette of possibilities. Let’s think about these two basic types of applications: 1.

562

We produce an environment for our learners to experience, somewhat passively. In this scenario, we have essentially scripted what will happen and our learners have a ‘sensing’ role in this environment. They are to experience something we have orchestrated, and then explore the meaning of that in a guided debriefing session to follow. Let’s call this scripted access.

We produce an environment for our learners to use actively. In this scenario, we have thoughtfully created an environment in which our learners are to co-create content and meaning. They are to follow a set of guiding instructions, and they are largely left to their own devices to address the challenge at hand. Let’s call this open access.



Guided tours: Learners are taken through a guided tour of the virtual world to illustrate relevant points about this new environment, which leads to a relective dialogue among the participants about the potential impact this environment will have on their business, and what it means for their role in particular.For example,we have already conducted several guided tours of Second Life for clients for this very reason – to expose them to the virtual economy and generate dialogue around future business scenarios. Interviews: Learners are introduced to an in-world attendee who is interviewed by a host, followed by a question and answer period. Potential topics include virtual world business practices, hiring practices, economics, compensation, innovation, teaming, etc., with the two-fold focus on the virtual economy and the relections

Advancing Learning Through Virtual Worlds





on the experience of interacting with an avatar. This model could also be used to introduce participants to famous historical igures (from within the organization or from other realms). Examples of this include training by CNN of young reporters as well as school students learning about history through interviewing avatars representing prominent igures such as Abraham Lincoln. Acting a scenario: Learners assume a particular role in a pre-scripted scenario (metaphoric or realistic) in which they are to have an interaction with another person or persons. This could be developed to give learners the opportunity to practice certain roles and the interactions required, or to explore behaviors in a metaphoric environment that can be debriefed and linked to their real-world situation later. This could also be useful for creating an experience to highlight diversity issues (gender, position of authority, etc.) in which individuals may be asked to take on an avatar of a different race, gender, religion, etc. Modeling: Learners are provided with a three-dimensional simulation of an object (molecule, three-dimensional organization chart, concept map, solar system, machine, anything in between) that otherwise would be dificult to render or make available in the real world. The presenter can even manipulate the object in real time. This can be used in collaborative work on new product models, rendering a model of existing physical spaces or machinery for training purposes (e.g., training new store managers), or just as an effective mechanism for illustrating conceptual or operational models.

examples of open Access •

Identities: Learners are each given an assignment to create their avatars, and then











present themselves in-world to each other. Design and debrieing focuses on the choices they make, why, and what their experience in-world is like and what that can tell them about their real-life choices, values, behaviors, etc. Scavenger Hunt: Learners are assigned a list of items to collect in-world. This hunt leads them through a series fact-inding missions and relationship-building experiences. This can be used for individuals or groups, with debrieing on what the experience was like for them, and sharing of what they have learned or created through the hunt. Team Building: Literally, building. Teams of learners are assigned a task of co-creating an object, event or environment in the virtual world. Learning the basic building techniques would be required, and a coach may be present in-world to help the team. This assignment could be very simple or more complex, depending on the audience and desired outcomes, with debrieing both on the work product and on the process. Team Meetings: Distributed teams or functional groups gather in the virtual world to share artifacts and stories from recent work. This helps to build connections across geographies and enhance organizational knowledge and memory. Create a Club: A group is formed in Second Life to host events related to a functional area of expertise, product area or other topic. These group-sponsored events provide venues for engaging discussions and gathering of input from others in-world. Business Simulation: Learners assume control of a “real” business within Second Life, and are given targets and a budget. This would likely be a small business selling clothing, for example. Assessing the market, understanding the dynamics of the economy, allocating inances, maintaining

563

Advancing Learning Through Virtual Worlds

supply chains, innovating, and building relationships are all potential topics to be debriefed.

In every one of the examples above, there is an implicit requirement that the virtual world activities and experiences are integrated into a larger learning program or into real work processes. Many of the activities provide an immersive practice field for your participants in which they can test new skills and build the experience around the course content or around the behaviors they will need to be more effective at work. That practice needs to be very thoughtfully positioned as a logical step in moving from concepts (delivered through conventional means) through to application (implementing these concepts in context).

3d teaming - one example of an educational Activity in second life Here we describe one example of how we have used Second Life toprovide an educational experience for academic and corporate populations: 3D teaming. One of the open access applications Figure 5. 3D teaming exercise in Second Life

564

of Second Life is 3D teaming. Simply, this is a teaming event in Second Life that capitalizes on the ability to collaboratively build objects and use in-world communication tools to organize, strategize and solve problems together with other avatars. Essentially, imagine what we might normally provide our learners as an outdoor teaming exercise, but instead doing this entirely in a virtual space.For our model, we developed the simple task of working as a team to build a bridge from the edge of a body of water out to a platform in the center of a bay (figure 5). We arranged for four teams to do this exercise simultaneously with a coach assigned to each team to give some pointers and observe the behaviors of the team. We also integrated preparatory assignments, the teaming activity, debriefing of the teaming exercise and some follow-up discussions and assignments into a 5-week learning process. Specifically, we developed the following learning outcomes for this exercise: •



To Improve the Participant’s Understanding of the Complexity and Dynamics of Virtual Teaming Practices To experience and learn about what it takes

Advancing Learning Through Virtual Worlds



• •





to work effectively as a virtual team in a 3D environment To develop an understanding of some concepts relating to effective virtual teaming practices To develop the participant’s leadership skills To provide participants with an interactive and relective team experience in which everyone (faculty and participants) learns together about personal and team effectiveness given the challenge at hand. To improve the participant’s understanding of new alternatives for leveraging globally distributed talent To explore new media such as Second Life as an alternative for leveraging globally distributed talent

Preparation – learning the basics of second life For the participants, we asked them to do a number of things in advance of the event (figure 6). First, we wanted to be sure that the participants would be proficient in the basics of Second Life, so we asked them to download Second Life on their computer (can be downloaded for free at www. secondlife.com), create a Second Life account

and an avatar, and then to learn some of the basics of navigation and communication at Orientation Island. In addition, we sent them links to a couple of articles related to effective teaming behaviors, virtual teaming, and virtual world technologies. We produced a one-page user guide for them to print and have handy for the event, and we gave them the technology requirements for the event itself: 1. 2. 3. 4. 5.

Laptop meeting the minimum requirements as listed at the Second Life website Internet connection (wired preferred) Headset External mouse (optional) Power cord

For the coaches, we held preparatory meetings to orient them to the project and to Second Life, and to provide clarity on their roles and the mechanics of how the event would run. Coaches for this event were to observe the teaming behaviors in which we were interested and actively participate in the debriefing of the experience. They were also available to their teams to provide light guidance if they got stuck either in the mechanics of the environment or had questions about the assignment. For the environment, we built a cove – a space of virtual water nearly encircled by landon SSE

Figure 6. Preparation guidelines

565

Advancing Learning Through Virtual Worlds

MBA, an island in Second Life owned by the Stockholm School of Economics. On the shoreline, we placed four flags equally distributed, each representing a different country. This was simply to provide differing landmarks for the teams as they arrived and received their starting location assignment. Additionally, we built small canisters at the base of each flag that held the instructions for the team. We wanted to maintain elements of discovery in this experience, and maintain their curiosity. In an interesting experiment in outsourcing, we contracted in Second Life with a builder to create the components of a variety of bridges. This proved to be highly economical, and resulted in a set of bridge components that we then placed, in subsets, in each of the avatars’ inventories. Sourcing the capabilities to provide such items was achieved purely by happenstance, as is often the case in such a socially-based, events-driven environment. During an announced and open-invitation meeting for all avatars interested in running a business in Second Life, Ace Carson read profiles and found an avatar in the audience who claimed to be a builder of furniture and household objects. While the meeting was happening, Ace initiated a private chat session with the other avatar, “Kitten”, who answered questions regarding her skills and interest in building bridge components. A verbal agreement was quickly struck regarding number and properties of the items required, payment terms, and a date for initial delivery. Within days Ace received a full set of bridge parts from Kitten, and paid her the full sum: L$ 4000, roughly the equivalent of US$ 15. For the avatars, we decided to create a set of our own avatars specifically for this event. As noted, we had assigned the participants as preparatory work the task of creating their own accounts and avatars in order to learn the basics of communication and navigation. Knowing what we know about Second Life and the myriad costumes and items one can possess, we decided to bypass any potential distractions by providing our own avatars

566

for temporary use by the participants. In a nutshell, we did this for the following reasons: • •

• •





To limit potential distractions To provide a stock set of materials for the participants to use in the exercise (in the avatar inventories), thereby bypassing the need to ‘hand out’ materials to avatars during the event To pre-populate the Contacts list for each avatar with links to their team members To dress the avatars with color-coded tshirts to let them visually identify their team mates To pre-populate the Landmarks list with the location of the event (in case one of the participants wandered off) To create a group speciically for the participant avatars, enabling us to eficiently send messages to the entire group as things were beginning and as we wrapped things up.

time to team! There are two ways of approaching the actual running of this event. If all participants are located physically in the same location, then they may be gathered in a classroom for a brief overview of the basics of Second Life and the teaming exercise. They may then be asked to go to some other location where they have an internet connection and can sit individually away from the other participants. Alternatively, all participants may be distributed across the world with their computer and an internet connection and logged into Second Life in which case no face to face contact is necessary with any of the participants nor they with each other. Thestarting process was fairly simple:a participant was to email us once he/she was online at the designated time, and we would then reply with account information for one of the Second Life avatars we had prepared. (We then changed the

Advancing Learning Through Virtual Worlds

avatar passwords after this event) As the participants logged on with the avatars we had created, they found their avatar already at the teaming site as we had preprogrammed this. The coaches in avatar form were then ready to orient them to our teaming space and to direct them to find their teammates (via t-shirt color) and proceed to their assigned flagpole. There they were to click on the canister at the bottom of the flagpole, inside of which was a notecard (figure 7): The avatars in their teams were then to follow these instructions and build the bridge with all team members then walking across the bridge without falling into the water. As the action unfolded, we observed the emergence of several common issues across the teams: • • •



Lack of discussion within the team of what they were to do and who would do what Poor communication within the team to check everyone’s opinions Assumption regarding leadership – in some cases one person would take control, or there were competing team members vying for control Assumption that the teams were all competing against each other, with some people even “stealing” another team’s inventory In addition, some other interesting things

happened. First the technology was a hurdle for some. Their headsets did not work well, and some of the teams had to struggle with dual modality for communication (voice and text chat). Second, within Second Life, objects can float in the virtual space. Several of the teams did not realize this, and continued diligently to try and build a bridge reflective of real-world realities, which took more time than the one team who, once they realized the walkway pieces could float, quickly created their bridge and had all members stroll across. Third, some teams could not get all the team members to collaborate with some team members merely floating away or even joining other teams. After the building session, which took about 30 minutes, we then asked all the avatar participants and coaches to move to a nearby outdoor auditorium. In this auditorium we had built a large presentation board on which we had uploaded a set of debriefing questions (figure 8). We then proceeded to have a 30 minute discussion in which participants were asked to reflect and discuss their experience. Additionally, we asked all participants to submit an assignment related to this teaming exercise – a three to five page discussion of the following questions: 1. 2.

How can you apply virtual 3D experience to real lives? What did you enjoy?

Figure 7. Notecard with event instructions

567

Advancing Learning Through Virtual Worlds

3. 4.

What did you learn? What could we do to make exercise better?

In these submissions, participants discussed a number of 3D teaming issues such as time management, communication, decision making, trust, and team identity as well as technology issues such as the state of the technology and potential uses and advantages of 3D collaboration environments. Some comments included the following: While in the exercise we carried out, there was some confusion when using VOIP for group communication - who speaks when, and so on - this is not dissimilar from the problems encountered by early telephone adopters (DynamixGEL, 2007, p. 2)!

People act differently when their identity is hidden, which makes it attractive for those that are shy, but can also make people behave badly (Ibid, p. 2).

Figure 8. Debriefing session

568

We can learn a lot from people’s strengths and weaknesses, because of the limitations of the environment, so people who are good at communicating will find a way to communicate. (Ibid, p. 2). One of the stumbling blocks that we observed during the teaming exercise is that there were some participants who were lagging behind due to technical issues or unfamiliarity. This is unsurprising due to the relatively new concept of virtual teaming on a 3D internet application. A method to alleviate this stumbling block and to smoothen the pacing to a suitable level for all participants is to ensure that the participants have some introductory knowledge and basic familiarity in prior to the teaming exercise. However, this doesn’t ensure that the problem will subside away. Online, interactive help must be available at all times and the team leader must always take into consideration such issues when planning the pacing and time needed to complete a virtual team project (Rafi et al, 2007, p. 2).

