Human-Computer Interaction. HCI Applications and Services: 12th International Conference, HCI International 2007, Beijing, China, July 22-27, 2007, ... IV (Lecture Notes in Computer Science, 4553) 3540731091, 9783540731092

Citation preview

Lecture Notes in Computer Science Commenced Publication in 1973 Founding and Former Series Editors: Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen

Editorial Board David Hutchison Lancaster University, UK Takeo Kanade Carnegie Mellon University, Pittsburgh, PA, USA Josef Kittler University of Surrey, Guildford, UK Jon M. Kleinberg Cornell University, Ithaca, NY, USA Friedemann Mattern ETH Zurich, Switzerland John C. Mitchell Stanford University, CA, USA Moni Naor Weizmann Institute of Science, Rehovot, Israel Oscar Nierstrasz University of Bern, Switzerland C. Pandu Rangan Indian Institute of Technology, Madras, India Bernhard Steffen University of Dortmund, Germany Madhu Sudan Massachusetts Institute of Technology, MA, USA Demetri Terzopoulos University of California, Los Angeles, CA, USA Doug Tygar University of California, Berkeley, CA, USA Moshe Y. Vardi Rice University, Houston, TX, USA Gerhard Weikum Max-Planck Institute of Computer Science, Saarbruecken, Germany

4553

Julie A. Jacko (Ed.)

Human-Computer Interaction HCI Applications and Services 12th International Conference, HCI International 2007 Beijing, China, July 22-27, 2007 Proceedings, Part IV

13

Volume Editor Julie A. Jacko Georgia Institute of Technology and Emory University School of Medicine 901 Atlantic Drive, Suite 4100, Atlanta, GA 30332-0477, USA E-mail: [email protected]

Library of Congress Control Number: 2007930202 CR Subject Classification (1998): H.5.2, H.5.3, H.3-5, C.2, I.3, D.2, F.3, K.4.2 LNCS Sublibrary: SL 2 – Programming and Software Engineering ISSN ISBN-10 ISBN-13

0302-9743 3-540-73109-1 Springer Berlin Heidelberg New York 978-3-540-73109-2 Springer Berlin Heidelberg New York

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. Springer is a part of Springer Science+Business Media springer.com © Springer-Verlag Berlin Heidelberg 2007 Printed in Germany Typesetting: Camera-ready by author, data conversion by Scientific Publishing Services, Chennai, India Printed on acid-free paper SPIN: 12078042 06/3180 543210

Foreword

The 12th International Conference on Human-Computer Interaction, HCI International 2007, was held in Beijing, P.R. China, 22-27 July 2007, jointly with the Symposium on Human Interface (Japan) 2007, the 7th International Conference on Engineering Psychology and Cognitive Ergonomics, the 4th International Conference on Universal Access in Human-Computer Interaction, the 2nd International Conference on Virtual Reality, the 2nd International Conference on Usability and Internationalization, the 2nd International Conference on Online Communities and Social Computing, the 3rd International Conference on Augmented Cognition, and the 1st International Conference on Digital Human Modeling. A total of 3403 individuals from academia, research institutes, industry and governmental agencies from 76 countries submitted contributions, and 1681 papers, judged to be of high scientific quality, were included in the program. These papers address the latest research and development efforts and highlight the human aspects of design and use of computing systems. The papers accepted for presentation thoroughly cover the entire field of Human-Computer Interaction, addressing major advances in knowledge and effective use of computers in a variety of application areas. This volume, edited by Julie A. Jacko, contains papers in the thematic area of Human-Computer Interaction, addressing the following major topics: • • • • • •

Business Applications Learning and Entertainment Health Applications Work and Collaboration Support Web-Based and Mobile Applications and Services Advanced Design and Development Support The remaining volumes of the HCI International 2007 proceedings are:

• Volume 1, LNCS 4550, Interaction Design and Usability, edited by Julie A. Jacko • Volume 2, LNCS 4551, Interaction Platforms and Techniques, edited by Julie A. Jacko • Volume 3, LNCS 4552, HCI Intelligent Multimodal Interaction Environments, edited by Julie A. Jacko • Volume 5, LNCS 4554, Coping with Diversity in Universal Access, edited by Constantine Stephanidis • Volume 6, LNCS 4555, Universal Access to Ambient Interaction, edited by Constantine Stephanidis • Volume 7, LNCS 4556, Universal Access to Applications and Services, edited by Constantine Stephanidis • Volume 8, LNCS 4557, Methods, Techniques and Tools in Information Design, edited by Michael J. Smith and Gavriel Salvendy

VI

Foreword

• Volume 9, LNCS 4558, Interacting in Information Environments, edited by Michael J. Smith and Gavriel Salvendy • Volume 10, LNCS 4559, HCI and Culture, edited by Nuray Aykin • Volume 13, LNAI 4562, Engineering Psychology and Cognitive Ergonomics, edited by Don Harris • Volume 14, LNCS 4563, Virtual Reality, edited by Randall Shumaker • Volume 15, LNCS 4564, Online Communities and Social Computing, edited by Douglas Schuler • Volume 16, LNAI 4565, Foundations of Augmented Cognition 3rd Edition, edited by Dylan D. Schmorrow and Leah M. Reeves • Volume 17, LNCS 4566, Ergonomics and Health Aspects of Work with Computers, edited by Marvin J. Dainoff I would like to thank the Program Chairs and the members of the Program Boards of all Thematic Areas, listed below, for their contribution to the highest scientific quality and the overall success of the HCI International 2007 Conference.

Ergonomics and Health Aspects of Work with Computers Program Chair: Marvin J. Dainoff Arne Aaras, Norway Pascale Carayon, USA Barbara G.F. Cohen, USA Wolfgang Friesdorf, Germany Martin Helander, Singapore Ben-Tzion Karsh, USA Waldemar Karwowski, USA Peter Kern, Germany Danuta Koradecka, Poland Kari Lindstrom, Finland

Holger Luczak, Germany Aura C. Matias, Philippines Kyung (Ken) Park, Korea Michelle Robertson, USA Steven L. Sauter, USA Dominique L. Scapin, France Michael J. Smith, USA Naomi Swanson, USA Peter Vink, The Netherlands John Wilson, UK

Human Interface and the Management of Information Program Chair: Michael J. Smith Lajos Balint, Hungary Gunilla Bradley, Sweden Hans-Jörg Bullinger, Germany Alan H.S. Chan, Hong Kong Klaus-Peter Fähnrich, Germany Michitaka Hirose, Japan Yoshinori Horie, Japan Richard Koubek, USA Yasufumi Kume, Japan

Robert Proctor, USA Youngho Rhee, Korea Anxo Cereijo Roibás, UK Francois Sainfort, USA Katsunori Shimohara, Japan Tsutomu Tabe, Japan Alvaro Taveira, USA Kim-Phuong L. Vu, USA Tomio Watanabe, Japan

Foreword

Mark Lehto, USA Jiye Mao, P.R. China Fiona Nah, USA Shogo Nishida, Japan Leszek Pacholski, Poland

Sakae Yamamoto, Japan Hidekazu Yoshikawa, Japan Li Zheng, P.R. China Bernhard Zimolong, Germany

Human-Computer Interaction Program Chair: Julie A. Jacko Sebastiano Bagnara, Italy Jianming Dong, USA John Eklund, Australia Xiaowen Fang, USA Sheue-Ling Hwang, Taiwan Yong Gu Ji, Korea Steven J. Landry, USA Jonathan Lazar, USA

V. Kathlene Leonard, USA Chang S. Nam, USA Anthony F. Norcio, USA Celestine A. Ntuen, USA P.L. Patrick Rau, P.R. China Andrew Sears, USA Holly Vitense, USA Wenli Zhu, P.R. China

Engineering Psychology and Cognitive Ergonomics Program Chair: Don Harris Kenneth R. Boff, USA Guy Boy, France Pietro Carlo Cacciabue, Italy Judy Edworthy, UK Erik Hollnagel, Sweden Kenji Itoh, Japan Peter G.A.M. Jorna, The Netherlands Kenneth R. Laughery, USA

Nicolas Marmaras, Greece David Morrison, Australia Sundaram Narayanan, USA Eduardo Salas, USA Dirk Schaefer, France Axel Schulte, Germany Neville A. Stanton, UK Andrew Thatcher, South Africa

Universal Access in Human-Computer Interaction Program Chair: Constantine Stephanidis Julio Abascal, Spain Ray Adams, UK Elizabeth Andre, Germany Margherita Antona, Greece Chieko Asakawa, Japan Christian Bühler, Germany Noelle Carbonell, France Jerzy Charytonowicz, Poland Pier Luigi Emiliani, Italy

Zhengjie Liu, P.R. China Klaus Miesenberger, Austria John Mylopoulos, Canada Michael Pieper, Germany Angel Puerta, USA Anthony Savidis, Greece Andrew Sears, USA Ben Shneiderman, USA Christian Stary, Austria

VII

VIII

Foreword

Michael Fairhurst, UK Gerhard Fischer, USA Jon Gunderson, USA Andreas Holzinger, Austria Arthur Karshmer, USA Simeon Keates, USA George Kouroupetroglou, Greece Jonathan Lazar, USA Seongil Lee, Korea

Hirotada Ueda, Japan Jean Vanderdonckt, Belgium Gregg Vanderheiden, USA Gerhard Weber, Germany Harald Weber, Germany Toshiki Yamaoka, Japan Mary Zajicek, UK Panayiotis Zaphiris, UK

Virtual Reality Program Chair: Randall Shumaker Terry Allard, USA Pat Banerjee, USA Robert S. Kennedy, USA Heidi Kroemker, Germany Ben Lawson, USA Ming Lin, USA Bowen Loftin, USA Holger Luczak, Germany Annie Luciani, France Gordon Mair, UK

Ulrich Neumann, USA Albert "Skip" Rizzo, USA Lawrence Rosenblum, USA Dylan Schmorrow, USA Kay Stanney, USA Susumu Tachi, Japan John Wilson, UK Wei Zhang, P.R. China Michael Zyda, USA

Usability and Internationalization Program Chair: Nuray Aykin Genevieve Bell, USA Alan Chan, Hong Kong Apala Lahiri Chavan, India Jori Clarke, USA Pierre-Henri Dejean, France Susan Dray, USA Paul Fu, USA Emilie Gould, Canada Sung H. Han, South Korea Veikko Ikonen, Finland Richard Ishida, UK Esin Kiris, USA Tobias Komischke, Germany Masaaki Kurosu, Japan James R. Lewis, USA

Rungtai Lin, Taiwan Aaron Marcus, USA Allen E. Milewski, USA Patrick O'Sullivan, Ireland Girish V. Prabhu, India Kerstin Röse, Germany Eunice Ratna Sari, Indonesia Supriya Singh, Australia Serengul Smith, UK Denise Spacinsky, USA Christian Sturm, Mexico Adi B. Tedjasaputra, Singapore Myung Hwan Yun, South Korea Chen Zhao, P.R. China

Foreword

Online Communities and Social Computing Program Chair: Douglas Schuler Chadia Abras, USA Lecia Barker, USA Amy Bruckman, USA Peter van den Besselaar, The Netherlands Peter Day, UK Fiorella De Cindio, Italy John Fung, P.R. China Michael Gurstein, USA Tom Horan, USA Piet Kommers, The Netherlands Jonathan Lazar, USA

Stefanie Lindstaedt, Austria Diane Maloney-Krichmar, USA Isaac Mao, P.R. China Hideyuki Nakanishi, Japan A. Ant Ozok, USA Jennifer Preece, USA Partha Pratim Sarker, Bangladesh Gilson Schwartz, Brazil Sergei Stafeev, Russia F.F. Tusubira, Uganda Cheng-Yen Wang, Taiwan

Augmented Cognition Program Chair: Dylan D. Schmorrow Kenneth Boff, USA Joseph Cohn, USA Blair Dickson, UK Henry Girolamo, USA Gerald Edelman, USA Eric Horvitz, USA Wilhelm Kincses, Germany Amy Kruse, USA Lee Kollmorgen, USA Dennis McBride, USA

Jeffrey Morrison, USA Denise Nicholson, USA Dennis Proffitt, USA Harry Shum, P.R. China Kay Stanney, USA Roy Stripling, USA Michael Swetnam, USA Robert Taylor, UK John Wagner, USA

Digital Human Modeling Program Chair: Vincent G. Duffy Norm Badler, USA Heiner Bubb, Germany Don Chaffin, USA Kathryn Cormican, Ireland Andris Freivalds, USA Ravindra Goonetilleke, Hong Kong Anand Gramopadhye, USA Sung H. Han, South Korea Pheng Ann Heng, Hong Kong Dewen Jin, P.R. China Kang Li, USA

Zhizhong Li, P.R. China Lizhuang Ma, P.R. China Timo Maatta, Finland J. Mark Porter, UK Jim Potvin, Canada Jean-Pierre Verriest, France Zhaoqi Wang, P.R. China Xiugan Yuan, P.R. China Shao-Xiang Zhang, P.R. China Xudong Zhang, USA

IX

X

Foreword

In addition to the members of the Program Boards above, I also wish to thank the following volunteer external reviewers: Kelly Hale, David Kobus, Amy Kruse, Cali Fidopiastis and Karl Van Orden from the USA, Mark Neerincx and Marc Grootjen from the Netherlands, Wilhelm Kincses from Germany, Ganesh Bhutkar and Mathura Prasad from India, Frederick Li from the UK, and Dimitris Grammenos, Angeliki Kastrinaki, Iosif Klironomos, Alexandros Mourouzis, and Stavroula Ntoa from Greece. This conference could not have been possible without the continuous support and advise of the Conference Scientific Advisor, Prof. Gavriel Salvendy, as well as the dedicated work and outstanding efforts of the Communications Chair and Editor of HCI International News, Abbas Moallem, and of the members of the Organizational Board from P.R. China, Patrick Rau (Chair), Bo Chen, Xiaolan Fu, Zhibin Jiang, Congdong Li, Zhenjie Liu, Mowei Shen, Yuanchun Shi, Hui Su, Linyang Sun, Ming Po Tham, Ben Tsiang, Jian Wang, Guangyou Xu, Winnie Wanli Yang, Shuping Yi, Kan Zhang, and Wei Zho. I would also like to thank for their contribution towards the organization of the HCI International 2007 Conference the members of the Human Computer Interaction Laboratory of ICS-FORTH, and in particular Margherita Antona, Maria Pitsoulaki, George Paparoulis, Maria Bouhli, Stavroula Ntoa and George Margetis.

Constantine Stephanidis General Chair, HCI International 2007

HCI International 2009 The 13th International Conference on Human-Computer Interaction, HCI International 2009, will be held jointly with the affiliated Conferences in San Diego, California, USA, in the Town and Country Resort & Convention Center, 19-24 July 2009. It will cover a broad spectrum of themes related to Human Computer Interaction, including theoretical issues, methods, tools, processes and case studies in HCI design, as well as novel interaction techniques, interfaces and applications. The proceedings will be published by Springer. For more information, please visit the Conference website: http://www.hcii2009.org/

General Chair Professor Constantine Stephanidis ICS-FORTH and University of Crete Heraklion, Crete, Greece Email: [email protected]

Table of Contents

Part I: Business Applications The Experiential Preferences of the Online Consumers in Different Internet Shopping Lifestyles Towards Online Shopping Websites . . . . . . . YenLee Chu and Jim Jiunde Lee

3

Guidelines of Quality for Industrial Online Tourism . . . . . . . . . . . . . . . . . . Miguel Cipolla-Ficarra and Francisco Cipolla-Ficarra

12

Provide Context-aware Advertisements with Interactivity . . . . . . . . . . . . . Qin Gao, Pei-Luen Patrick Rau, and Gavriel Salvendy

22

Skills Matter: A Tale of the Anxious Online Shopper . . . . . . . . . . . . . . . . . Yi Maggie Guo and Barbara D. Klein

32

ATM Advertisement and Financial Preferences with Sensory Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Naotsune Hosono, Sayuri Gotanda, Hiromitsu Inoue, and Yutaka Tomita

42

Multi-users and Multi-contextuality – A Mobile Tourism Setting . . . . . . . Carina Ihlstr¨ om Eriksson and Maria ˚ Akesson

48

“My Money in E-Purse” Searching Problems in Self Service User Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Eylem Kılı¸c

58

Traffic Classification - Towards Accurate Real Time Network Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zhu Li, Ruixi Yuan, and Xiaohong Guan

67

First Impressions with Websites: The Effect of the Familiarity and Credibility of Corporate Logos on Perceived Consumer Swift Trust of Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Paul Benjamin Lowry, Tom L. Roberts, and Trevor Higbee

77

Online Ad Intrusiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scott McCoy, Andrea Everard, Peter Polak, and Dennis F. Galletta

86

An Examination of Online Product Comparison Service: Fit Between Product Type and Disposition Style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fiona Fui-Hoon Nah, Weiyin Hong, Liqiang Chen, and Hong-Hee Lee

90

Measuring Service Quality of Online Bookstores with WebQual . . . . . . . . Heejun Park and Seung Baek

95

XIV

Table of Contents

Designing Product List on E-tailing Websites: The Effect of Sorting on Consumer Decision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shun Cai and Yunjie Xu

104

Analysis of Web Page Complexity Through Visual Segmentation . . . . . . . Guangfeng Song

114

An Investigation of Online Group-Buying Institution and Buyer Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chuan-Hoo Tan, Khim-Yong Goh, and Hock-Hai Teo

124

Challenges and Solutions of Multilingual and Translingual Information Service Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hans Uszkoreit, Feiyu Xu, and Weiquan Liu

132

The Antecedents of Online Consumers’ Perceived Usefulness of Website: A Protocol Analysis Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cheng Yi and Zhenhui Jiang

142

A Study on the Characteristics for the Day and Night Time Consumer Groups of Internet Shopping Malls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sung hee Yoon and Soo ae Kweon

150

Part II: Learning and Entertainment Interaction Design Patterns for Classroom Environments . . . . . . . . . . . . . . Henning Breuer, Nelson Baloian, Christian Sousa, and Mitsuji Matsumoto

163

Predicting the Outcome of a Computer Literacy Course Based on a Candidate’s Personal Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andries J. Burger and Pieter J. Blignaut

173

An Essay About the Relevance of Educational Interactive Systems in the Learning Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alessandra Carusi and Cl´ audia Mont’Alv˜ ao

183

Interacting Play—Design as a Metaphor for Developing Interactive Games . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Teng-Wen Chang

190

Usability Assessment of an E-Learning Courseware for Basic Cataloging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . X. Sherab Chen and Tingting Lu

198

Suppressing Competition in a Computer-Supported Collaborative Learning System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kwangsu Cho and Bosung Kim

208

Table of Contents

XV

A Study of Acteme on Users Unexpert of Videogames . . . . . . . . . . . . . . . . Francisco Cipolla-Ficarra