Advancing Learning Through Virtual Worlds

Virtual teaming is very different from real-world teaming. One of the major differences is that maintaining open communications virtually is much harder than in the real-world. That is probably due to the understanding that having an online presence is weaker and less compelling than a real-world presence. There needs to be a standard protocol of appropriate behaviours when giving out and consuming information so that instructions, comments, and responses are communicated effectively (ibid, p. 2.). Overall the exercise was seen as beneficial by the participants as noted in the following quotations: We enjoyed working together, trying to overcome communication barriers, in order to construct something tangible (the bridge) (DynamixGEL, 2007, p. 4). This exercise challenged conventional thinking in strategically designing a solution adapted for the unique environment (no physics). We enjoyed the entertainment aspect: Second Life is like a game, it is always fun to play around especially in multiplayer settings (Ibid, p. 4). The fun part of this experience was, looking at people moving tools, trying to reach their goal (i.e. building the bridge) and seeing each other virtually as in a real life. Unlike other online media (e.g. video conferencing) it was possible for more than 20 people to come to the same platform and communicate at the same time. It was also possible to use expressions and to easily interact with each other. (Akayezu et al., 2007, p. 2)

overCoMing ChAllenges to iMPleMentAtion First it is important to address the technical and security concerns that accompany the use of new

collaborative technologies such as virtual worlds. There is software to install, hardware requirements for operation to fulfill, and security issues in terms of reaching through corporate firewalls to public virtual worlds (such as Second Life) and in terms of what data is shared in those spaces to consider. Credentialing and validating online identities is the topic of much debate and development at present. All of this very rightly brings shivers to the spines of corporate IT and security groups, but they may take comfort as new platforms that are designed for closed communities come online. To deal with this challenge, users and builders may assign building and access rights for areas within Second Life to individual avatars as well as report abusive behavior to Linden Lab. Additionally, for the momentSecond Life provides private islands for organizations to buy that allows for controlled access to those spaces, and other providers will license software for installation behind the corporate firewall. Not only are hardware and software profiles a concern, but there are also challenges that educators face in orchestrating events in Second Life. Other virtual worlds (e.g.,Protosphere, Olive, Qwaq, etc.) are licensed by organizations and will therefore provide more control for the educators and staff implementing them. The benefit of Second Life, however, is that it does provide a very rich toolset (building and sharing objects, multiple communication channels, group formation, avatar inventories, and access to public areas as well as private areas) and a standard, free account provides full capabilities, except owning land.

AreAs for fUtUre reseArCh Results from our teaming experiment as well as our investigation of other activities in virtual worlds reveal several areas for future research. For example, we need to investigate the various roles that educators and learners should take and in which settings as well as what source and format of content are best adapted for which virtual world experiences, keeping in mind the knowledge, 569

Advancing Learning Through Virtual Worlds

behaviors, attitudes framework. In line with this, we should investigate the best interplay between virtual world and real life events. The above leads to the need for a better understanding of what skills are required of educators and in what ways educators should be educated themselves to ensure the required skillset. Finally, further research should investigate the issue of virtual identities of learners and how these interplay with real life identities in educational settings as well as beyond in other settings.

ConClUsion In conclusion, our purpose here was to demonstrate that similar to what we have seen with the integration of internet-based resources into curricula and the development of corporate intranets as core learning and business environments (where learning and work gets done), virtual worlds provide another space in which learning can happen. Through our introduction of virtual worlds and the potential for learning activities within them, we hope that we have raised interest among the chapter’s readers to explore the possibilities these worlds provide.

referenCes Akayezu Josee, B., Bajwa, I., Ung, J., Wondemu, K., Tabish, W., & Plenet, Y. (2007). Second Life Assignment. Beck, J. C., & Wade, M. (2006). The kids are alright: How the gamer generation is changing the workplace. Boston: Harvard Business School Press. DynamixGEL. (2007). 3D virtual teaming team: Second Life assignment.

570

Hustad, E., & Teigland, R. (2008). Social networking and Web 2.0 in multinationals: Implications for KM. Paper presented at the 9th European Conference on Knowledge Management (ECKM), Southampton, UK. Johnson, S. (2006). Everything bad is good for you. New York: Berkley Publishing Group. Proserpio, L., & Gioia, D. (2007, March). Teaching the virtual generation. Academy of Management Learning & Education, 69–80. Rafi, A., Sakr, Y., Ben Jemia, S., Alhasanat, A., Tsaalbi, A., Lin, H., &Bin Mohamed, A. H. (2007). Team confluence: Second Life assignment. Wasko, M., Donnellan, B., & Teigland, R. (2007). Can regional innovation systems go virtual? Paper presented at American Conference on Information Systems (AMCIS).

AdditionAl reAding Bartle, R. A. (2004).Designing virtual worlds, Indianapolis, Ind: New Riders Pub. Biocca, F., & Levy, M. R. (1995).Communication in the Age of Virtual Reality, Lawrence Erlbaum Associates. Castronova, E. (2001).Virtual worlds: a firsthand account of market and society on the cyberian frontier, CESifo Working Paper No. 618, Munich:CESifo. Castronova, E. (2005).Synthetic Worlds: The Business and Culture of Online Games, Chicago: University of Chicago Press. Castronova, E. (2007).Exodus to the Virtual World: How Online Fun is Changing.

Advancing Learning Through Virtual Worlds

Davis, A., Murphy, J., Owens, D., Khazanchi, D., & Zigurs, I. (2009). Avatars, People, and Virtual Worlds: Foundations for Research in Metaverses . Journal of the Association for Information Systems, 10(2), 90. Donath, J. S. (1999). Identity and Deception in the Virtual Community.In M. A. Smith & P. Kollock (Eds.),Communities in Cyberspace.Routledge, 29-56. Duarte, D. L., & Snyder, N. T. (2001).Mastering Virtual Teams. San Francisco, CA: Jossey-Bass. Dubé, L., & Paré, G. (2004). The multi-faceted nature of virtual teams. In D. J. Pauleen (Ed.) Virtual teams: Projects, protocols, and processes. Hershey: Idea Group Publishing. Guo, Y., & Barnes, S. (2007). Why People Buy Virtual Items in Virtual Worlds with Real Money. The Data Base for Advances in Information Systems, 38(4). IBM. (2007).Virtual Worlds, Real Leaders: Online games put the future of business leadership on display. Global Innovation. Kahai, S. S., Carroll, E., & Jestice, R. (2007). Team collaboration in virtual worlds. The DATA BASEfor Advances in Information Systems, 38(4), 61–68. Lanier, J., & Biocca, F. (1992). An insider’s view of the future of virtual reality. The Journal of Communication, 42(4), 150–172. doi:10.1111/j.1460-2466.1992.tb00816.x Linebarger, J. M., Janneck, C. D., & Kessler, G. D. (2005). Leaving the world behind: supporting group collaboration patterns in a shared virtual environment for product design. Presence (Cambridge, Mass.), 14(6), 697–719. doi:10.1162/105474605775196625 Mennecke, B. E. (2008). Second Life and other Virtual Worlds: A Roadmap for Research, Communications of the AIS, 20 (20).

Naone, E. (2008).One Avatar, Many Worlds -Companies want to let users carry their avatar identities online. http://www.technologyreview. com/Infotech/20529/page1/. Pollitt, D. (2008). Learn-while-you-play programme gets IBM recruits up to speed. Training & Management Development Methods, 22(1), 401. Steuer, J. (1992). Defining Virtual Reality: Dimensions Determining Telepresence. The Journal of Communication, 4(24), 73–93. doi:10.1111/ j.1460-2466.1992.tb00812.x Steuer, J. (1995). Defining virtual reality: Dimensions determining telepresence. In F. Biocca& M. R. Levy (Eds.) Communications in the Age of Virtual Reality, Hillsdale, NJ: Lawrence Erlbaum Associates, 33-56. Wagner, C. (2008). Learning Experience with Virtual Worlds. Journal of Information Systems Education, 19(3), 263. Yoo, Y., & Alavi, M. (2004). Emergent leadership in virtual teams: what do emergent leaders do? Information and Organization, 14(1), 27–58. doi:10.1016/j.infoandorg.2003.11.001

Key terMs And definitions 3D Teaming: the act of collaborating in a 3D environment, e.g., virtual world Avatar: a virtual identity within a computerbased, simulated environment In-World: within a virtual world Machinima: a video filmed within a virtual world Open Access: an environment in which participants actively learn as they co-create content and meaning within virtual worlds Scripted Access: an environment in which participants experience somewhat passively a preplanned set of activities within virtual worlds

571

Advancing Learning Through Virtual Worlds

Teleport: to transport one’s avatar directly from one location to another within a virtual world without flying there. Virtual World: a computer-based, simulated environment where individuals assume a virtual identity called an avatar Web 2.0: internet-enabled communication technologies such as social networking sites (e.g., Facebook, LinkedIn), microblogging (e.g., Twitter), multimedia files (e.g., YouTube), co-creation content (e.g., blogs, wikis), and virtual worlds (e.g., Second Life)

2

3

4

5

6

7

8

endnotes 1

572

Second Life Herald, http://www.secondlifeherald.com/slh/2006/11/its_official_an.html, and CNNMoney.com, http:// money.cnn.com/blogs/legalpad/2006/11/ anshe-chung-first-virtual-millionaire.html

9

Gartner Group, April 2007, http://www. gartner.com/it/page.jsp?id=503861 Second Life, http://www.secondlife.com/ whatis/economy_stats.php Presentation by S. Mahaley, Human Resource Planning Society, April 2008 http://advancedmedialab.files.wordpress. com/2007/06/barcamp_second_life.jpg There are often contests held for the best machinima films Exchange rate is around 275 Lindens per US dollar. Current exchange rates here. For a machinima tour of the solar system created in Second Life, see Aimee Weber’s movie at http://alt-zoom.com/movies/azpresents/aweber/SolarSystem.html. Duke Corporate Education, Know, Do, Believe, http://www.dukece.com/how-wework/working-together.php

573

Chapter 32

Virtual Reality 2.0 and Its Application in Knowledge Building Johannes Moskaliuk University of Tuebingen, Germany Joachim Kimmerle University of Tuebingen, Germany Ulrike Cress Knowledge Media Research Center, Germany

AbstrACt In this chapter, we will point out the impact of user-generated online virtual realities on individual learning and knowledge building. For this purpose, we will irst explain some of the central categories of virtual realities (VRs) such as presence and immersion. We will also introduce the term virtual reality 2.0 (VR 2.0), which refers to those new types of VRs that are characterized by typical features of the Web 2.0, such as the opportunity that exists for users to create content and objects themselves. We will explain why we think the term VR 2.0–as a combination of Web 2.0 and VR–is a good label for currently existing user-generated online VRs. This chapter will also explain the concept of knowledge building, both in general terms and in the Web 2.0 context. The main emphasis of the chapter is on the signiicance of knowledge building for online VRs. In this context, we will describe the visualization of educational content, learner-object interaction, as well as personal, social, and environmental presence as its main features. We will also describe online VRs as a toolbox for user-generated content, and explain why the integration of different tools and seeing “living and learning” in context are relevant for applying user-generated online VRs in educational contexts. In conclusion, we will look at future trends for VR 2.0 environments. DOI: 10.4018/978-1-60566-384-5.ch032

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Virtual Reality 2.0 and Its Application in Knowledge Building

introdUCtion

virtual realities

Virtual Reality 2.0 (VR 2.0) is a new generation of online environment where users can communicate and interact with each other using avatars, and can define and generate its content. VR 2.0 is based on Web 2.0 concepts such as mashups of different applications and tools, the concepts of social networking and user-generated content, and the idea that the Web may replace the desktop as the main operating system and become the central entity for different applications. Our assumption is that the future of the Web could lie in a VR 2.0 which combines VR features and the ideas of the Web 2.0. In this context, we will describe under what conditions the best use can be made of VR 2.0 for purposes of individual learning and for collaborative knowledge building. We will first define what VR is, what its key features are and how they may be classified. The concepts of presence and immersion will be explained. VR 2.0 applications are described as systems that emphasize user communication and interaction, embracing Web 2.0 concepts to VRs. The virtual online world Second Life is presented as a prototype of a VR 2.0 tool. Next, we will introduce the knowledge building concept suggested by Scardamalia and Bereiter (1994, 2006), which is quite appropriate to describe and explain individual learning and collaborative knowledge building. We will also present an adaptation of this model to Web 2.0 environments by Cress and Kimmerle (2008), and explain why this model provides a suitable explanation, for knowledge building in the VR 2.0 context. We will point out the key factors of successful individual learning and knowledge building in VR 2.0. The chapter will conclude by looking briefly at the potential future development of VRs and their effects on individual learning and knowledge building.