215

A Natural Language Interface for a 2D Networked Game . . . . . . . . . . . . . . Andrea Corradini, Adrian Bak, and Thomas Hanneforth

225

Detecting Learning Difficulties on Ubiquitous Scenarios . . . . . . . . . . . . . . . Marco de S´ a and Lu´ıs Carri¸co

235

Synchronous Reading in Real-Time Environments . . . . . . . . . . . . . . . . . . . . Yoram Eshet and Eran Chajut

245

Design and Development of Mixed Mobile Education System Based on SIFT Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kwangjin Hong and Keechul Jung

255

The Effects of Visual Versus Verbal Metaphors on Novice and Expert Learners’ Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yu-chen Hsu

264

The Effect of Tangible Pedagogical Agents on Children’s Interest and Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jun-hee Kim, Dea-hyun Jung, Haeng-suk Chae, Ji-young Hong, and Kwang-hee Han

270

An Entertainment System for Improving Motivation in Repeated Practice of Musical Instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Itaru Kuramoto, Yuya Shibata, Yu Shibuya, and Yoshihiro Tsujino

278

What Makes Game Players Want to Play More? A Mathematical and Behavioral Understanding of Online Game Design . . . . . . . . . . . . . . . . . . . . De Liu, Xun Li, and Radhika Santhanam

284

Breaking the Traditional E-Learning Mould: Support for the Learning Preference Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fang Liu, Jasna Kuljis, and Lorna Lines

294

KaLeSy-CJ: Kanji Learning System Focusing on Differences Between Chinese and Japanese . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sa Lu, Naoko Yamashita, Hiroyuki Tominaga, Toshihiro Hayashi, and Toshinori Yamasaki

302

Towards an Effective Evaluation Framework for IMS LD-Based Didactic Materials: Criteria and Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . Carmen L. Padr´ on, Paloma D´ıaz, and Ignacio Aedo

312

Mobile Game-Based Methodology for Science Learning . . . . . . . . . . . . . . . Jaime S´ anchez, Alvaro Salinas, and Mauricio S´ aenz

322

XVI

Table of Contents

Group Collaboration and Learning Through Online Assessments: Comparison of Collaborative and Participatory Online Exams . . . . . . . . . Jia Shen, Starr Roxanne Hiltz, and Michael Bieber

332

A New Framework of Usability Evaluation for Massively Multi-player Online Game: Case Study of “World of Warcraft” Game . . . . . . . . . . . . . . Seungkeun Song, Joohyeon Lee, and Insun Hwang

341

A Study of Learners’ Perceptions of the Interactivity of Web-Based Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jui-ni Sun and Yu-chen Hsu

351

Perceived Usefulness and Usability of Weblogs for Collaborating Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yin-Leng Theng and Elaine Lew Yee Wan

361

Support Case-Based Authentic Learning Activities: A Collaborative Case Commenting Tool and a Collaborative Case Builder . . . . . . . . . . . . . Lu Xiao, John M. Carroll, and Mary Beth Rosson

371

The Practice of Combining Cinematic Narrative with 3D Gameplay . . . . Charles Shih-I Yeh

381

Creating Computer Supported Collaborative Learning Activities with IMS LD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dan Yu and XinMeng Chen

391

Face to Face Communications in Multiplayer Online Games: A Real-Time System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ce Zhan, Wanqing Li, Farzad Safaei, and Philip Ogunbona

401

Part III: Health Applications Human Factors: User Experience Design Guidelines for Telecare Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alejandro Rodr´ıguez-Ascaso, Bruno von Niman, Steve Brown, and Torbjørn Sund Usability Issues in the Development of a User Interface for an Alerts and Reminders System for a Nursing Documentation Application . . . . . . Jos´e A. Borges, Nestor J. Rodriguez, Carlos Perez, and Gilberto Crespo Managing Group Therapy Through Multiple Devices . . . . . . . . . . . . . . . . . Lu´ıs Carri¸co, Marco de S´ a, Lu´ıs Duarte, and Jo˜ ao Carvalho

413

418

427

Table of Contents

Challenge for Preventing Medication Errors -Learn from Errors- : What Is the Most Effective Label Display to Prevent Medication Error for Injectable Drug ? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hiroyuki Furukawa

XVII

437

A Study on Display to Improve the Visibility of PTP Sheets . . . . . . . . . . . Akira Izumiya, Michiko Ohkura, and Fumito Tsuchiya

443

Persuasive Interaction Strategy for Self Diet System: Exploring the Relation of User Attitude and Intervention by Computerized Systematic Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Youngho Jeen, Jiyoun Han, Hyodong Kim, Kyungwon Lee, and Peom Park

450

Augmented Virtual Reality for Laparoscopic Surgical Tool Training . . . . Kanav Kahol, Jamieson French, Troy McDaniel, Sethuraman Panchanathan, and Mark Smith Integrating Technology into the Medical Environment of a Developing African Country . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . James Katzenstein and Barbara Chrispin The Analysis of Near-Miss Cases Using Data-Mining Approach . . . . . . . . Masaomi Kimura, Kouji Tatsuno, Toshiharu Hayasaka, Yuta Takahashi, Tetsuro Aoto, Michiko Ohkura, and Fumito Tsuchiya

459

468 474

Ergonomic Study of Medicament Packages for the Improvement of Safety in Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Akinori Komatsubara

484

Industrial Design and Human Factors: Design Synergy for Medical Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Merrick Kossack, Andrew Gellatly, and Alice Jandrisits

489

How New Technologies Can Help Create Better UI’s for Medical Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . William H. Muto and Edmond W. Israelski

499

The Usability Assessment of Pharmaceutical Package . . . . . . . . . . . . . . . . . Michiko Ohkura, Toshiaki Harikae, and Fumito Tsuchiya

508

The Practices of Scenario Observation Approach in Defining Medical Tablet PC Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chien Yu Peng, Wei Shin Kao, You Zhao Liang, and Wen Ko Chiou

518

Interfacing Safety and Communication Breakdowns: Situated Medical Technology Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Svetlena Taneva and Effie Law

525

XVIII

Table of Contents

Medication Errors Caused by Order Entry System and Prevention Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fumito Tsuchiya

535

Motivation in Home Fitnessing: Effects of Immersion and Movement . . . . Joyce Westerink, Marko de Jager, Ronald Bonants, Marijn Bruinink, Jan van Herk, Yvonne de Kort, Wijnand Ijsselsteijn, and Fren Smulders

544

Dependability and Security in Medical Information System . . . . . . . . . . . . Xukai Zou, Yuan-Shun Dai, Bradley Doebbeling, and Mingrui Qi

549

Part IV: Work and Collaboration Support A Role-Based Multiple View Approach to Distributed Geo-Collaboration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gregorio Convertino, Dejin Zhao, Craig H. Ganoe, John M. Carroll, and Mary Beth Rosson

561

The Project Browser: Supporting Information Access for a Project Team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anita Cremers, Inge Kuijper, Peter Groenewegen, and Wilfried Post

571

Design and Implementation of a Personalized Business Activity Monitoring System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hoontae Kim, Yong-Han Lee, Hongsoon Yim, and Nam Wook Cho

581

Visual Feedback: Its Effect on Teleconferencing . . . . . . . . . . . . . . . . . . . . . . Atsunobu Kimura, Masayuki Ihara, Minoru Kobayashi, Yoshitsugu Manabe, and Kunihiro Chihara

591

Document Management and Information Organizing Method Using RFID Tags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hiroyuki Kojima, Ken Iwata, and Naoki Nishimura

601

Affordance Table – A Collaborative Smart Interface for Process Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jari Laarni, Leena Norros, and Hanna Koskinen

611

Experimental Effect Estimation of an Integrated Decision Support System to Aid Operator’s Cognitive Activities for Nuclear Power Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seung Jun Lee and Poong Hyun Seong Designing an Ergonomic Computer Console on the Ship . . . . . . . . . . . . . . Chiuhsiang Joe Lin, Yung-Tsan Jou, Chih-Wei Yang, Chun-Lang Huang, Tse-Jung Chen, and Ming-Zhe Tsai

620

629

Table of Contents

XIX

A Human-Centered Systems Approach to Ship Facility Design . . . . . . . . . Chiuhsiang Joe Lin, Yung-Tsan Jou, Chih-Wei Yang, Shi-Bin Lin, Yao-Hung Hsieh, and Hung-Jen Chen

639

UX Office. A New Software Application for User Experience Services . . . Li Ma and Xiaowei Yuan

649

Some Empirical Results on a Multimedia Work Support System . . . . . . . Jarmo Makkonen and Ari Visa

654

Experimenting Wearable Solutions for Workers’ Training in Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I˜ naki Maurtua, Miren Unceta, and Miguel Angel P´erez

663

Excellence in Ease of Use with Rich Functionality How Enterprise Software Applications with Rich Functionality Can Be Built to Excel in Ease of Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Abbas Moallem

672

Perspectives on Reuse Process Support Systems for Document-Type Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yukiko Morimoto, Hisao Mase, and Hiroshi Tsuji

682

ICT-Mediated Synchronous Communication in Creative Teamwork: From Cognitive Dust to Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vivian Nguyen and Paul Swatman

692

Project Evaluation by E-Mail Communication Pattern . . . . . . . . . . . . . . . . Jugo Noda, Ryosuke Saga, and Hiroshi Tsuji

702

E-Stores Design: The Influence of E-Store Design and Product Type on Consumers’ Emotions and Attitudes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Talya Porat, Rinat Liss, and Noam Tractinsky

712

Design Discussion of the [braccetto] Research Platform: Supporting Distributed Intensely Collaborating Creative Teams of Teams . . . . . . . . . . Claudia Schremmer, Alex Krumm-Heller, Rudi Vernik, and Julien Epps Augmented Desk System: The Information Table of Collaborative and Cooperative (InfoTable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cheng-Wei Shih, Shu-Feng Pan, Ming-Xian Lee, and Hong-Sheng Chen

722

735

A Framework for Mobile Services Supporting Mobile Non-office Workers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bente Skattør, Lasse Berntzen, Tore Engvig, and Per Hasvold

742

Process-Oriented User Support for Workflow Applications . . . . . . . . . . . . . Dian Tan and Hartmut Wandke

752

XX

Table of Contents

Decentralized Information Aggregation and Central Control in Networked Production Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simon Thiel, Thomas Schlegel, and Dieter Spath

762

With User Interface Design from Socio-Cultural Trend to a Innovative Office Phone Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Claude Toussaint

771

Complexity in Enterprise Applications vs. Simplicity in User Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Matthias Uflacker and Daniela Busse

778

Focus+Roles: Socio-Organizational Conflict Resolution in Collaborative User Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Davy Vanacken, Chris Raymaekers, Kris Luyten, and Karin Coninx

788

Information Complexity in Air Traffic Control Displays . . . . . . . . . . . . . . . Jing Xing

797

HOPE: Extensible System for Automatic & Periodic Diagnosis of Offshore Software Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xuan Zhang, Lei Zhang, Meiping Chai, Shigeru Miyake, and Ryota Mibe

807

Playwright Agent as Adaptive Scripts for Organization-Communication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yu-Pei Yeh and Teng-Wen Chang

816

A User-Oriented Assessment of Enterprise Information Systems . . . . . . . . Eun Jung Yu, Youngjoon Choi, and Choon Seong Leem

825

The Role of Cognitive Styles in Groupware Acceptance . . . . . . . . . . . . . . . NorolHamiza Zamzuri and WanAdilah WanAdnan

835

Part V: Web-Based and Mobile Applications and Services Effect of Physical Consistency of Web Interface Design on Users’ Performance and Satisfaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ahamed AlTaboli and Mohammad Raafat Abou-Zeid A Semiotic-Based Framework for the Development of Tailorable Web Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rodrigo Bonacin, M. Cecilia C. Baranauskas, and Thiago Medeiros dos Santos Travelling Stories: Mobile Applications for Storytellers . . . . . . . . . . . . . . . . Anxo Cereijo Roib´ as, Nina Sabnani, and Riccardo Sala

849

859

869

Table of Contents

XXI

Analysis of Internet Users’ Interests Based on Windows GUI Messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jinhyuk Choi, Geehyuk Lee, and Yonghoon Um

881

Deconstructing Web Experience: More Than Just Usability and Good Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chandra Harrison and Helen Petrie

889

Report on Project to Identify Excellent Local Government Web Sites: An Attempt to Educate Web Systems Developers on the Importance of Human-Centered Design Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seiji Hayakawa, Haruhiko Urokohara, Toshikazu Shinohara, Naoto Aizawa, and Yayoi Kitamura

899

A Survey of Factors Influencing People’s Perception of Information Security . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ding-Long Huang, Pei-Luen Patrick Rau, and Gavriel Salvendy

906

Towards Web Services Oriented Unified Supervisory HCI . . . . . . . . . . . . . Djilali Idoughi and Christophe Kolski

916

Everyday Storytelling: Supporting the Mediated Expression of Online Personal Testimony . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aisling Kelliher and Glorianna Davenport

926

Chat-Robot Based Web Content Presentation Interface and Its Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yumi Kikuchi(Tomioka), Kota Hidaka, Shinya Nakajima, and Minoru Kobayashi Understanding Influence of Mobile Internet Services on Life Behavior of Mobile Users . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sang Min Ko, Yong Gu Ji, and Dongsoo Kim Interactive Design of Memory Sharing Applications for Families . . . . . . . . Minako Kubo, Akihiko Kodama, Etsushi Takaishi, Rikio Chiba, Keita Watanabe, and Michiaki Yasumura Personal Life Logger and Belonging Monitor Using Reliable ZigBee Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . KwangHee Lee, KyoungJu Noh, and Changseok Bae A Three-Level Approach for Analyzing User Behavior in Ongoing Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Enric Mor, Muriel Garreta-Domingo, Juli` a Minguill´ on, and Sheena Lewis Relevance Measurement on Chinese Search Results . . . . . . . . . . . . . . . . . . . Hua Qin and Pei-Luen Patrick Rau

934

944 954

961

971

981

XXII

Table of Contents

Multiple Viewed Search Engine for an e-Journal – A Case Study on Zoological Science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Takahiro Seki, Taiki Wada, Yasuhiro Yamada, Nozomi Ytow, and Sachio Hirokawa Exploring Multi-dimensional Conceptualization of Social Presence in the Context of Online Communities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kathy Ning Shen and Mohamed Khalifa

989

999

Does the Web Design Disconnect the Emotional Connection? . . . . . . . . . . 1009 Tara Shrimpton-Smith and Bieke Zaman Ideas for Mobile Services Created and Explored in an Ethnographical Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1019 Bente Skattør The Effects of Mobile Service Quality and Technology Compatibility on Users’ Perceived Playfulness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1029 Felix B. Tan and Jacky P.C. Chou Emotional Web Usability Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1039 Sylvia Tzvetanova, Ming-Xi Tang, and Lorraine Justice A Successful Field Test of a Mobile and Multilingual Information Service System COMPASS2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1047 Hans Uszkoreit, Feiyu Xu, Weiquan Liu, J¨ org Steffen, Ilhan Aslan, Jin Liu, Christel M¨ uller, Bernhard Holtkamp, and Manfred Wojciechowski Love Dimensions on the Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1057 Hannu Vanharanta, Jussi Kantola, and Waldemar Karwowski

Part VI: Advanced Design and Development Support A Replicated CSCW Framework for Multi-user Interaction in 3D Collaborative Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1065 Fazhi He, Huajun Liu, Xiantao Cai, and Shuxu Jing A Review of Possibilistic Approaches to Reliability Analysis and Optimization in Engineering Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1075 Li-Ping He, Hong-Zhong Huang, Li Du, Xu-Dong Zhang, and Qiang Miao An Interactive Evolutionary Design System with Feature Extraction . . . . 1085 Xu Jiang, Sun Shouqian, Tan Zhengyu, and Shi Fuqian

Table of Contents

XXIII

Verification of Development of Scenarios Method and Visual Formats for Design Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1095 Haifu Lin, Masaaki Kurosu, Hideaki Takahashi, Hiroshi Kato, and Takeshi Toya A Camera-Based Multi-touch Interface Builder for Designers . . . . . . . . . . 1102 Han-Hong Lin and Teng-Wen Chang Design of Change-Absorbing System Architecture for the Design of Robust Products and Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1110 Sanghyun Oh, Beomjin Park, Shinae Park, and Yoo Suk Hong A Multifunctional VR-Simulator Platform for the Evaluation of Automotive User Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1120 Tony Poitschke, Markus Ablaßmeier, Stefan Reifinger, and Gerhard Rigoll Stripe-Cache: An Efficient Cache Scheme for Building Multimedia Oriented RAID System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1130 QinQi Wei, ChangSheng Xie, and Xu Li DiFac: Digital Factory for Human Oriented Production System . . . . . . . . 1140 Marco Sacco, Claudia Redaelli, Carmen Constantinescu, Glyn Lawson, Mirabelle D’Cruz, and Menelaos Pappas INT-MANUS: Interactive Production Control in a Distributed Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1150 Thomas Schlegel, Aravind Srinivasan, Maxim Foursa, Manfred Bogen, Rejin Narayanan, David d’Angelo, G´eza Haidegger, Istvan Mezgar, Joseph Canou, Damien Sall´e, Fabrizio Meo, Jon Agirre Ibarbia, and Anja Herrmann Praturlon 3D Modeling and Design Supported Via Interscopic Interaction Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1160 Frank Steinicke, Timo Ropinski, Gerd Bruder, and Klaus Hinrichs Capturing User Interface Events with Aspects . . . . . . . . . . . . . . . . . . . . . . . 1170 Yonglei Tao Research of Distributed Data Mining Association Rules Model Based on Similarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1180 Shengjun Xue and Zhengqiu Lu Applications of Virtual Reality in Product Design Evaluation . . . . . . . . . . 1190 Jilin Ye, Saurin Badiyani, Vinesh Raja, and Thomas Schlegel Performance Evaluation for Automatic Protection Switching in a CDMA-TDD Wireless Communication System . . . . . . . . . . . . . . . . . . . . . . . 1200 Li Zhang, Hiroyuki Okamura, and Tadashi Dohi

XXIV

Table of Contents

A Routing Algorithm for Random Error Tolerance in Network-on-Chip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1210 Lei Zhang, Huawei Li, and Xiaowei Li Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1221

Part I

Business Applications

The Experiential Preferences of the Online Consumers in Different Internet Shopping Lifestyles Towards Online Shopping Websites YenLee Chu and Jim Jiunde Lee Institute of Communication Studies National Chiao Tung University 1001 Ta Hsueh Road, Hsinchu, Taiwan 300, ROC [email protected], [email protected]

Abstract. Consumers nowadays are too often to purchase or consume through internet, thus, their requirements towards online shopping are seemingly higher than before. Subsequently the online shopping vendors are supposed to satisfy the consuming experience of shoppers that come from different kinds of levels in order to compete in this market. This research is based on Schmitt’s (1999/2003) Strategic Experiential Modules and Lifestyles Theory. By using the quantitative method, this study figured out the experiential strategy and designed elements that were preferred by the consumers from dissimilar lifestyle who surfed in different kind of online shopping websites. According to the results of this study, Taiwan online consumers could be divided into three lifestyle groups: online shopping avoider, online shopping lovers and fearful online shoppers. Among the online specialty consumers, the online shopping avoider, the online shopping lovers and the fearful online shoppers paid attention on sensory experiences, affective experiences and relational experiences, respectively. Whereas, among the online shopping mall shoppers, online shopping avoider and online shopping lovers insisted on sensory experiences and interaction experiences respectively. Keywords: Lifestyles, Experiential Marketing, Online Shopping.