Virtual Realities are artificial worlds that were generated digitally. In its simple form, a VR is an interface between humans and machines that will allow human beings to perceive computergenerated data as reality (Lanier & Biocca, 1992). The feature that defines VRs is interaction by a user with the virtual world, or in other words, immediate feedback (output, as immediate as possible) from the system to user input, creating a perception of some reality which is as realistic as possible by using three-dimensional presentation. Most definitions of VR also imply that data generated by the computer may be perceived with more than one sensory organ (i.e. at least seeing and hearing). The terms Artificial Reality (Krueger, 1991) and Cyberspace (Novak, 1991) are frequently used as synonyms of VR. Talking of an “Artificial Reality” implies that it is possible to represent content or data which have no corresponding “real” existence in the real world. “Cyberspace” refers not so much to technical aspects but to the concept of a world-wide data network between individuals. Located in different places, they can interact and communicate in a “social setting that exists purely within a space of representation and communication” (Slater, 2002, p. 535). There is a broad range of existing and potential (future) VR applications, differing mainly in the extent of technical requirements for input and output devices. Some VR systems require a user to wear a Head Mounted Display (HMD) in which stereoscopic projection creates a perception of space. DataGloves or DataSuits are worn as output devices, allowing the user to interact with the virtual world. In so-called Cave Automatic Virtual Environments (CAVE) events are projected into a room, moving objects in the representation in line with movements of the user’s body. The use of 3D glasses can increase depth perception. Flight or driving simulations locate users inside a

574

Virtual Reality 2.0 and Its Application in Knowledge Building

vehicle or cockpit and require them to use a steering wheel, control stick or other steering device as input medium. Either the surrounding space or an integrated screen are used for projection of what happens. Some systems also provide some feedback of real motion of the vehicle or plane to create a realistic impression. Desktop Virtual Realities have the lowest technical requirements for input and output media, in that they use a standard mouse, joystick or three-dimensional mouse which allows easy navigation in a three-dimensional space. In this case, 3D glasses will also increase three-dimensional perception. The terms Augmented Realities or Mixed Realities refer to systems in which a presentation of the real world is overlaid with computer-generated data, objects or representations. In principle, it is possible to use all VR applications for Mixed Realities. The term is also used for overlaying real and animated presentations in films.

Presence and immersion Classifying VRs by technical complexity of a system does not take into account their users’ perception, which should, however, also be among the relevant criteria. Distinguishing VRs by the degree of presence, which they allow appears to be more useful (Sheridan, 1992, Ijsselsteijn, 2004, Zhao & Elesh, 2008). Steuer (1992) uses the term telepresence, meaning “the extent to which one feels present in the mediated environment, rather than in the immediate physical environment” (p. 76/77). The main point here is the feeling of being there (Lombard & Ditton, 1997), i.e. the personal perception of an individual, which depends on the available sensory information, but also on this person’s control of attention, motivational factors and other mental processes. Steuer (1992) suggests two independent factors: vividness and interactivity. “Vividness means the representational richness of a mediated environment as defined by its formal features” (p. 81)

(cf. also Naimark, 1990; or Zeltzer, 1992). The definition of vividness includes sensory breadth, meaning the number of sensory dimensions, which are presented at the same time, and sensory, depth as the resolution in which these dimensions are presented. Interactivity is “the extent to which users can participate in modifying the form and content of a mediated environment in real time” (p. 84). The relevant factors of interactivity identified by Steuer (1992) include speed of system response to user input, range of attributes that can be manipulated within the system, and mapping between input from human users and system responses. In other words, this definition concentrates on those technical features of a media system which define its presence. A media system may be called a Virtual Reality if a high degree of presence is achieved, i.e. if there is a sufficient degree of vividness and interactivity that users have the impression to experience a “real” environment. Sheridan (1992) even named five factors that will influence the perception of presence: “extent of sensory information”, “control of sensors relative to environment”, “ability to modify the physical environment”, “task difficulty”, and “degree of automation”. An even higher degree of reflection of an individual user’s personal experience of a VR is contained in the concept of immersion, meaning the user’s feeling of, so to speak, being immersed in the virtual world which is provided by the technical system. So the concept of immersion not only takes into account technological aspects of a VR, but also emotional, motivational and cognitive processes of focusing attention. Obviously, a user’s intrinsic motivation (Ryan & Deci, 2000), personal involvement and interest in the respective topic may be considered as substantial factors for a high degree of immersion. The notion of flow (Csíkszentmihályi, 1990), meaning a mental state of operation in which the person is fully immersed in what he or she is doing, also has a great influence on the experience of immersion as defined here in the VR context. The degree of immersion

575

Virtual Reality 2.0 and Its Application in Knowledge Building

will depend, on the one hand, on what technology provides (thus overlapping with definitions of the notion of presence), but takes into account non-technological aspects as well.

4.

vr 2.0: Marriage of vr and web 2.0

6.

The inexpensive and simple availability of fast Internet has established online VRs as an additional form of VRs. These are network-based desktop applications (displaying presentations on a screen, using a mouse, joystick or 3D mouse), in which users are represented as avatars and may interact and communicate with each other. The VR system normally provides a platform and involves its users in content production. There is no longer any sense in distinguishing between authors or administrators as content producers, on the one hand, and users as consumers of that content on the other; “user-generated online VRs” would be a more appropriate term here. Our proposed terminology for such environments is VR 2.0 – a combination of technical facilities provided by an online VR with Web 2.0 concepts. VR 2.0 means, on the one hand, an expansion of VRs by adding Web 2.0 features, and on the other hand, the term implies that the degree of presence and immersion of which a VR is capable will not primarily depend on technical features and not necessarily on the number and fidelity of the input and output channels that it uses (Zeltzer, 1992). The following features characterize a VR 2.0:

7.

1.

2.

3.

576

A VR 2.0 is a 3D platform which is accessible online, not meant for playing a specific game or carrying out a specific program. Access is possible using a desktop computer which is connected to the Internet, without technical barriers, not requiring any specific equipment. The appearance and behavior of avatars may be determined and influenced by users.

5.

Avatars may interact and communicate with each other using spoken and written language. Avatars in a VR 2.0 environment share the same perception of this environment. Users are represented by avatars, giving them presence in the sense of being there. Content and objects may be generated by users in real time.

An important prototype for VR 2.0 is Second Life (http://www.secondlife.com). Development of this software by Linden Lab in San Francisco started in 1999 and the aim was to create a parallel world, imagined and created by its users (“residents”), for many different purposes. Second Life has been online since June 2003. It contains many of the above mentioned features and its technology has undergone a process of development in recent years. It integrates a voice chat that allows simple communication between users through spoken language. It is also possible to integrate external content (say, from web sites or video streams), and technical progress in this field is on the way. One peculiarity of Second Life is its micro-payment system which uses its own currency. Users can buy objects (say, clothes or scripts to animate an avatar) from other users and pay for these or for other services in Linden Dollars. Linden Dollars are convertible into real U.S. dollars (and vice versa) at a variable exchange rate. Second life is, however, only one example of a VR 2.0 application, there are now many others which are considerably different in the extent of facilities which they provide and technical requirements. At the same time, we are witnessing a permanent process of technical improvement of VR 2.0 applications, mainly in the field of more realistic photographic representation of three-dimensional worlds, integration of external services and applications and programming and developing facilities for scripts and objects within the VR 2.0. At the present stage, it is difficult to make reliable predictions about the future devel-

Virtual Reality 2.0 and Its Application in Knowledge Building

opment of this technology (cf. “Summary and Future Trends”).

educational Uses of vr 2.0 Many educational institutions have realized the benefit of VR 2.0 and are using Second Life as a platform for their own activities. Harvard University was one of the first to have its presence in Second Life and conducted a seminar for Law students in this environment. Ohio University, the Massachusetts Institute of Technology, Princeton University and the University of Edinburgh have their own virtual campuses in Second Life, providing various forms of academic teaching. The first in Germany to provide lectures and seminars through Second Life was RFH (Rheinische Fachhochschule) University of Applied Sciences in Cologne. It also offers a tutorial for new starters to explain the functionality and operation of Second Life. Many Higher Education and other educational institutions from all over the world are now represented in Second Life, ranging from adult education programs (German “Volkshochschule”) and language schools (including the German Goethe Institute) to libraries and cultural establishments. But the Second Life world is also inhabited by business companies, newspapers, marketing and advertising experts, concert and event agencies, training and coaching providers, financial services and staff providers, authors and musicians. The authors of this chapter have done some teaching of Psychology at the University of Tübingen within and through Second Life. This was carried out on the basis of a concept for presenting scientists and their work in the framework of VR 2.0 environments. It was also evaluated to what extent VR 2.0 is a suitable environment for scientific experiments and studies, and how these will have to be organized and carried out in such a context, including the question of how much time VR 2.0 users would be prepared to spend for their

participation in experimental studies and what payments or rewards they would expect. An experiment in Cognitive Psychology was also carried out in Second Life. It showed that VR 2.0 is indeed a suitable environment for controlled scientific studies. Currently the authors are improving this experimental environment and planning further studies. The idea is to use a software-controlled bot that will accompany users through the experiment and monitor and record their performance. The Knowledge Media Research Center and University of Tuebingen have their own islands in Second Life, in which experiments may be carried out. The islands also contain event and meeting rooms and a so-called sand box where users may create their own objects and content. The island also hosts an information area and an exhibition of teaching and learning tools.

Knowledge bUilding The following section will explain the concept of knowledge building in general terms: what it is, what it implies and what educational philosophy is behind it. We will introduce a model that transfers the ideas behind knowledge building to the Web 2.0 concept, and, finally, discuss the role of knowledge building in online VRs.

Concept of Knowledge building The concept of knowledge building, introduced by Scardamalia and Bereiter (1994, 1996, 2003, 2006), describes the creation of new knowledge in modern knowledge societies as a socio-cultural process. New knowledge is created in a social process and in concrete situations, and this will occur if a community has reached the boundaries of its existing knowledge, and if members of that community are no longer able to explain experiences in their environment with their existing

577

Virtual Reality 2.0 and Its Application in Knowledge Building

knowledge. Scardamalia and Bereiter compare that situation with a scientific community in which a group of scientists generates new knowledge and then shares it with the rest of the community. According to Scardamalia and Bereiter this ideal form of a knowledge-building community should also be the ideal for other forms of learning in schools, higher education and job training. Even if such a knowledge-building community will not necessarily create “new” knowledge in the scientific or academic sense, this knowledge – say, of a school class working on physical phenomena – will still be “new” to that respective community (i.e. that class or group of pupils). Knowledge building is always a discourse-oriented process. By participating in some common discourse, community members will share their knowledge with other members and in this way contribute to the advance of collective knowledge (Scardamalia, 2002). Knowledge building is based on a constructivist view of learning: only a person’s own experience with the environment will lead to the construction of new knowledge; knowledge cannot be passed on independently from one’s own experience. New experiences with one’s environment will necessitate the construction of new knowledge, regardless of whether or not this knowledge had previously been available to other individuals. The concept of knowledge building should be understood in the tradition of the Russian psychologist Lev Vygotsky. He regarded social interaction between learners as the key factor, learning is always a construction through a social process. According to Vygotsky (Vygotski, Cole, John-Steiner, Scribner, Souberman, 1978), even thinking should be understood as a social process, in that it reflects the culture in which individuals interact. Society or members of a community enable individuals to tackle tasks with requirements that go beyond the stage of development that they have reached themselves, thus extending their own range of skill and understanding (Vygotski, Cole, John-Steiner, Scribner, Souberman, 1978). The notion of cognitive apprenticeship is in the same

578

tradition. Here the acquisition of cognitive skills is compared to the acquisition of experience and skills in traditional training for craftsmanship. The underlying assumption is that the acquisition of cognitive skills is, in the last resort, only possible in a social context. By observing advanced learners or experts and – progressively – independent work coached by a teacher, accompanied by an exchange with other learners, an individual will acquire the competence to master complex tasks. Knowledge building plays a particularly important role in the field of computer-supported collaborative learning. Members of a knowledgebuilding community can use software tools to communicate with each other, exchange knowledge and give some structure to their cooperation. Scardamalia, Bereiter and Lamon (1994) developed the software CSILE (computer-supported intentional learning environment) – originally for school classes – with the aim of supporting a knowledge-building community. Its main component is a database in which users can save texts or graphical notes, browse through existing notes, and re-organize and comment these notes. The emphasis is on discussion of problems and appropriate approaches to their solution in order to gain deeper insights. Knowledge is meant to be exchanged openly, de-centralized, not with any central authority (say, a teacher) taking control or making an assessment. The underlying idea is that all users should participate in the knowledgebuilding community on equal terms and contribute to the growth of collective knowledge (Hewitt & Scardamalia, 1998; Scardamalia & Bereiter, 1994).

Adaptation of web 2.0 to Knowledge-building A new dimension of knowledge building may be observed in the framework of Web 2.0. Through active user participation in the production of content, individuals now have the opportunity to participate in a collective development of knowl-

Virtual Reality 2.0 and Its Application in Knowledge Building

edge and, at the same time, benefit from a vast amount of knowledge which is available world wide. Knowledge building is intensified by what is offered through the Internet: individuals participate in self-regulated learning through informal learning spaces, as members of a community of knowledge. The world-wide availability of (mainly free) software tools has opened up a new dimension of knowledge processes: large numbers of users can work jointly on shared digital artifacts (Tapscott & Williams, 2006). This will not only lead to cumulation of knowledge, by which the knowledge of many individuals is brought together and made available to others, but also to emergence or creation of new knowledge (Johnson, 2002). Cress and Kimmerle (2007, 2008) have proposed a model which takes into account, on the one hand, what the Web 2.0 can do, and takes up, on the other hand, Scardamalia and Bereiter’s approach in more depth, specifying the underlying processes. Their model describes how a large community of interest can use a shared digital artifact to produce knowledge jointly. In this context the authors refer to the wiki technology. Wikis are collections of web sites in the Internet or local intranets. These web sites may not only be read, but also edited by any user, and users may also create new content, add to, modify or even delete existing content (Leuf & Cunningham, 2001; Raitman, Augar, & Zhou, 2005). In doing so, several users can create one digital artifact together, and this activity will support the collaborative development of knowledge (Fuchs-Kittowski & Köhler, 2005; Köhler & Fuchs-Kittowski, 2005). The Cress and Kimmerle (2008) model explains this collaborative development of knowledge by integrating a constructivist and systems-theoretical perspective. The assumption is that wikis support learning and knowledge building in precisely the way that was described by Scardamalia and Bereiter. It is argued that people’s individual knowledge can be used as a supply for learning processes of other people (cf. Kafai, 2006) and that a wiki,

as a shared digital artifact, is perfectly suited for supporting this kind of mutual use and development of knowledge (cf. also Bruckman, cress). The authors distinguish in their model between cognitive systems of individual users and the social system represented by the community. These systems are distinguished by different modes of operation: while cognitive systems operate on the basis of cognitive processes, social systems are based on communication. Individual learning and knowledge building will occur, according to the model, if cognitive systems externalize their own knowledge, i.e. transfer it into a social system (say, by user cooperation on a wiki) and internalize new knowledge at the same time. An analogous process takes place with social systems, which externalize and internalize (in an exchange with cognitive systems of users) new information and new knowledge. Such a fruitful exchange between cognitive and social systems via shared digital artifacts is stimulated by socio-cognitive conflicts which exist between the prior knowledge of a cognitive system and the information which is contained in a social system. Such socio-cognitive conflicts may be solved by mutual adaptation of knowledge and information through the exchange processes as described above. In this way, according to the model, new knowledge will be generated. Apart from wikis, the model may also be applied to other software tools which individuals use to work jointly on a digital artifact. The authors have also described the joint development of individual and collective knowledge by using social-tagging systems or pattern-based task management systems (Kimmerle, Cress, & Held in press; Riss, Cress, Kimmerle, & Martin, 2007). One step further is applying this model to individual learning and collective knowledge building in VR 2.0. Like other Web 2.0 tools, VR 2.0 provides a shared digital artifact and environment with the opportunity to cooperate with other individuals and get access to their knowledge. In this way, VR 2.0 may induce socio-cognitive

579

Virtual Reality 2.0 and Its Application in Knowledge Building

conflicts and provide a framework to solve these conflicts. This may occur, for example, if users work jointly on a model, prepare a simulation or represent their knowledge in a three-dimensional mind map and notice in this context that they have different points of view or different degrees of prior knowledge, and try to find a solution together. Other than text-based Web 2.0 tools like Wikis, VR 2.0 provides additional multi-media facilities for knowledge representation, including three-dimensional representation. Even bearing in mind that the presentation and readability of text in existing VR 2.0 prototypes is still unsatisfactory, the VR 2.0 concept as such permits using the third dimension for text presentation. Single web pages may not only be shown side by side on a screen (say, in different tabs of a browser), but also behind and on top of each other. Showing three-dimensional clusters of documents, sorted by belonging together or affinity to some keyword, may also help to find single documents and understand their context.

Using vr 2.0 for Knowledge bUilding The following section will describe some features of VRs that may support knowledge building. We will characterize different forms of visualization of educational content, and describe why such visualizations can support learning and knowledge building. Whereas current Web 2.0 applications only use 2D multimedia material, VR 2.0 can also show the third dimension for more details and a higher degree of immersion. Furthermore, VR 2.0 can create virtual environments that could not normally be visited by real human beings (say, an active volcano, or a city in the Middle Ages), or to visualize metaphoric content that needs to be translated into real life (e.g. a map of a city or some statistical distribution). In addition, users may interact with learning objects, move around them, look below or behind them, or manipulate

580

the form, shape or behavior of an object (say, electric circuitry). The social presence of users (i.e. the fact that avatars of other users are also present in the VR at the same time) and opportunities provided for social interaction are features of the system that support knowledge building. In addition to simple ‘who’s online’ information in Web 2.0 applications, VR 2.0 offers more specific information about other users and a wide scope of social interaction. Environmental presence allows users to meet each other in the same environment; they are able to point to things or work together directly on learning objects. This is a very important difference between previous Web 2.0 communication and VR 2.0. As a result, VR 2.0 is more similar to real-world communication. VR 2.0 offers a toolbox for user-generated content. As in other Web 2.0 applications, users are invited to introduce their own content and build their own objects. Taking a constructivist view of learning, one might describe this integration of learners in the role of active constructors of knowledge as an important and relevant factor of successful learning and knowledge building. VR 2.0 also provides the opportunity to integrate different tools and applications. This so-called mashup of content in one application, which is a central feature of Web 2.0, is particularly valuable for users of VR 2.0 (e.g. by providing the integration of a learning management system in Second Life). The future of this technology may be seen as a seamless integration of different media. Another important feature of VR 2.0 is the fact that it brings together learning and living in a way that will encourage lifelong learning, situational learning and implicit or informal learning.

visualization of educational Content In a VR it is possible to visualize or imitate objects from the real world. The technical features which VRs provide can make such a visualization look very realistic, close to the real thing, and in this

Virtual Reality 2.0 and Its Application in Knowledge Building

way users may experience a high degree of immersion in the learning environment. Consequently, learning is a more immediate experience and may be more effective (Gay, 1994). Compared to two-dimensional presentations or mere descriptions, the more realistic three-dimensional type of presentation gives learners an additional benefit because this resembles what they look at every day and can be understood more easily. An advantage also results from the fact that visual learning content is processed and remembered more easily than text (Paivio, 1990, Shepard, 1967). A presentation which is close to reality is also a good way to “anchor” content from a learning environment in a context which is close to real-life situations. The benefits of this type of learning content have been described in research on anchored instruction (Bransford, Sherwood, Hasselbring, Kinzer, & Williams, 1990; Cognition and Technology Group at Vanderbilt, 1992). At the same time, this content is more authentic and refers to some concrete application of what is being learned, which may be an important requirement for motivating learners. This will allow situated learning – knowledge will not remain inert, but may be applied directly. Realistic visualization of learning objects in a VR 2.0 is particularly suitable in those cases in which actual observation of the object or a visit to the real place would be too complicated, expensive or dangerous, or if learners are separated from each other.

is available but could not normally be perceived by humans without a change or transduction of scale (see below). Apart from displaying real content, a VR may also visualize or simulate abstract concepts or translate them into some concrete shape. It is possible to materialize data, processes or semantic structures and make mental models explicit. In this context a VR is a cognitive tool for problem solving and it extends the scope of a person’s perception and cognition. (Biocca, 1996). Understanding abstract concepts, a complex cognitive process, is easier with a concrete representation. Recognizing connections and patterns requires a smaller extent of mental effort. In the sense of embodied cognition (Clark, 1997) thinking is not regarded as a formal operation based on abstract symbols, but it is embedded in a situational and cultural context (Anderson, 2003). A VR also permits a representation of content which could not be perceived or registered by human beings without changing its scale or transduction (Winn, 1993). Scaling may be necessary because the size of the learning object (say, a human cell or the solar system) would rule out direct observation without appropriate enlargement or reduction. The term transduction refers to representations of information which could not normally be perceived by the sensory system of human beings (say, by using different colors for showing a body’s emission of different degrees of warmth).

Example: A geography class visits a virtual volcano and the students are able to watch an eruption. Their avatars can move freely inside the volcano and observe subterranean seismic activity.

learner object interaction

A VR may also be enriched with information that would normally not be visible. This may consist of schematic or abstract information that is not available in real life, like street names on a satellite view of a city, or information which

The benefit of three-dimensional representation from the learner’s point of view is increased by the opportunity to interact with objects in a virtual world, manipulate and change them. First of all, learners in a VR can inspect a learning object from all sides, go around it, look at it from underneath, from above or from the other side. This is an advantage from the point of view of discovery learning (Bruner, 1961).

581

Virtual Reality 2.0 and Its Application in Knowledge Building

Example: Students of Mechanical Engineering study a virtual model of an electrical engine to understand its functionality. They may not only inspect the engine from all angles, but also make adjustments and watch the results in its operation. They may also disassemble the entire engine and practice its correct re-assembly. What is particularly relevant in the context of knowledge acquisition is an adequate representation of the transition from abstract visualization with schematic diagrams etc to other forms of representation which depict and closely resemble reality. This provides external models for mental processes which can be internalized by learners more easily (Salomon, 1994). The VR provides the model of a cognitive operation which learners have to carry out mentally in order to acquire their own mental model of certain facts or of a topic of instruction. A dynamic overlay of realistic and abstract representations of the same thing may be controlled by learners through an interactive process, say, by replacing a schematic presentation of an object step by step with more realistic pictures, depending on the individual progress of learning or the extent of prior knowledge. Example: School students of Biology study the structure of human organs. The irst representation of the organs which they see is a schematic drawing, just meant to explain its main characteristics. With increasing knowledge, the representation of each organ resembles more closely its actual appearance. At the same time, an interactive functionality permits to overlay the realistic picture with the schematic representation, in order to link what was learned about the structure with a view of the real organ. Scaling of visualized objects may (ideally) also be performed as an interactive process, in order to

582

enable learners, say, to start with the original size of an object and zoom into more detail.