1 Introduction Nowadays, consumers highly rely on online shopping, and are gradually strict with the quality of it. In 2005, Taiwan Market Intelligence Center analyzes online shoppers in Taiwan, and also compares their satisfaction in 2001, 2004 and 2005. According the result of the questionnaire, these online shoppers can be grouped into four categories, which are "very satisfied’, 'rather satisfied",” average" and "dissatisfied" As to the proportion of” very satisfied" group, the momentum decreases from 79% to 60% during the 2001-2005 period. The proportion of “average" group goes up from 5% to 35.5%.That is because the competition between online shops is more and more drastic, and information explosion can always distract consumers' attention. Moreover, consumers expect to obtain deeper satisfaction. J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 3 – 11, 2007. © Springer-Verlag Berlin Heidelberg 2007

4

Y. Chu and J.J. Lee

Unfortunately, many of the online store operators or designers only pay attention on the functional attributes of products, neglect the importance of experience. In the age of information explosion it seems difficult to attract consumer attention. As Zaitman stated in 1997 at Procter & Gamble's Future Forces Conference in Cincinnati, Ohio, "Consumer preference and motivation are far less influenced by the functional attributes of products and services than the subconscious sensory and emotional elements derived by the total experience.” Therefore, director of Yahoo! Value Lab, Tim Sanders suggests that “experience” is the basis of a new economy. The commercial websites are requisite to provide an unforgettable and amazing experience in order to prove their worthiness. Furthermore, Zarem(2000) addresses that creating experience is the new battle ground because people today are motivated by experience. Besides, based on the idea “the more you know and understand about consumers, the more effectively you can communicate and market to them” (Plummer, 1974), the study of people’s values and lifestyles has become a standard tool for both social scientists and marketers around the world. Bellman, Lohse & Johnson(1999) pointed out that the most important information for predicting shopping habits—online and offline—are measures of consumer lifestyles, not demographics. In other words, to effectively conduct a shopping website, we should be acquainted with the consumer lifestyles and characters. Hence, this research is based on Schmitt’s (1999/2003)SEMs, Strategic Experiential Modules and Lifestyles Theory. By using the quantitative method, this study figured out the experiential strategy and designed elements that were preferred by the consumers from dissimilar lifestyle who surfed in different kind of online shopping websites. Consequently, these results will probably provide the website designers an immediate and accurate direction, practically.

2 Literature Review 2.1 Experiential Marketing and Strategic Experiential Modules Experiential Marketing is distinct from conventional market, which is emphasizing product features and benefits, and considering consumer rational. Experiential Marketing centers on customer experience, it provides consumer sensory, affective cognitive, interaction, relational function to replace the value of function. (Schmitt 1999, 2003) As Schmitt (1999, 2003) mentioned in his Experiential Marketing, He takes experiential forms as Strategic Experiential Modules, SEMs. Strategic Experiential Modules includes Sense, Feel, Think, Act (interaction) and Related, which are strategic foundations of Experiential Market. Every Strategic Experiential Modules are provided with distinct structure and marketing principle, for example, the strategy of Sense, Feel, and Think provides consumer stimulation consciousness, generating their emotional connection to the company, and then encourage team to more complex thinking. Otherwise, the strategy of Interaction and Related is to design and create customer’s experience related to body, action modules and lifestyle, this strategy makes consumer produce action activity with experience, and offer them some communication relation experience, like console, assistance, value, etc. Now the following are the brief accounts of these five experience forms:

The Experiential Preferences of the Online Consumers

5

：

Sense stimulates as many of the five senses or modalities as possible virtually through sight/vision (visual), sound (aural or auditory), touch (tactile or hap tic), smell (olfactory), and taste (gustatory), providing customer aesthetic pleasure and excitement. (Schmitt 1999, 2003). Feel Emphasize the store and purchase atmosphere, try to stresses the emotional component of shopping (Schmitt 1999, 2003). This includes how a web site places users in a certain mood, makes them respond in an emotional fashion, and appeals to their feelings (Schmitt 1999, 2003). Think Engages users in creative and problem-solving ways and impacts their thinking (Schmitt 1999, 2003).These experiences include how a web site or shop intrigues users, stimulates their curiosity, and appeals to their creative cognition (Schmitt 1999, 2003). Act (Interaction) Provide creative associations as well as lifestyle marketing. Through increasing sensory experience and opportunities of interact others, It affects customer’s action, lifestyle and enrich user’s life. (Schmitt 1999, 2003). Relate Allow individual make connection with ideal-self, others and social culture, meanwhile, establish the sense of identity and belonging. (Schmitt 1999, 2003).

：

：

：

：

These Experience form varies according to the degree of customer’s engaging and immersion. The seller must apply these forms sophisticatedly to maximize the whole effect of experiential marketing. (Schmitt 1999, 2003). 2.2 Internet Lifestyles There is a lot amount of researches about internet shopping based on the approaches of socio-economic, and demographic, etc. However, in the flourishing of internet shopping nowadays, socio-economic, demographic is not the best approach and indication to developing Internet shopping market. That is because one’s above-the-line and belowthe-line personality would be different, we must realize consumer’s profile through their lifestyle on the web, and then we can make the correct marketing decision. Therefore, to respond the flourishing of internet, some researchers propose the ideas about Internet-related psychographic or Internet shopper lifestyle, they consider that demographic and socio-economic descriptors have been neither the most effective in developing segments (Wedel and Kamakura, 2000) nor a good predictor of the propensity to buy on-line (Vellido,2000). For example: what the past researches believe seems more closely related to actual online purchase behavior. Those who actually purchase online appear to have been using the Internet for a longer time (Dahlén, 1999; Novak, Hoffman and Yung, 2000), to be more frequent web users (Hoffman, Kalsbeek and Novak, 1996) and to spend more time on the Internet (Rangaswamy and Gupta 1999). The research of Swinyard & Smith 2003 uses the dimension of internet user’s lifestyle of shopping, and they took US online household as research object, they analyzed the reason why their research object are internet shopper or non-shopper, and their difference in internet and shopping behavior. Their research attempted to arrange a lifestyle profile of internet shopper/non-shopper and provide a strategy of online store design according to different shopping group.

（

）

6

Y. Chu and J.J. Lee

This questionnaire was mailed to a probability sample of 4000 US online households, of which 1738 (43.5 percent) replied and e-mailed to 20,000 e-mail addresses, of which 2477 (12.4 percent).The instrument contains 38 Internet shopping psychographic statements (Internet shopper lifestyle scale), 14 measures of Internet behavior, and 13 themes of Internet usage. As the result shows, Internet shoppers were divided into four segments: shopping lovers (11.1 percent), Internet explorers (8.9 percent), suspicious learners (9.6 percent), and business users (12.4 percent). Internet non-shoppers were similarly identified as: fearful browsers (10.7 percent), shopping avoiders (15.6 percent), technology muddlers (13.6 percent), and fun seekers (12.1 percent). Shopping Lovers (11.1%) ： their age are general younger, frequently buy online and really enjoy doing so. To them, Internet shopping is a new and activity of fun.

：

Adventuresome Explorers(8.9%) this kind of user are varied in their internet behavior, and it is most obvious in shopping behavior, they would seek other advise on internet, preview the product, search for lowest price in everything they buy before they shop on line, and they prefer that their order could be send to them directly.

：

Suspicious Learners (9.6%) They are just learning how to use the Internet. They are frustrated by it, struggling to complete tasks found easy by others. To become a significant on-line shopper group, this segment needs direction and hands-on guidance.

：

Business Users (12.4%) They mainly use the Internet for business purposes and look at the Internet in terms of what it can do for their professional life. But they don’t often make personal online purchases.

：

Fearful Browsers (10.7%) They are very computer literate and often practice “Internet-window-shopping.” However, being afraid of their credit-card number stolen, shipping charges, and wishing they could see products in person before they buy, they less shop on internet.

：

Shopping Avoider(15.6%) shopper in this group prefer experience the pleasure of shopping personally, they hate online shopping and do not understand its procedure, so they are very difficult to turn into online shoppers.

：

Technology Muddlers(19.6%) They have lower computer literacy,and they have a little excitement about increasing their computer and on-line comfort level.