Personal and social Presence When describing the main features of a VR, reference was made to the notion of presence. Heeter (1992) has proposed personal presence, social presence and environmental presence as the main dimensions of this concept. The following subsection describes personal and social presence, whereas environmental presence will be dealt with in a separate section. Users in a VR have to be represented by avatars. This is a requirement both for personal presence of an individual, i.e. the personal feeling of a user to be there in a world created by media, and for social presence of other individuals as “sense of being with another” (Biocca, Harms, & Burgoon, 2003, p. 456). In other words, learning in a VR 2.0 is embedded in a social environment. Social-psychological aspects such as identification with the group, anonymity of group members and the perception of social identity (Tajfel & Turner, 1986) are extremely important. What is evident here is the fact that presence will not only depend on the degree of realism of which a VR is capable, or in other words, a technologically sophisticated VR will not automatically lead to a higher degree of presence. The key factor of a feeling of social presence is the amount and resolution of available information. Avatars will not need to be as realistic as possible – the point is that users should perceive them as valid representations of real people. As far as the significance of personal and social presence for knowledge building is concerned, both are decisive for establishing a knowledgebuilding community. If knowledge building is regarded as a socio-cognitive process, the perception of presence of one’s own self and other group members in a learning environment is necessary for discourse-oriented forms of learning. The existence of media-based representations of other group members makes it easier for socio-cognitive

Virtual Reality 2.0 and Its Application in Knowledge Building

conflicts to occur and to perceive these conflicts. At the same time, VR 2.0 provides a framework for solving such conflicts by offering a broad range of activity and communication options. Such realistic forms of interaction and communication within a VR 2.0 make it easier to establish some common ground. This term refers to knowledge about information which is shared between participants of a conversation, their shared understanding (Clark, 1996). In face-to-face communication, the existence of some common ground is demonstrated by grounding activities like nodding, shaking one’s head, giving an immediate reply or simply by paying attention. In media-based communication the effort required for grounding depends on the type of media and is relatively small in VR 2.0, resembling natural face-to-face communication. Generally, there is a great similarity between VR 2.0 and face-to-face arrangements. Even if sensory perception is restricted in comparison to real life, the perception of one’s own self as part of a learning environment and of the presence of other people is similar as in a setting in the real world. This is even more the case if people are affected personally and see some connection between their own person and what happens in a VR 2.0. This will increase their feeling of presence. It will also increase collective cognitive responsibility of a group for succeeding together (Scardamalia, 2002), a key factor for efficient learning. Learning in a community will only be successful if individual learners perceive themselves as important members of the group and jointly accept responsibility for achieving the targets of the group. In this way a genuine learning community will be formed in which all members of a group of users with different backgrounds and experiences can bring in their knowledge to the benefit of all. The observation of what other members of such a group are doing will lead to a form of social observational learning (Bandura, 1977). Bandura’s assumption is that individuals (as observers) learn

by observing other individuals (models), and that consequences of the model’s behavior (acceptance or punishment) encourage or discourage the observer as well. Observational learning works with procedural knowledge (know-how, skills), but may also support the acquisition of factual knowledge (cognitive apprenticeship approach, cf. above). A VR 2.0 provides opportunities, close to reality, to acquire new knowledge by observing other users. Example: Prospective police oficers of the Border Guard are meant to be prepared for dangerous situations during a cross-border control, as part of their training. They observe the behavior of their instructor and two colleagues in a VR 2.0 simulation of such a situation. Then they rehearse their own behavior in similar situations. This type of VR 2.0 practice is good preparation for real situations, but obviously less complicated and dangerous than in real life.

environmental Presence Environmental presence is closely linked with personal and social presence. Different learners represented by their avatars are simultaneously present in the VR, share the same (or similar) awareness about their situation and environment. This contains two substantial benefits for cooperative knowledge building: VR 2.0 users may refer to objects in their (learning) environment without any ambiguity (say, by pointing at whatever it is). These learners will find it easier to enter into an exchange and discussion on learning content and objects, and in this way cooperative knowledge building will occur. At the same time, it is relatively easy to create group awareness in a VR 2.0. Group awareness means the perception and knowledge of the presence (who and where) of other people and of what they are doing at this moment (Gutwin & Greenberg, 2002; Kimmerle, Cress, & Hesse, 2007) – one of the basic require-

583

Virtual Reality 2.0 and Its Application in Knowledge Building

ments for cooperative knowledge building. The shared environment (as the external representation of what goes in a user’s mind) facilitates nonverbal communication by allowing, for example, manifesting actions that make explicit verbal backchanneling unnecessary, or observation of other avatars’ behavior, from which the conclusion may be drawn that everything was understood by all (Clark, 1996). Grounding is also much easier in such a shared environment. In this way VR 2.0 is very close to the opportunities which face-to-face learning settings provide, and at the same time it solves the classical communication problem in computer-mediated learning arrangements that results from the absence of a shared environment. While in computer-mediated communication fewer social stimuli are available (Kiesler, Siegel, & McGuire, 1984) and those involved in interaction have fewer opportunities to express themselves and understand the background of their partner than in face-to-face communication (Culnan & Markus, 1987), such social stimuli and background information exist abundantly in a VR 2.0, in fact, very similar to a face-to-face communication. The context in which knowledge is acquired is of paramount importance for effective learning and later recall of what has been learned (Godden & Baddeley, 1975). Learning and cognition are always situated (Greeno, 1998), what people know depends on the context in which this knowledge was acquired and is being used (Brown, Collins, & Duguid, 1989). The distributed cognition approach (Hutchins, 1995) goes even one step further by regarding artifacts, as parts of a socio-technical system, as the main components: Cognition is always distributed between the individual and some artifact, so dealing with artifacts is the main requirement for any knowledge acquisition and knowledge building. The context and situation in which knowledge is acquired is even more relevant for the acquisition of procedural knowledge, which exists implicitly but cannot be made explicit and passed on easily.

584

Here, learning takes place through observation of other individuals (as in observational learning, cf. above), observation of their interaction with the environment and learning objects, and trying it out together as a form of learning by doing. This is impossible without a shared environment. The VR 2.0 context with its close affinity to reality makes it easier to transfer what has been learned into situations outside this learning context. So a VR 2.0 provides ideal conditions for acquiring skills and concepts in realistic situations and contexts, even if it would be too costly or dangerous in reality. Example: A company would like to expand business with Chinese enterprises and train their staff in inter-cultural competence to prepare cooperation with their respective business partners. In a VR 2.0 house in the typical style of the country, company staff meet a Chinese trainer and practice appropriate behavior and manners.

Platform for User-generated Content VR 2.0 is a platform and provides tools to its users that enable them to create their own content and objects. It will depend on the users what they use the platform for and what content and objects it contains. Integrated 3D editors enable them to create three-dimensional objects from within the VR 2.0. But it is also possible to create other multi-media components (e.g. films or audio files) or text within a VR 2.0. Example: A learner group is preparing a supervised written examination. The participants produce jointly a three-dimensional concept map of various theoretical concepts that were meant to be learned. The nods of the concept map are linked with brief summaries or references, which are available in the shared digital library.

Virtual Reality 2.0 and Its Application in Knowledge Building

Active construction of learning content has the decisive advantage that active involvement in a learning environment leads to deeper understanding of its content and promotes knowledge acquisition (Craik & Lockhart, 1972). In the process of active construction, users acquire a mental model of their learning content. There is no need to create abstractions, as the learning content may be experienced in an environment that looks close to reality, and may be manipulated. This makes learning by design (Kolodner et al., 2003; Kolodner et al., 2004) possible, understood in the sense of actual construction of real objects in a VR 2.0. Ideally, the same laws of physics apply in a VR 2.0 as in the real world. So it is possible to put hypotheses to an immediate test in “reality” and learn by experience.

in part) by programmed virtual agents that support learners in their knowledge acquisition process or check and correct their steps and results. Another benefit for learning that results from the construction of objects and creation of content is due to the self-explanation effect (Chi, Bassok, Lewis, Reimann, & Glaser, 1989). Explaining learning content to other users leads to deeper insights of the person who does the explanation. Externalizing knowledge supports elaboration. The opportunity to create content is not restricted to objects and content of the environment. Users can also influence the representation of their own self in the VR 2.0 environment, the appearance of their avatar, its behavior and style of presentation.

integration of different tools Example: Students of Architecture deal with the construction of multistory buildings which are safe during earthquakes. Based on their previous knowledge, they draft a plan of such a building and calculate its static ability to withstand earthquakes. Then they construct such a building in a VR 2.0. The VR 2.0 allows simulation of an earthquake to test the stability of this construction. If breaking points have been identiied through the simulation, the students can modify their draft and test their modiied construction. At the same time, construction of an environment will always take place in some context of cooperation. Content is produced from within the VR 2.0 and shared between users from the very beginning. Other users may watch the process of construction, comment on it or even interfere. Experienced learners or experts in the role of tutors have a platform with VR 2.0 which they can use to support less experienced learners or novices in the sense of cognitive apprenticeship (Collins et al., 1986). But unlike face-to-face tutoring, in VR 2.0 settings tutors may be replaced (completely or

VR 2.0 is a platform that permits the integration of external content. An integrated web browser can display web sites in a VR 2.0, and films, pictures or other multi-media files may be included as well. Topical content provided through RSS feeds (e.g. blog entries, Twitter messages) may also be included in a VR 2.0. There is no break between different media, between VR 2.0 and other content, no change between different systems (say, between VR 2.0, web site on the Internet, media player …), and users are not required to adapt their search and navigation strategies to changing media. In other words, VR 2.0 is a central platform that may, in principle, integrate all other content. This will lead to more efficient information processing, avoiding interruption that would result from changing media. The integration of programs for two-dimensional text display (browser, PDF reader) avoids the restriction of not being able to read large quantities of text in a three-dimensional presentation. The social presence of other users continues while a user is reading in a VR 2.0, say, a web page in the Internet with the integrated browser. The benefit of having communication and user

585

Virtual Reality 2.0 and Its Application in Knowledge Building

interaction in a form that resembles reality will also exist when looking at external content. The learning community also has access to content that was not originally part of the VR 2.0. At the same time, this external content becomes part of the learning environment, is embedded in the learning context, and the advantages of environmental presence (as described above) will come to bear. Example: Students of a Literature course study the works of William Shakespeare and try to ind interesting information in a VR 2.0. In a virtually rebuilt Globe Theatre they can watch ilms with some of the author’s most important dramas, and replay some key scenes themselves. The system also contains a large library with Shakespeare’s complete works, available in the form of e-books which may be read in the VR 2.0. They also ind a chat bot that is able to answer student questions by quoting from Shakespeare’s works. This chat bot uses a large database of quotations which exists outside VR 2.0. The students will enter all results of their research into a wiki, which is part of the VR 2.0 and on which they can work together in cooperative effort. It is possible to integrate even an entire learning management system (LMS) into a VR 2.0. A LMS supports the provision and use of various types of learning content and provides tools for user cooperation. At the same time it provides administration tools for users and learning events, access control for different types of material and – in most cases – also tools to run tests or quizzes to monitor one’s learning progress. One example of this type of software is the Open Source project Sloodle, which links the VR 2.0 Second Life with the LMS Moodle, providing simple access to and allowing uncomplicated work on text-based learning material (Kemp & Livingstone, 2006). If a VR 2.0 is used consistently as a platform for

586

the integration of all content that is relevant for the learning process, it may also be an option to use the VR 2.0 for organizing and structuring the learning process itself. Different steps or activities may be allocated to different rooms on a virtual campus, and such a VR 2.0 may contain lecture halls, laboratories, rooms for group work, a library and media center, thus giving learners an idea of the resources and activities that they need for their learning process. The allocation of different steps or learning units to different “rooms” may also help learners to give a structure to their learning and understand what is going on.

sUMMAry And fUtUre trends If the previous trend continues, the costs for providing desktop computers with advanced technology will continue to go down. At the same time, graphics cards will be more powerful, the machines will have more computing power and Internet connections will be faster. It will soon be common to have a computer that has capabilities to access a VR 2.0 application. So we can assume that the use of VR 2.0 will increase and many people can access VR 2.0 online environments. Another trend that we can expect is a more and more seamless technological linkage of different online worlds in which people live. As VR 2.0 is a platform that will neither set goals of a game nor prescribe any specific uses, it may be used in a wide range of different worlds. While in classical learning settings (schools and universities in particular) the acquisition and application of knowledge are two separate things, VR 2.0 allows an integration. Learning – ideally – is not restricted to a specific place and time, but embedded in other activities, as a process of life-long learning. What may also be expected is an improvement of input and output devices. Apart from increasingly flatter screens and higher resolution, we are witnessing the development of capable touch screens. These will allow direct interaction with a

Virtual Reality 2.0 and Its Application in Knowledge Building

VR 2.0 by touching or moving the hand in front of a screen or projection space. We can also expect an advance in the development of a 3D mouse that allows navigation in a three-dimensional space. At the moment we can only speculate about the use of data gloves, which have, so far, not been used in conjunction with VR 2.0 systems, but if the appropriate technology exists and components can be produced at reasonable costs, these may turn out to be a potential input device for VR 2.0 systems. While, so far, users of VR 2.0 systems depend on the servers of their provider (i.e. cannot store their data on their own computer infrastructure), there are now some providers that offer “closed” VR 2.0 systems that are only accessible by a restricted group of users. Business corporations or universities might operate VR 2.0 systems which can only be used by their own staff or students, and where users have full control of their own data. Most VR 2.0 systems still require special client software, which may, however, not be available for all operating systems and which requires (additional) installation by their users. Access to more sophisticated VR 2.0 systems by using a standard Internet browser is still a technical barrier. A related problem is access to a VR 2.0 using mobile telephones or PDAs. A first step in this field was made by Linden Lab who made the source code of their Second Life client freely available under GPL, thus allowing potential further development of that client by interested users. Another problem exists in that different VR 2.0 systems from different providers are incompatible with each other, so content from one VR 2.0 is hardly portable to another one. Avatars only apply to one VR 2.0, so users of different systems have to use different avatars. Standardization between different VR 2.0 systems will be absolutely necessary to gain wider acceptance and find more users for VR 2.0.

ConClUsion All VR 2.0 systems which are now available have been tested in various contexts and are under development. They are suitable tools for collaborative knowledge building and individual learning. The main difference to classical VR applications lies in the platform character of VR 2.0 and the role of user-generated content. VR 2.0 allows users to produce their own content. This permits learning in the form of active construction of knowledge, in a realistic applied context. What is also important is the social and communicative aspect. Users have online access to a VR 2.0 and meet other users from all over the world. This allows communication and interaction with other users, a key requirement for socio-cultural learning. With these features, VR 2.0 has a great potential for knowledge building in schools, universities and job training. Many aspects of a VR 2.0 learning platform may be compared, in terms of what they can achieve, to face-to-face settings, and some of their built-in facilities even go far beyond that. In this way VR 2.0 systems are important milestones for ubiquitous and life-long learning.

referenCes Anderson, M. L. (2003). Embodied cognition: A field guide. Artificial Intelligence, 149(1), 91–130. doi:10.1016/S0004-3702(03)00054-7 Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall. Biocca, F. (1996). Intelligence augmentation: The vision inside virtual reality. In B. Gorayska & M. Jakob L. (Eds.), Cognitive technology: In search of a humane interface (pp. 59-78). Amsterdam: Elsevier.

587

Virtual Reality 2.0 and Its Application in Knowledge Building

Biocca, F., Harms, C., & Burgoon, J. K. (2003). Toward a more robust theory and measure of social presence: Review and suggested criteria. Presence (Cambridge, Mass.), 12(5), 456–480. doi:10.1162/105474603322761270 Bransford, J. D., Sherwood, R. D., Hasselbring, T. S., Kinzer, C. K., & Williams, S. M. (1990). Anchored instruction: Why we need it and how technology can help. In D. Nix & R. Spiro (Eds.), Cognition, education, and multimedia: Exploring ideas in high technology (pp. 115-141). Hillsdale: Lawrence Erlbaum Associates. Brown, J. S., Collins, A., & Duguid, P. (1989). Situated cognition and the culture of learning. Educational Researcher, 18(1), 32–42. Bruckman, A. (2006). Learning in online communities. In R. Sawyer (Ed.), The Cambridge handbook of the learning sciences (pp. 461-472). New York: Cambridge University Press. Bruner, J. S. (1961). The act of discovery. Harvard Educational Review, 31, 21–32. Chi, M. T. H., Bassok, M., Lewis, M. W., Reimann, P., & Glaser, R. (1989). Self-explanations: How students study and use examples in learning to solve problems. Cognitive Science, 13(2), 145–182. Clark, A. (1997). Being there: Putting brain, body, and world together again. Cambridge, MA: MIT Press. Clark, H. H. (1996). Using language. Cambridge: Cambridge University Press. Cognition & Technology Group at Vanderbilt. (1992). Anchored instruction in science and mathematics: Theoretical basis, developmental projects, and initial research findings. In R. A. Duschl & R. J. Hamilton (Eds.), Philosophy of science, cognitive psychology, and educational theory and practice (pp. 244-273). Albany, NY: State University of New York Press.

588

Collins, A., Brown, J. S., & Newman, S. E. (1986). Cognitive apprenticeship: Teaching the craft of reading, writing, and mathematics. Cambridge, MA: University of Illinois. Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11(6), 671–684. doi:10.1016/S00225371(72)80001-X Cress, U., & Kimmerle, J. (2007). A theoretical framework for collaborative knowledge building with wikis: A systemic and cognitive perspective. In C. Chinn, G. Erkens, & S. Puntambekar. (Eds.), Proceedings of the 7th Computer Supported Collaborative Learning Conference (pp. 153-161). New Brunswick: International Society of the Learning Sciences. Cress, U., & Kimmerle, J. (2008). A systemic and cognitive view on collaborative knowledge building with wikis. International Journal of Computer-Supported Collaborative Learning, 3(2), 105–122. doi:10.1007/s11412-007-9035-z Csíkszentmihályi, M. (1990). Flow: The psychology of optimal experience. New York: Harper and Row. Culnan, M. J., & Markus, M. L. (1987). Information technologies. In F. M. Jablin, L. L. Putnam, K. H. Roberts & L. W. Porter (Eds.), Handbook of organizational communication: An interdisciplinary perspective (pp. 420-443). Newbury Park, CA: Sage Publications. Fuchs-Kittowski, F., & Köhler, A. (2005). Wiki communities in the context of work processes. WikiSym 2005–Conference Proceedings of the 2005 International Symposium on Wikis 2005 (pp. 33-39). New York: ACM Press. Gay, E. (1994). Is virtual reality a good teaching tool. Virtual Reality Special Report, 1(4), 51–59.

Virtual Reality 2.0 and Its Application in Knowledge Building

Godden, D. R., & Baddeley, A. D. (1975). Contextdependent memory in two natural environments: On land and underwater. The British Journal of Psychology, 66(3), 325–331. Greeno, J. G. (1998). The situativity of knowing, learning, and research. The American Psychologist, 53(1), 5–26. doi:10.1037/0003-066X.53.1.5 Gutwin, C., & Greenberg, S. (2002). A descriptive framework of workspace awareness for real-time groupware. [CSCW]. Computer Supported Cooperative Work, 11(3), 411–446. doi:10.1023/A:1021271517844 Heeter, C. (1992). Being there: The subjective experience of presence. Presence (Cambridge, Mass.), 1(2), 262–271. Hewitt, J., & Scardamalia, M. (1998). Design principles for distributed knowledge building processes. Educational Psychology Review, 10(1), 75–96. doi:10.1023/A:1022810231840 Hutchins, E. (1995). How a cockpit remembers its speeds. Cognitive Science, 19(3), 265–288. Ijsselsteijn, W. A. (2004). Presence in depth. Eindhoven: University of Technology. Johnson, S. (2002). Emergence: The connected lives of ants, brains, cities, and software. New York: Scribner. Kafai, Y. B. (2005). Constructionism. In R. K. Sawyer (Ed.), The Cambridge handbook of the learning sciences (pp. 35-46). New York: Cambridge University Press. Kemp, J., & Livingstone, D. (2006, August). Putting a Second Life “metaverse” skin on learning management systems. In Livingsone & J. Kemp (Eds.) Second Life Education Workshop at the Second Life Community Convention, San Francisco (pp. 13-18). San Francisco: The University of Paisley.

Kiesler, S., Siegel, J., & McGuire, T. (1984). Social psychological aspects of computermediated communication. The American Psychologist, 39(10), 1123–1134. doi:10.1037/0003066X.39.10.1123 Kimmerle, J., Cress, U., & Held, C. (in press). The interplay between individual and collective knowledge: Technologies for organisational learning and knowledge building. Knowledge Management Research and Practice. Kimmerle, J., Cress, U., & Hesse, F. W. (2007). An interactional perspective on group awareness: Alleviating the information-exchange dilemma (for everybody?). International Journal of Human-Computer Studies, 65(11), 899–910. doi:10.1016/j.ijhcs.2007.06.002 Köhler, A., & Fuchs-Kittowski, F. (2005). Integration of communities into process-oriented structures. Journal of Universal Computer Science, 11(3), 410–425. Kolodner, J. L., Camp, P. J., Crismond, D., Fasse, B., Gray, J., & Holbrook, J. (2003). Problembased learning meets case-based reasoning in the middle-school science classroom: Putting learning by design into practice. Journal of the Learning Sciences, 12(4), 495–547. doi:10.1207/ S15327809JLS1204_2 Kolodner, J. L., Camp, P. J., Crismond, D., Fasse, B., Gray, J., Holbrook, J., et al. (2004). Promoting deep science learning through case-based reasoning: Rituals and practices in learning by design classrooms. In N. M. Seel & S. Dijkstra (Eds.), Curriculum, plans, and processes in instructional design: International perspectives. New York: Lawrence Erlbaum Associates. Krueger, M. W. (1991). Artificial reality. New York: Addison-Wesley.

589

Virtual Reality 2.0 and Its Application in Knowledge Building

Lanier, J., & Biocca, F. (1992). An insider’s view of the future of virtual reality. The Journal of Communication, 42(4), 150–172. doi:10.1111/j.1460-2466.1992.tb00816.x Leuf, B., & Cunningham, W. (2001). The wiki way. Quick collaboration on the Web. Boston: Addison-Wesley. Lombard, M., & Ditton, T. (1997). At the heart of it all: The concept of presence. Journal of ComputerMediated Communication, 3(2), Article 4. Naimark, M. (1990). Realness and interactivity. In B. Laurel (Ed.), The art of human computer interface design (pp. 455-459). Boston: Addison Wesley. Novak, M. (1991). Liquid architectures in cyberspace. In M. Benedict (Ed.), Cyberspace: First steps (pp. 225-254). Cambridge, MA: MIT Press. Paivio, A. (1990). Mental representations: A dual coding approach. Oxford: Oxford University Press. Raitman, R., Augar, N., & Zhou, W. (2005). Employing wikis for online collaboration in the e-learning environment: Case study. Proceedings of the 3rd International Conference on Information Technology and Applications, ICITA 2005 II (pp. 142-146). Washington, D.C.: IEEE Computer Society. Riss, U. V., Cress, U., Kimmerle, J., & Martin, S. (2007). Knowledge transfer by sharing task templates: Two approaches and their psychological requirements. Knowledge Management Research and Practice, 5(4), 287–296. doi:10.1057/palgrave.kmrp.8500155 Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. The American Psychologist, 55(1), 68–78. doi:10.1037/0003-066X.55.1.68

590

Salomon, G. (1994). Interaction of media, cognition, and learning: An exploration of how symbolic forms cultivate mental skills and affect knowledge acquisition. Mahwah: Lawrence Erlbaum Associates. Scardamalia, M. (2002). Collective cognitive responsibility for the advancement of knowledge. In B. Smith (Ed.), Liberal education in a knowledge society (pp. 67-98). Chicago: Open Court. Scardamalia, M., & Bereiter, C. (1994). Computer support for knowledge-building communities. Journal of the Learning Sciences, 3(3), 265–283. doi:10.1207/s15327809jls0303_3 Scardamalia, M., & Bereiter, C. (1996). Student communities for the advancement of knowledge. Communications of the ACM, 39(4), 36–37. doi:10.1145/227210.227220 Scardamalia, M., & Bereiter, C. (2003). Knowledge building. In J. W. Guthrie (Ed.), Encyclopedia of Education (pp. 1370-1373). New York: Macmillan Reference. Scardamalia, M., & Bereiter, C. (2006). Knowledge building: Theory, pedagogy, and technology. In K. Sawyer (Ed.), The Cambridge handbook of the learning sciences (pp. 97-115). New York: Cambridge University Press. Scardamalia, M., Bereiter, C., & Lamon, M. (1994). The CSILE project: Trying to bring the classroom into world 3. In K. McGilly (Ed.), Classroom lessons: Integrating cognitive theory and classroom practice (pp. 201-228). Cambridge, MA: MIT Press/Bradford Books. Shepard, R. N. (1967). Recognition memory for words, sentences, and pictures. Journal of Verbal Learning and Verbal Behavior, 6(1), 156–163. doi:10.1016/S0022-5371(67)80067-7 Sheridan, T. B. (1992). Musings on telepresence and virtual presence. Presence (Cambridge, Mass.), 1(1), 120–126.