：

Fun Seekers (12.1%) Members of this segment look to the Internet for its entertainment value. But because of their consideration to the safety of internet shopping and privacy, and their lower economic foundation, they do not shop on line. Furthermore Brengman, Geuens, Weijters, Smith, & Swinyard(2003) replicate the study of Swinyard & Smith 2003 in Belgium, to test the Internet shopper lifestyle scale validity and evaluate the similarity and difference of online shopper/nonshopper in Belgium and America. However, considering national and cultural condition of Belgium and its difference between America, the researcher revised Swinyard & Smith’s 2003 questionnaire and used online e-mail survey, they e-mail 20000 to e-mail user, and of which 11% replied. According to the result of research, there is not major difference in the classification of online shopper in Belgium and in America. The classifications are: Shopping

（ ）

（ ）

The Experiential Preferences of the Online Consumers

7

Lovers, Business Users, Suspicious learners and Tentative Shoppers. Especially, the Shopping Lovers and Business Users have almost same lifestyle with these shoppers in America. However, Suspicious learners in Belgium do not use credit card, because of their distrust with online store, it is different in America. Furthermore, in Swinyard & Smith 2003 research, the group of Adventuresome Explorers is replaced by Tentative Shoppers, which is willing to shop online, use credit card but not interested in online shopping. As for internet non-shopper in Belgium, this research classifies them as: Fearful Browsers, Positive Technology Muddlers, Negative Technology Muddlers and Adventurous Browsers, one the Fearful Browser is similar to Swinyard & Smith (2003) classification. It is a cross-culture difference, leading different internet non-shoppers, and their lifestyle. Therefore issue of cross-cultural is an influential factor of the definition of consumer’s lifestyle in specific region, and also of the design strategy of relative experience. The articles researching shopping style of internet user is very rare in Taiwan to help designer for creating electronic commercial website, the internet user shopping style is what we try to analyze in this article, furthermore, we use Swinyard & Smith’s approach (2003), emphasizing the behavior of internet use, shopping and preference of user. We refer Swinyard & Smith (2003) to design our questionnaire. The research questions are:

（

）

1. What is the distinction and characteristic of internet shopper’s lifestyle in Taiwan? 2. What are the experience forms (sensory experiences, affective experiences, cognitive experiences, interaction experiences and relational experiences) adopted by consumer with different lifestyle when they shop on online specialty shop and online shopping mall?

3 Research Method The research boundaries in this article are online specialty shop (for example, online bookstore) and online shopping mall (for example, Yahoo! Online Shopping Mall), so Internet users who had shopped online in these two categories are our research object. As for sampling method, because of the unclear size and boundary of the matrix in the Internet, and the strong diversity in internet, convenience sampling is more available for this research. On-line survey used in this research, the period is 12 days, from December 20th to 31st, 2005, recycled questionnaire are 396, 350 of these are valid. We used SPSS10.0 for windows to process the quantitative analysis and the application methods includes Reliability analysis, Descriptive statistic, Factor analysis, Cluster analysis, Cross tab, Chi-square Test and One Way ANOVA.

4 Results 4.1 Internet Lifestyles

：

Through the Factor analysis, this article classified five major structures love to shop online, fear online financial theft, internet convenience, love the store atmosphere, hassle to return merchandise bought on-line and the online shipping charges, then

8

Y. Chu and J.J. Lee

using Cluster analysis to define the internet user’s group with different internet lifestyle, the title of these group is according to the factors of their Internet Usage, demographic characteristics, online shopping related lifestyle, the present situation of online shopping, etc . In the result, three distinct lifestyle of Internet shopper are defined: online shopping avoider, online shopping lovers, and fearful online shoppers. Online shopping avoider: They prefer real store and personal experience of shopping, among the three groups. They like the atmosphere and pleasure of shopping in real store, enjoy the service there, they consider the service system of real stores is better the online ones, therefore, they do not identify on line shopping, neither agree online shopping as a joyful experience. They seldom shop online. For example, in our research, among the users who in this group, most of them expend less than NT1000 dollars purchasing things online. Comparing Swinyard &Smith’s research, this group is similar to their Shopping Avoider, which enjoy personal shopping experience, detesting online shopping, they are both difficult to persuade to shop online. Online shopping lovers: This group, which prefers to shop online most, is mainly composed by students. Their amount on times and money spending on internet shopping is more than the other researching groups (most of them spend NT10000 to NT50000 on internet shopping). Concerning user in this group, online shopping is a very interesting way for shopping. They like browsing all kinds of merchandises on the web, searching a lower price; they consider online shopping is quite enjoyable and convenient. Online shopping lover is similar to Shopping Lovers in others’ research (Swinyard&Smith,2003;Brengman,Geuens,Weijters,Smith,&Swinyard,2003), they are mostly young user, and interested in shopping online, they are also the ideal opinion leaders (Brengman,Geuens,Weijters,Smith,&Swinyard,2003), the website manager should pay more attention on them. Fearful online shoppers: Users in this group are mostly in the age bracket from 26 to 30, and the majorities are working people. They would search on the web, and also consider internet shopping as a new and interesting way of shopping, but they have misgiving about the delivery and merchandise-returning system of online shopping, their desire for shopping online is uncertain. This group is similar to Fearful Browsers of Brengman,Geuens,Weijters,Smith,&Swinyard,2003.Website designer would convince them choose to merchandise on internet, if they are sure about the safety and complete deliver-return system. 4.2 Experiential Preferences In our research, we found that it is different experience for each group shopping in different shopping website. Online shopping avoider: As for online shopping avoider, no matter rather are in online specialty shop or online shopping mall, they care sensory experiences most. They experience the aura through music, smell, decoration of the actual store; seeing the actual goods, they can even try on when they are choose clothes. It is a pleasure for them to shop at an actual store, it is about sense perception, which can not completely be experienced on the web, and usually shopper can only experience the

The Experiential Preferences of the Online Consumers

9

merchandise by imaging. If the website could enhance the experience more than imaging, such as, offering online conversation with service staff, comfortable music, even animation, video, or 3D image of their product, it would help for the online store to attract the online shopping avoider. Online shopping lovers: in the online specialty shop, the online shopping lovers mind affective experiences most, they like a website which can offer the feeling of free, comfortable, satisfaction and thoughtful, Norman (2004) believes that emotion is influential in the process of data processing, even change the module. Other psychologists also confirms that the emotional transfer influence attitude, memory and decision of action, conspicuously(Morris, Woo, Geason, & Kim,2002； Allen, Machleit&Kleine,1992), when a website could offer the user a enjoyable searching experience, user’s trustfulness, loyalty, and satisfaction to the website will raise. Therefore, in the approach to marketing, online specialty shop manager should set up a free, satisfying and thoughtful experience for their web store. In the online shopping mall, online shopping lovers care interaction experience most. Interaction experience can be derived from some service, such as FAQ, ordertracking, keyword searching, nearest selling location searching, user friendly interface, completed guidance, personal assistance, e-coupon, promotion campaign, e-mail feedback, etc. In a website, the greater is its degree of interaction, the better its ménage achievement, and also, consumer’s attitude to the website would be more positive. (Teo, Oh, Liu & Wei, 2003). Fine interaction experiences are crucial for online shopping lovers, especially in a complex online shopping mall, whose production information is enormous. Designer of online shopping mall should offer more interaction experience according to the online shopping lover group, to enhance their desire to shop on their website. Fearful online shoppers: fearful online shoppers care most about relational experience when shopping on online specialty; it is concerning their misgiving about the safety of online shopping. According to the research of Yen, Yung-shen (2000), consumer with a lower credibility is more sensitive to professional qualification of website. we suggest the designer of online specialty shop to offer more relational experiences design issues, such as, sharing the reflections of shopping, information of buyer’s appraisal, establishing a higher brand values of the website, etc, in order to built user’s identity to the website and connection to the other.

5 Concluding Discussion According to the result, user with different lifestyle, attach their importance on different experience, when shopping on different internet shop, therefore, the website manager could design diversified experience situation in terms of distinct user lifestyle, as we analyze and classify in this article. For example, online specialty shop and online shopping mall would offer more physically sensible experience as their market approach for online shopping avoiders: the language should be easy to read, visually aesthetics feeling, and with proper text size; the color should be design with consistence; the background music should match the theme of different group. Furthermore,

10

Y. Chu and J.J. Lee

on the subject of fearful online shoppers, the raise of their identity to the website is the key to attract them to shop on the online specialty shop, it is related to the label position, the service after merchandize, the function for user to share their feel and experience, etc. Because of the method of questionnaire design (literal description), a gap between the description and consumer’s cognition would be existing in this research, therefore, an example website for the research object is suggested for other research following this topic, secondly, different experience strategies is the elements for building a experiential website, they should finally come to a ultimate state of total experiences of flow experience (Schmitt, 1999/2000). Mahfouz(2004) also claims, sensory, affective and cognitive experience strongly and positively influenced flow experience in terms of control and cognitive enjoyment, and moderately impacted attention focus (Hoffman & Novak, 1996 ), we suggest the follow-up research should focus on the influence of experience strategies to different kind on website, and to the flow experience to user with different lifestyle, it would help us to understand how to design the most ideal experiential online shopping website.