Virtual Reality 2.0 and Its Application in Knowledge Building

Slater, D. (2002). Social relationships and identity online and offline. In L. A. Lievrouw & S. M. Livingstone (Eds.), Handbook of new media: Social shaping and consequences of ICTs (pp. 533-546). London: Sage. Steuer, J. (1992). Defining virtual reality: Dimensions determining telepresence. The Journal of Communication, 42(4), 73–93. doi:10.1111/j.1460-2466.1992.tb00812.x Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In S. Worchel & W. Austin (Eds.), Psychology of intergroup relations (vol. 2, pp. 7-24). Chicago: Nelson-Hall Publishers. Tapscott, D., & Williams, A. D. (2006). Wikinomics: How mass collaboration changes everything. New York: Portfolio. Vygotski, L. S., Cole, M., John-Steiner, V., Scribner, S., & Souberman, E. (1978). Mind in society: The development of higher psychological processes. Cambridge, MA: Harvard University Press. Winn, W. (1993). A conceptual basis for educational applications of virtual reality. (HITLab Tech. Rep. R-93-9). Seattle: University of Washington, Human Interface Technology Laboratory. Zeltzer, D. (1992). Autonomy, interaction, and presence. Presence (Cambridge, Mass.), 1(1), 127–132. Zhao, S., & Elesh, D. (2008). Copresence as ‘being with’: Social contact in online public domains. Information Communication and Society, 11(4), 565–583. doi:10.1080/13691180801998995

AdditionAl reAding

Clark, A. (1997). Being There: Putting brain, body, and world together again. Cambridge, MA: MIT Press. Cress, U., & Kimmerle, J. (2008). A systemic and cognitive view on collaborative knowledge building with wikis. International Journal of Computer-Supported Collaborative Learning, 3(2), 105–122. doi:10.1007/s11412-007-9035-z Heeter, C. (1992). Being there: the subjective experience of presence. Presence (Cambridge, Mass.), 1(2), 262–271. Kafai, Y. B. (2005). Constructionism. In R. K. Sawyer (Ed.), The cambridge handbook of the learning sciences (pp. 35-46). New York: Cambridge University Press. Kalawsky, R. S. (1993). The Science of virtual reality and virtual environments: A technical, scientific and engineering reference on virtual environments. Wokingham: Addison-Wesley. Lanier, J., & Biocca, F. (1992). An insider’s view of the future of virtual reality. The Journal of Communication, 42(4), 150–172. doi:10.1111/j.1460-2466.1992.tb00816.x Naimark, M. (1990). Realness and interactivity. In B. Laurel (Ed.), The Art of Human Computer Interface Design (pp. 455-459). Boston: Addison Wesley. Scardamalia, M., & Bereiter, C. (1994). Computer support for knowledge-building communities. Journal of the Learning Sciences, 3(3), 265–283. doi:10.1207/s15327809jls0303_3 Scardamalia, M., & Bereiter, C. (2006). Knowledge building: Theory, pedagogy, and technology. In K. Sawyer (Ed.), The Cambridge handbook of the learning sciences (pp. 97-115). New York: Cambridge University Press.

Boellstorff, T. (2008). Coming of age in Second Life: An anthropologist explores the virtually human. Princeton: Princeton University Press.

591

Virtual Reality 2.0 and Its Application in Knowledge Building

Stanney, K. M. (2002). Handbook of virtual environments: Design, implementation, and applications. Mahwah, New Jersey: Lawrence Erlbaum Associates. Steuer, J. (1992). Defining virtual reality: Dimensions determining telepresence. The Journal of Communication, 42(4), 73–93. doi:10.1111/j.1460-2466.1992.tb00812.x Vince, J. (2004). Introduction to virtual reality. Berlin: Springer.

Key terMs And definitions Environmental Presence: Environmental presence is closely linked with personal and social presence. Different learners represented by their avatars are simultaneously present in the VR, share the same (or similar) awareness about their situation and environment. Immersion: Immersion is the user’s feeling of, so to speak, being immersed in a virtual world which is provided by the technical system. So the concept of immersion not only takes into account technological aspects of a VR, but also emotional, motivational and cognitive processes of focusing attention. Knowledge Building: The concept of knowledge building describes the creation of new knowledge in modern knowledge societies as a socio-cultural process. New knowledge is created in a social process and in concrete situations, and this will occur if a community has reached the boundaries of its existing knowledge, and if members of that community are no longer able to explain experiences in their environment with their existing knowledge. Scardamalia and Bereiter compare that situation with a scientific community in which a group of scientists generates new knowledge and then shares it with the rest of the community.

592

Personal and Social Presence: Users in a VR have to be represented by avatars. This is a requirement both for personal presence of an individual, i.e. the personal feeling of a user to be there in a world created by media, and for social presence of other individuals as sense of being with another. Presence: The term presence refers to the extent to which somebody has the impression to be present in a mediated environment. Presence is a matter of the feeling of being there, i.e. the personal perception of an individual, which depends on the available sensory information, but also on this person’s control of attention, motivational factors and other mental processes. Transduction: The term transduction refers to representations of information which could not normally be perceived by the sensory system of human beings (say, by using different colors for showing a body’s emission of different degrees of warmth). Virtual Reality (VR): VRs are artificial worlds that were generated digitally. In its simple form, a VR is an interface between humans and machines that will allow human beings to perceive computer-generated data as reality. The feature that defines VRs is interaction by a user with the virtual world, or in other words, immediate feedback (output, as immediate as possible) from the system to user input, creating a perception of some reality which is as realistic as possible by using three-dimensional presentation. VR 2.0: VR 2.0 is a combination of technical facilities provided by an online VR with Web 2.0 concepts. VR 2.0 means, on the one hand, an expansion of VRs by adding Web 2.0 features, and on the other hand, the term implies that the degree of presence and immersion of which a VR is capable will not primarily depend on technical features and not necessarily on the number and fidelity of the input and output channels that it uses.

593

Chapter 33

Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education Haya Ajjan University of North Carolina at Charlotte, USA Richard Hartshorne University of North Carolina at Charlotte, USA Richard E. Ferdig Kent State University, USA

AbstrACt In this chapter, the authors provide evidence for the potential of Web 2.0 applications in higher education through a review of relevant literature on educational technology and social networking. Additionally, the authors report the results and implications of a study exploring student and faculty awareness of the potential of Web 2.0 technologies to support and supplement classroom instruction in higher education. Also, using the decomposed theory of planned behavior as the theoretical foundation, the authors discuss factors that inluence student and faculty decisions to adopt Web 2.0 technologies. The chapter concludes with a list of recommendations for classroom use of Web 2.0 applications, as well as implications for policy changes and future research.

introdUCtion The use of Internet technologies such as websites, newsgroups, and e-mail have had a significant impact on the way courses are delivered and designed in higher education (Barnett, Keating, DOI: 10.4018/978-1-60566-384-5.ch033

Harwook, & Saam, 2004). Recently a new wave of Internet technologies, named Web 2.0 technologies (O’Reilly, 2005; Murugesan, 2007), has emerged with the potential to further enhance teaching and learning in many colleges and universities. With the use of Web 2.0 technologies, students are able to access the web for more than just static course information; they are now able to access and create

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education

collective knowledge though social interactions with their peers and faculty (Maloney, 2007). Web 2.0 technologies also enable students to connect multiple pieces of information and in doing so create new information that is shared with others (Maloney, 2007). Web 2.0 technologies have many theoretical affordances to improve teaching and learning (Ferdig, 2007). These affordances include the ability to support scaffolding and active learner participation, provide opportunities for student publication, feedback, and reflection, and the potential for development of a community of learners (Ferdig, 2007). Additionally, while students today are embracing emerging technologies such as cell phones, text messaging, YouTube, wikis, social networks, and other Web 2.0 applications, we also know that many faculty still have not made the switch to these emerging technologies; they prefer course websites and e-mail as their predominant

means of connecting with their students (Ajjan & Hartshorne, 2008). In this chapter, the results and implications of a study exploring student and faculty awareness of the potential of Web 2.0 technologies to supplement classroom learning are discussed. Also, using the decomposed theory of planned behavior (DTPB) as the theoretical foundation (Taylor & Todd, 1995), factors that influence student and faculty decisions to adopt such technologies are examined. This chapter extends the existing literature by providing new insights on factors that influence student and faculty adoption of Web 2.0 technologies. Understanding these factors will be useful in formulating effective strategies and recommendations to increase the likelihood of adoption and effective use of Web 2.0 technologies.

Figure 1. The decomposed theory of planned behavior (**student (or subordinate) influence is only considered in the faculty model)

594

Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education

bACKgroUnd why Use web 2.0 in higher education? Web 2.0 provides online users with interactive services and control over their own data and information (Madden & Fox, 2006; Maloney, 2007). Examples of Web 2.0 technologies include wikis, blogs, instant messaging, internet telephony, social bookmarking, and social networking sites. These new technologies change the way documents are created, used, shared, and distributed and make sharing content among participants much easier than in the past (Dearstyne, 2007). In the study addressed in this chapter, there was a focus on the following four types of Web 2.0 collaboration tools: wikis, blogs, social bookmarks, and social networking. Although many Web 2.0 applications are not designed specifically for educational purposes, Web 2.0 tools have a number of affordances that make them useful in teaching and learning environments and are rooted in strong pedagogical underpinnings of constructivism (Ferdig, 2007). There are at least four important theoretical considerations that indicate social software will be useful tools for teaching and learning. First, social networking tools provide opportunities to scaffold student learning in the student’s Zone of Proximal Development (Brown & Ferrara, 1985; Vygotsky, 1978). The Zone of Proximal Development is the distance between what a student could learn on their own and what they could learn with the assistance of a more knowledgeable other (Vygotsky, 1978). Web 2.0 technologies not only allow more direct interaction between teacher, student, and content, but it also opens up the role of more knowledgeable other to other students, parents, and even the computer (Scardamalia & Bereiter, 1991). A second theoretical consideration for the use of Web 2.0 technologies comes from the notion of learning as active participation in a shared

endeavor with others (Rogoff, 1994; Linn, 1991). Collaboration and cooperative learning can be supported with technology in meaningful ways (Denning & Smith, 1997). These technologies allow users to manage and organize their input in a effective way and thus support constructive learning (Jonassen et al, 1999). Examples of Web 2.0 technologies that promote such collaboration include wikis and collaborative writing spaces (e.g. Google Documents). A third important reason for higher education to consider the use of Web 2.0 technologies is that feedback is critical to learning. As students publish artifacts, teachers “can infer the process by which students transform meanings and strategies appropriated within the social domain” (Gavelek & Raphael, 1996, p. 188). Teachers do not need Web 2.0 technologies to give feedback to their students. However, Web 2.0 technologies provide an authentic environment for students to receive feedback from their teachers and from outside sources. Student blogs are excellent examples of opportunities for students to publish authentic material that receives internal and external feedback (teacher and outsiders). A fourth (but by no means final) theoretical consideration of the use of social software is that “learning occurs through centripetal participation in the learning curriculum of the ambient community” (Lave & Wenger, 1991, p. 100). Social software like Facebook and Myspace provide opportunities for students to create and try out ideas within communities of practice. They are able to explore their identity within society. Although there is relatively little empirical work, these theoretical considerations suggest Web 2.0 tools can and should be explored by educators.

theoretical framework In this study, the decomposed theory of planned behavior (DTPB) was used to examine student and faculty intentions to use Web 2.0 tools in the classroom. The DTPB (Figure 1) originated from

595

Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education

the theory of planned behavior (Ajzen, 1991) and, in the past, has been applied to understand the adoption behavior of information technology tools (Taylor & Todd, 1995; Thompson, Compeau, & Higgins, 2006). DTPB suggests that attitudes, subjective norms, and perceived behavioral control will influence a user’s behavioral intention, which will in turn influence an individual’s actual behavior (Ajzen, 1991). The theory further decomposes the constructs of attitude, subjective norms, and perceived behavioral controls into lower level belief constructs, allowing us to better understand and examine factors that impact the use of new technologies (Taylor & Todd, 1995). This decomposition can generate administrative information about specific factors that influence adoption intention. Therefore, this theoretical framework was selected to explain the adoption intention and use of Web 2.0 technologies to supplement in-class teaching and learning by faculty and students.

Attitude Attitude is the degree to which an individual favors the behavior of interest (Ajzen, 1991). In this chapter, three low-level belief constructs related to attitudinal components are considered: perceived usefulness, perceived ease of use, and compatibility. Perceived usefulness can be defined as the extent to which users believe that the adopted technology will improve his/ her job performance (Davis, 1989). The greater the perceived usefulness, the more likely it is for the user to adopt the new technological application (Rogers, 2003). Ease of use represents the degree to which the technology is easy to use and understand (Rogers, 2003). Technologies that are perceived to be easy to use have a higher possibility of adoption by potential users. Compatibility is the extent to which technology fits with the potential users’ existing values and practices (Rogers, 2003). Tornatzky and Klein (1982) found that an innovation is more likely to be adopted

596

if it is perceived to be compatible with the value system and work practices of an individual. As ease of use, usefulness, and compatibility increase, the attitude toward using the technology is also likely to become more positive.

Subjective Norms Subjective norms examine the perceived expectations from others that influence a user to perform a particular behavior (Ajzen, 1991). When it comes to adopting a new technology, different social groups might have different opinions regarding the adoption of a particular technology (Taylor & Todd, 1995). In the faculty research model, pressures from three groups--superiors, peers (other faculty), and students--were considered. While superiors might feel that adopting Web 2.0 technology may improve student’s learning or satisfaction with a course, other faculty might feel that it requires an undesired change in the current process. Students, on the other hand, might be more supportive since their level of comfort with Web 2.0 technologies tends to be higher than that of most faculty (Prensky, 2001). In the student research model, pressures from two groups--faculty and peers (other students)--were considered. While faculty might feel that the new technology introduces changes to the current teaching process, peers might be supportive of the use of Web 2.0, given that they are typically more comfortable than faculty in using Web 2.0 technologies (Prensky, 2001). One reason for the difference in use could be the age difference between students and faculty. Several studies have shown that younger participants are more likely to use Web 2.0 technologies than older participants such as wikis and social networking (Lenhart & Madden, 2007; Madden & Fox, 2006).

Perceived Behavioral Control Perceived behavioral control captures the user’s perceptions of the availability of required resourc-

Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education

es and opportunities to perform the behavior of interest and is made of three components (Ajzen, 1991). Facilitating conditions make up the first two components and reflect the availability of resources and tools needed to use the technology (Triandis, 1979). Two types of facilitating conditions were considered in this study: the availability of resources (i.e. time and money) and the availability of compatible hardware and software tools. According to Taylor and Todd (1995), the absence of facilitating conditions can negatively impact the intention and usage of technology. The final component is self efficacy, or a reflection of one’s personal comfort when using technology (Bandura, 1982). Greater self efficacy to use technological applications is positively related to behavioral intentions and actual usage (Compeau & Higgins, 1995; Taylor & Todd, 1995).

university in the southeastern United States. The second survey was intended for graduate and undergraduate students at the same large southeastern university. Two email invitations were sent, one to students and one to faculty members at the university inviting them to participate in the survey. Participation in both surveys was completely voluntary. In sum, 136 faculty members participated in the study (Table 1) and 429 students participated in the study (Table 2).

instruments Both survey instruments for faculty and students were designed using the DTPB as a guiding framework (Taylor & Todd, 1995). The survey instruments were then pilot tested by small subsections of the intended samples (faculty and students). The instruments were updated based on their feedback in order to establish its face and content validity (Nunnally, 1978). The two surveys focused on items exploring comfort level with Web 2.0 technologies (blogs, wikis, and social networking, social bookmarking), actual usage of specific Web 2.0 technologies to supplement inclass learning, and attitudes toward specific Web 2.0 technologies. Additionally, the instruments consisted of a series of items using a five point

Methods In order to determine the awareness of students and faculty members of Web 2.0 technologies and their intention to adopt Web 2.0 technologies as tools to supplement in class learning, two surveys were conducted during the fall semester of 2007. The first survey was intended for faculty at a large Table 1. Profile of faculty respondents Variable Gender

Age

Role at university

Value

Frequency

Percentage

Male

61

43

Female

81

57

Under 30

3

2

30-39

46

34

40-49

32

23

Over 50

58

41

Lecturer

28

20

Assistant Professor

53

37

Associate Professor

35

25

Professor

16

11

Other

11

7

597

Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education

Table 2. Profile of student respondents Variable Gender

Age

Year at University

Value

Percentage

Male

166

43

Female

257

57

16-21

168

39

22-27

129

30

28-33

46

11

34-40

32

7

Over 40

51

12

Freshman

61

14

Sophomore

49

12

Junior

69

16

Senior

108

25

Graduate

126

30

Other

11

3

Likert-scale (strongly disagree to strongly agree) to examine factors that influence participant’s intentions to utilize Web 2.0 technologies in an educational setting. Items focused on areas of actual usage, behavioral intention, attitude, ease of use, perceived usefulness, subjective norms, perceived behavioral control, peer influence, superior influence, compatibility, facilitating conditions (technology and resources), and self efficacy. The internal reliability of all measures for both surveys was tested using Cronbach’s alpha and found satisfactory ranges from 0.67 to 0.98 (Nunnaly, 1978).

statistical Procedure for Analysis Descriptive statistics measures were used to understand frequency patterns related to the comfort level, actual usage, and expected benefits of using Web 2.0 technologies. The other focus of this chapter is to understand factors that influence students and faculty behavioral intentions to use Web 2.0 technologies using the DTPB. Thus, and given the multivariate nature of the variables, path analysis models were used to test the relationships proposed by the DTPB (Wright, 1921).

598

Frequency

Path analysis was used to estimate the magnitude of the linkage between variables and to provide information regarding underling causal processes. The findings of descriptive analysis for faculty and students, as well as path analysis results for both faculty and students, are presented in the next section of this chapter.

findings faculty descriptive statistics Some faculty respondents acknowledged that the use of Web 2.0 applications to supplement in-class learning could provide several benefits (Table 3, Figure 2). About 46% felt that the use of blogs would increase the interaction between faculty and students, while 23% felt that the same benefits would be attained from using social networks. A much smaller percentage of faculty respondents (16% and 7%, respectively) felt that wikis or social bookmarks would increase student-faculty interaction. 39% of faculty respondents felt that blogs had the potential to improve student satisfaction with a course. Similarly, 32% felt that the use of

Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education

Table 3. Faculty perceptions of the pedagogical benefits of Web 2.0 applications Improve student learning

Increase student-faculty interaction

Increase student-student interaction

Improve student satisfaction with course

Improve student writing

Easy to integrate

Blogs

47%

46%

52%

39%

41%

46%

Wikis

42%

23%

20%

22%

29%

38%

Social Networks

16%

16%

56%

32%

8%

23%

Social Bookmarks

9%

7%

26%

13%

1%

12%

blogs would increase student satisfaction with a course, while only 22% felt the use of wikis could positively influence student course satisfaction, and only 7% felt the use of social bookmarks would increase student satisfaction with a course. About 41% of the respondents felt that the use of blogs would improve students writing, while 29% felt the same way about wikis, and only 8% held the same opinion about social networking applications, while 4% felt this way about social bookmarks. In terms of integrating specific Web 2.0 technologies with course content, 46% felt that the use of blogs could be easily integrated, while 38% felt that wikis could be easily integrated, 23% felt that social networking tools could be easily integrated, and 12% felt that social bookmarks would be easy to integrate into an existing course structure.

The data indicated that while some faculty participants felt that the use of Web 2.0 applications could provide benefits (Table 4, Figure 3), only a small percentage chose to use them to supplement their in-class instruction. In fact, 55% of the faculty did not use wikis and did not plan to use them in the near future to supplement in-class learning, compared with 20% that either currently use, or plan to use, wikis in the near future. Also, 62% of faculty respondents did not use blogs and do not plan to use them in the near future, compared with only 16% that currently use them, or plan to use them, in the near future. Similarly, 74% of faculty respondents did not use social networks and did not plan to use them in the near future, compared with 9% that either currently use, or plan to use, social networks in the classroom in the near future. Finally, 80% of

Figure 2. Faculty perceptions of the pedagogical benefits of Web 2.0 applications

599

Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education

Table 4. Faculty use of Web 2.0 applications Don’t use and don’t plan to use

Use occasionally

Frequently use

Always use

Blogs

62%

9%

5%

2%

Wikis

55%

20%

4%

6%

Social Networking

74%

6%

1%

2%

Social Bookmarks

80%

13%

1%

1%

Figure 3. Faculty use of Web 2.0 applications

faculty respondents did not use social bookmarking applications and do not plan to use them in the future, compared with 15% that indicated some use of social bookmarks. The low usage of some Web 2.0 technologies (i.e. blogs, social networks, and social bookmarks) among faculty members might be partially explained by their level of comfort with such technologies (Table 5, Figure 4). Most respondents had never used many of these Web 2.0 technologies. In fact, 81% had never used social bookmarks, 56% had never used blogs, and 59% had never used social networks. On the other hand, many faculty members felt more comfortable using wikis, with approximately 72% reporting to have had some experience with these tools.

600

student descriptive statistics As with faculty, some student participants felt that the use of various Web 2.0 applications held a number of pedagogical benefits and would be useful in supplementing their in-class learning experience (Table 6, Figure 5). For example, 69%, 27%, 21%, and 12% felt that blogs, wikis, social networks, and social bookmarks respectively, held potential to improve their learning in a course. Also, approximately 27% felt that the use of blogs would increase the interaction between them and faculty. Likewise, 14% felt that the same benefits would be attained from using social networks, and 24% felt that the use of wikis could improve these interactions. However, only 14% percent felt the use of social bookmarks would increase studentfaculty interaction. 23% of student respondents felt that blogs had the potential to improve student satisfaction with a course. Similarly, 28% felt that

Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education

Table 5. Faculty comfort level in using Web 2.0 applications Never Use

Novice

Competent

Proficient

Blogs

56%

20%

13%

10%

Wikis

28%

26%

27%

18%

Social Networking

59%

17%

13%

11%

Social Bookmarks

81%

6%

6%

6%

Figure 4. Faculty comfort level in using Web 2.0 applications

the use of blogs would increase student satisfaction with a course, while only 18% felt the use of wikis could positively influence student course satisfaction. Only 8% of student participants felt social bookmarks could increase course satisfaction. About 34% of student respondents felt that the use of blogs would improve students writing, while 29% felt the same way about wikis, and only 15% and 4% held the same opinion about social networking and social bookmarking applications, respectively. In terms of integrating the specific Web 2.0 technologies with the course content, 38% felt that the use of blogs could be easily integrated, 45% felt that wikis could be easily integrated, 23% felt that social networking tools could be easily integrated, and 12% felt that social bookmarks would be easy to integrate into an existing course structure. Web 2.0 use data from the student survey indicated slightly more use, or planned future use, than indicated in the results of the faculty respon-

dents. Wikis were the most frequently used Web 2.0 applications with student results indicating that most students, approximately 73%, reported using wikis to supplement their in-class learning (Table 7, Figure 6). On the other hand, and more in line with faculty responses, 71% did not use social bookmarks and do not plan to use them in the future, 56% did not use blogs and do not plan not to use them in the near future, and 46% did not use social networking and do not plan to use social networks for instructional purposes in the near future. Unlike the faculty respondents, there was a seemingly high comfort level of students in the use of the Web 2.0 technologies (Table 8, Figure 7). In fact, 54% reported to have some experience using blogs, 87% have used wikis, and 78% claimed to be comfortable using a social network. However, only 29% of student respondents reported any experience with social bookmarks.

601

Student and Faculty Use and Perceptions of Web 2.0 Technologies in Higher Education

Table 6. Student perceptions of the pedagogical benefits of Web 2.0 applications Improve student learning

Increase student-faculty interaction

Increase student-student interaction

Improve student satisfaction with course

Improve student writing

Easy to integrate

Blogs

27%

27%

27%

23%

34%

38%

Wikis

69%

14%

28%

28%

29%

45%

Social Networks

21%

24%

62%

18%

15%

23%

Social Bookmarks

12%

7%

13%

8%

4%

12%

Figure 5. Student perceptions of the pedagogical benefits of Web 2.0 applications

Path Analysis findings

Behavioral Intention

The findings of the path analysis indicated that the DTPB was useful for explaining much of the variance in the use of Web 2.0 technologies by faculty and students. Additionally, most paths in both models were statistically significant. In this section, the influence of each factor on actual behavior for both student and faculty respondents, as illustrated by path analysis, is discussed (Figure 8 and Figure 9).

For faculty participants, regression results confirmed each of the three factors--attitude, behavioral intention, and subjective norm--explained a significant variance (75.4%) in behavioral intention (adjusted R2). Path analysis confirmed that attitude (β=0.830, t=12.334, P0.05) had no significant effect on the behavioral intention. Finally, path analysis results indicated the perceived behavioral control (β=0.128, t=2.218, P