References 1. Allen, C.T., Machleit, K.A., Kleine, S.S.: A comparison of attitudes and emotions as predictors of behavior at diverse levels of behavioral experience. Journal of Consumer Research 18(4), 493–504 (1992) 2. Bellman, S., Lohse, G.L., Johnson, E.J.: Predictors of online buying behavior. Communications of the ACM 42(12), 32–38 (1999) 3. Brengman, M., Geuens, M., Weijters, B., Smith, S.M., Swinyard, W.R.: Segmenting Internet shoppers based on their web-usage-related lifestyle: a cross-cultural validation [Electronic Version] (2003) Retrieved 1/7/2005 from http://ideas.repec.org/p/rug/rugwps/ 03-205.html 4. Dahlén, M.: Closing in on the Web consumer. In: Bohlin, Brodin, Lundgren, Thorngren. (eds.) Ch.12, Convergence in Communications & Beyond, Elsevier Science, Amsterdam (1999) 5. Hoffman, D.L., Novak, T.P.: Marketing in hypermedia computer-mediated environments: conceptual foundations. Journal of Marketing 60, 50–68 (July 1996) 6. Mahfouz, A.Y.M.: Interactive Eshopping Experience: An Empirical Investigation. Texas A&M University Texas (2004) 7. Morris, J., Woo, C., Geason, J., Kim, J.: The power of affect: Predicting intention. Journal of Advertising Research 42(3), 7–17 (2002) 8. Norman, D.: Emotional Design. Basic Books, New York (2004) 9. Novak, T.P., Hoffman, D.L., Yung, Y.: Measuring the Customer Experience in Online Environments: A Structural Modeling Approach. Marketing Science 19(1), 22–42 (2000) 10. Plummer, J.: The Concept and Applica tion of Lifestyle Segmentation. Journal of Marketing 38, 35–58 (1974) 11. Rangaswamy, A., Gupta, S.: Innovation Adoption and Diffusion in the Digital Environment: Some Research Opportunities. eBusiness Research Center Working Paper (1999) 12. Schmitt, B.H.: Customer Experience Management: A Revolutionary Approach to Connecting With Your Customers. John Wiley & Sons, Inc., Hoboken, NJ (2003) 13. Schmitt, B.H.: Experiential Marketing: How to Get Customers to Sense, Feel, Think,Act, and Relate to Your Company and Brands. The Free Press, New York (1999)

The Experiential Preferences of the Online Consumers

11

14. Swinyard, R., Smith, S.M.: Why people (don’t) shop online: a lifestyle study the internet consumer. Psychology & Marketing 20(7), 567–597 (2003) 15. Teo, H., Oh, I., Liu, C., Wei, K.K.: An empirical study of the effects of interactivity on web user attitude. International Journal of Human-Computer Studies 58, 281–305 (2003) 16. Vellido, A.: A methodology for the characterization of Business-to-consumer Ecommerce, Ph.D., Thesis Liverpool John Moores University (September 2000) 17. Wedel, M., Kamakura, W.: Market segmentation: conceptual and methodological foundations, second edition, Kluwer Academic Publishers (2000) 18. Yen, Y.-s.:An Empirical Study of the Impact of Internet Store’s Service Quality on Consumer Attitude Toward Web Sites, Master Thesis National Chengchi University, Department of Business Administration (2000) 19. Zarem, J.E.: Experience marketing. Folio: The Magazine for Magazine Management 1, 28–32 (2000)

Guidelines of Quality for Industrial Online Tourism Miguel Cipolla-Ficarra1 and Francisco Cipolla-Ficarra1,2 1 2

Ainci –Asociación Internacional de Comunicación Interactiva Alaipo – Asociación Latina de Interacción Persona – Ordenador Via Pascoli, S. 15, C.P. 7 - 24121 Bergamo, Italy [email protected], [email protected]

Abstract. We present guidelines to foster quality industrial online tourism. This is the first guideline ever made where the component of presentation of the contents prevails over the other categories of design in a hypermedia system, such as navigation or structure. As well as this, a heuristic table has been determined to show whether the website correctly represents the online content for tourist promotion. The main features that a heuristic assessor in the industrial multimedia/hypermedia systems has to consider are those locations where there is a crossing point among tourism, communicability, design industry and trade. Keywords: Industrial, Tourism, Design, Layout, Evaluation, Guidelines, Museum.

1 Introduction In the last few years many traditional industries have ceased their activities for different reasons or have settled down in new great emerging economical areas, such as the Asian countries. The machines, tools, office furniture, etc. have deteriorated with the passing of time. Others, however, have been restored and they occupy important places in museums, schools, labs, etc. These are objects that can be visited in situ or virtually, through the Internet. At the beginning of the Internet age, many people predicted the end of the visits to the art museums. However, it has been a wrong hypothesis, because many museums have doubled and even tripled the number of visitors every year. In these years the great art museums made important investments in CD-ROM supports [1], [2], [3], and later on in online websites, etc. Exhibiting art objects online is an activity that may bring great financial benefits for the museum managers, the rural or city areas where these exhibits are to be found, the adept at the updating and maintenance of the web site, etc.

2 Interdisciplinary Sciences in Communicative Quality Era In the literature concerning interface design, it can be observed that there is an interaction between software and hardware, which is studied within human-computer interaction (HCI). The notion of usability is one of the chief factors used to carry out quality assessment of a computer system. In 1984 Eason produced a study concerning J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 12 – 21, 2007. © Springer-Verlag Berlin Heidelberg 2007

Guidelines of Quality for Industrial Online Tourism

13

usability in which he suggested three main elements used to carry out the assessment of a system: the user, the system, and the tasks [4]. However, in his research he did not make any clear distinction between usability and usefulness of a system. At the beginning of the 90s, Nielsen determined this difference: that is, the elements for the functioning pf the system “usefulness” and the usage of the system usability [5]. Since the year 2000 we have entered the communicative quality era [6]. In a heuristic assessment method such as the one described in this work, a model is necessary to analyse efficiently each one of the components of the system examined. A model which is also useful to establish a code or common language among all those who partake in the assessment process, with the purpose of wiping out ambiguities which entail a loss of time and of economical resources. The quality of a hypermedia system depends basically on internal factors, such as the different types of access to database or contents organization; and external ones, such as the use of certain colours in the user-computer interface or the easiness of a system usage. However, in the particular case of the heuristic assessment of multimedia/hypermedia systems it is necessary to add the subfactor of the temporal context; more than a decade ago, the response time between user and computer in searching a record in the database was over a minute, without this being considered a usability problem of the system. Nowadays, with the speed reached by existing processors, there are persons who see it as a failure in the system, since the interaction between human and computer tends to drop to fractions of thousandths of a second, hence the need to take into consideration this subfactor known as “temporal context”. Orientation is one of the most frequent problems when it refers to usability and communicability in hypermedia systems. As for Nielsen says, in order to improve orientation in those systems destined to the general public, designers generally turn to indexes, guided tours, backtrack keys, maps and hot lists [5]. A heuristic assessment table for multimedia/hypermedia systems is suggested here, obtained through the analysis of online industrial hypermedia. The table is an important tool in the suggested method because it enables the detection of the first usability and communicability failures in multimedia systems while examining the main components. Moreover, it is possible to have a global view of the system, that is, whether the dynamic or the static means prevail in the interfaces, the navigation is unidirectional or bidirectional in guided tours, the access to the database is done through search index, etc. [6]. The table is used by the heuristic assessor when he is interacting with the multimedia system. At that moment, he has to determine in the first place which are the links, the indexes, the guided tours, etc., of the hypermedia system. The more the assessor navigates through the structure of the system, the more he can compile or fill in the table; the action consists of placing a > sign where the considered component exists. The compilation of the table makes it possible to know in a simple, reliable and quick way; the approximate dimensions of the system, what components are present, the first mistakes, the classification which the multimedia system belongs to, and who is the target user. In order to guarantee the quality of the software, apart from creating the measurement techniques, it is necessary to check observation and measurement mistakes (systemic mistakes in the execution of the measurement and random errors). The modern systems, being more complex, are more likely to fail where the design requires a special attention, because of the costs they can reach before they are fixed.

14

M. Cipolla-Ficarra and F. Cipolla-Ficarra

Therefore, it is interesting to discern the different kinds of existing mistakes according to the definition of Kit at the moment of the assessment; mistake, a human action that causes an incorrect result; fault, an imperfect stage, process or data definition in the computer programme, for instance; failure, an incorrect result. The result (manifestation) of a fault. Error, the meaning by which the result is incorrect [7].

3 Towards Industrial Website The main obstacle to be overcome when designing a website related to industrial tourism is to determine whether all the pieces of the exhibit will go online or whether it will be a selection thereof. If the second option is chosen, the new dilemma is to determine which pieces of the collection must prevail over the rest. Some choose to transfer the guideline followed in the paper leaflet to the computer screen. In some cases, this can be an adequate solution when the number of representative objects in the exhibit are between 10 and 20. However, it is not an adequate solution, especially if we bear in mind the dynamic means that the hypermedia systems include, and which boost the quality criteria of richness, motivation, etc. These criteria make up the MEHEM methodology [6], and they have been applied successfully in distance teaching courses, promotion of cultural heritage and implementation of new measurements of quality, etc. Later on, there is the decision of how to organize the most representative content. A solution is to use the guided tour links, referential links and collections links. In the first case, usually designers put the most relevant objects in the first place, and continue in decreasing order. In the second solution, the user has certain freedom to go deeper in the fruition. The last alternative entails a prior and thorough assessment of the contents and the way that the access to that collection is structured. This collection can be either partial or total, but its assessment must always be carried out in a comprehensive way. The guidelines that we suggest may help to carry out that task –see Appendix 1. Once the quantity aspect of the objects to be shown online is determined, it is important to consider the context in which the industrial objects are related. That is, if they are objects related to the textile business, electronics, mechanics, cinematography, the rural world, etc. This context can work as a metaphor of the interface. A correct interface design can strengthen the attention and the motivation in fruition. Moreover, the notion of isotopy between the context and the objects that make up the exhibit online must always be present, especially in the design of the interface. 180 degree panoramic photography should be used in the shelves where the objects are exhibited, with the possibility of making a zoom-in or zoom-out and analysing the piece in detail. In those objects of a great size, such as a cine camera, the engine of an aeroplane or a tractor from the early 20th century, it is advisable to use the 360º technique. In the case of using footage for these kind of objects, the user values positively the 360º rotation plus the zoom-in for certain details of it. The surprise factor or discovery factor of the object can be achieved through a detailed shot, in the upper or lower part of the object, at the start of the filming and move slowly to an overall shot. Later on, this can be followed with several detailed shots. Sometimes, inserting a

Guidelines of Quality for Industrial Online Tourism

15

movement of the camera in front of mechanical objects, mimicking the rotation of a gear or the up and down movement of some engine pistons can be an entertaining means to end the observation of objects in the museums.

4 Isotopies and Design in the Industrial Website’s For a better understanding of the isotopy concept it is necessary to describe briefly the area in which its study is considered, that is semiotics. Semiotics or semiology is the science that studies signs in general, whether they are coded or not, systematic or not, eventual or stable, natural or cultural [8]. Both terms, semiotics –sustained by Peirce, and semiology –defined by Saussure, are currently regarded as synonymous, although the use of the one and the other is variable, according to the cultural environment and the moment in history. The main difference between Saussure and Peirce lies in the fact that Peirce focused on the character and logical function of the sign (pragmatism), whereas Saussure focused on the psychological or social aspect (structuralism), but since both aspects are present in sign, semiology and semantics belong to a single discipline. Inside semiotics, the notion of isotopy describes the coherence and homogeneity of the sign at semantic level. The isotopies act as common denominators among the components of the multimedia system, aside from the location inside the structure of the system. That is, the isotopies make it possible to establish a link of equality, giving a coherence to the context from the presentation of the different graphical resources. With the notion of isotopy applied to the design it is intended to detect quickly the existence of equality links among aspects that at first may look trite, but which are not when carrying out the design since they can affect the communicability of the system. These links can be located in the correlation that must exist between the images and the text, or between the main and the supporting theme in systems whose contents are in the shape of a story. Table 1. Some classical elements that can help the isotopy in multimedia systems - Colour - Graphic - Picture - Drawing - Video - Music

- Screen Attention Element - Animation - Metamorphosis - Texture - Illumination - Object Presentation

- Transparency - Special Effects (FX) - Transition - Typography - Sound - Topology

As Brenda Laurel claims, there is a stage in multimedia systems and each of the components must be inserted in the right way [9]. The relationship between stage and objects can be twofold; real and creative/artistic. In the first there is a transfer of the elements in the rooms that make up the museum or the exhibit in the computer screen. In the second, the designer makes the stage or the interface according to the communication criteria he has adopted. In this case, the abstraction degree for the user can have several levels; low, intermediate and high. In the first stages, with photographed industrial objects, communicability is guaranteed. However, if there is a high

16

M. Cipolla-Ficarra and F. Cipolla-Ficarra

abstraction degree in the stage created by the designer, the objects may lose that communication quality because of the abstract context that introduces noise in the communicative and interactive process. In the case of industrial products it would be interesting to work with smooth backgrounds and clear colours. Otherwise, choosing a real stage, you can transfer the colour of the rooms of the exhibit to the interface. In both cases, the quality in the fruition would be guaranteed.

5 Dynamic and Static Media In the case of industrial objects the image must prevail over the text. There are two groups of images in multimedia systems: static and dynamic. In the first group, the photos, graphics, drawings, etc. are included. In the second group there are 2D, 3D animations, the combination of 2D and 3D, videos, etc. 5.1 Photography Photography is a key element in these websites, because it can convey 100% of the considered object. Besides, with the several self-edition techniques you can upgrade or age these objects. In relation to this, we advise not to change the state of the object, because it can entail a state of disappointment in the moment in which the user decides to make a visit in situ, that is, to the museum, exhibit, etc. (these kind of visits are generally made by two or more people). 5.2 Graphics The graphics usually have a didactic function, and they also work, like sketches, to give some additional information, along with narrative texts. In contrast to photography, the making of graphics and outlines is (higher) more expensive than the making of a photograph. Among dynamic means, video is the cheapest and quickest means to present the objects. In the case of 2D animation online, today there is the possibility of making it through commercial programs such as Flash and with lesser costs if we use 3D. The ideal solution in the context of showing mechanical objects is the combination of 2D and 3D animations. The results are excellent and the user prefers this kind of solution. Obviously, the costs are higher than video and with 2D made through Flash, for example. 5.3 Metamorphosis and Transitions Metamorphosis is one of the animation devices that has a high didactic value for objects. The didactic aspect of metamorphosis is due to the possibility of explaining all the stages of an evolution process simultaneously through a starting image and a goal image. These two images can be different from each other. Another attribute of metamorphosis is that it draws attention to the objects that are depicted in the screen. The objects can be disparate, such as the process that is described in what follows. This process has, as its goal, the drawing of the attention to a system of catalogues of products, which in principle do not belong to the same category. The metamorphosis is similar in some aspects to the transitions of multimedia systems since it implies an

Guidelines of Quality for Industrial Online Tourism

17

animation. However, between the origin and the goal of a metamorphosis there is a sequence of frames that is wider than in the transitions. The transitions serve to establish a coherence in the presentation of the objects in a card format, that is, image plus explanation text. The use of more than a kind of transition and the speed of the transition from a screen to another in a small guided tour link (a guided link is known as small when it has no more than five nodes) has an effect on the perception of the whole by the user. The user can regard every frame as something independent from the rest. In this case a loss of the continuity among the different frames that are linked to each other takes place. There are two kinds of transitions; irregular –in the shape of a puzzle, painting brush, etc. – and regular – in the shape of squares, lines or dots. The regular dot-shaped figures are the most suitable for promoting tourism in the industrial sector. 5.4 Illumination Effects The effects derived from illumination, such as shadows, reflections, diffusion, etc. must also be maintained constant in the whole system. The illumination effects stem from the kind of illumination that can be used in scenes with 3D objects. This kind of light can be environmental or directional. In the environmental illumination the objects that receive the light generate shadows, whereas in directional illumination a beam of light is cast over a given object in the scene. There is a tendency to use global illumination over static objects, whereas to underline any technical feature of the piece that is being exhibited, the tendency is to resort to directional illumination. When there is more than a beam of light, illumination is multidirectional. The effects of the type reflection or shine mimic the arrival of the light to a spot on an object. This light can be depicted in the shape of a star. The shine can be either static or dynamic. It is dynamic when there is an animation associated with it. The animation can consist in a 360º turn over its axis or in an increase or decrease of the shine. This effect is used mainly when in the background of the screen the black colour prevails, the scale of greys and blues and all the metallic colours are like gold, silver, copper, etc. The shining indicates prestige or superiority of an element in relation to the others that make up the interface. They can also be useful to draw or attract attention to an element of the screen. They are effects that must not be too used in those collections with over 20 objects on display. 5.5 Relief Effect This is about mimicking the shadows cast by ambiance illumination among the elements in the screen. The purpose is that these bi-dimensional elements "acquire" a threedimensional quality, without using 3D programmes. The emulation is obtained by correctly darkening the contours of those elements which are intended to be highlighted in the interface. Among these elements can be mentioned the navigation keys, the typography, the frames of the images, the photos, etc. It is a positive effect on the navigation keys since they are easier to recognize in an interface or in underscoring a static image such as photos. The interface of an industrial multimedia system has a higher degree of realism with the emulation of the shadows, which favours its acceptance.

18

M. Cipolla-Ficarra and F. Cipolla-Ficarra

5.6 Colour Through the interface’s colour a bigger or lesser coherence can be reached. The element of the interface where the same colour prevails the most is the background of the interface, especially if the background is a colour and not a texture. The colour of the background or pillar of the screen can be grouped basically in primary colours and secondary colours. Because of reasons related to online speed the white background is currently favoured in many European industrial websites, but in the North American countries there is a tendency to use dark colours, such as brown or black. In Latin America the blue and the light blue colours are preferred. However, the use of the colours must be previously analyzed at the moment of the design of the interface since they have different connotations. These connotations are related to the meaning and depend on the culture and geographical location of the users. 5.7 Basic Textures In some multimedia systems textures and images are used as backgrounds or bases of the screens. These textures can be related to the copyright of the images or, for example the logo of the museum. Nevertheless, it is necessary to maintain the same texture or image in the whole application, since constant change inside the same part of the structure can trigger confusion in the user when it comes to navigation. 5.8 Drawing The drawing, just as the typography and other elements that make up the interface, can have a greater or lesser degree of abstraction according to the purpose of the content of the multimedia system. Just as in the typography, if the goal of the drawings is to educate and inform, the drawing must be figurative. In contrast, if the purpose is a pastime, the drawing can be abstract. However, an increase in the use of figurative drawings, that is, those with little abstraction, is currently to be seen the multimedia applications. In a multimedia system with industrial content aimed at education, E-commerce, etc., it is advisable that it contains only figurative drawings. 5.9 Three-Dimensional Objects The tridimensionality is due to the presence of the coordinate “z”. There are multimedia systems that have the whole contents of the database in 3D. However, this content can be presented or viewed in different ways. The use or non-use of the coordinate “z” in the presentation of the content of the system is related to the higher or lesser realism of the image, the criteria of quality, richness and the notion of prospect. The maximum degree of realism in this coordinate is achieved in virtual environments.

6 Towards a New Professional The ideal professional for the websites aimed at the tourist promotion of industry must have a background in the formal sciences and the factual sciences. He /she should ideally have an education in engineering and social communication, for example. It is a

Guidelines of Quality for Industrial Online Tourism

19

mistake to leave the technical aspect of the design outside the project, through outsourcing, because this may lead to unexpected results. In many cases the timing of design, implementation and setting in motion last longer than was foreseen, which has an impact on cost [10]. With these kind of websites, it is interesting to develop and foster a new kind of professional, since in this way the use of usability labs can be avoided, the deadlines of the project can be shortened and, consequently, much may be gained in the time-cost equation. The factors cost and time have an essential importance when it comes to carrying out a test or assessment of the communicability of a hypermedia system. As a rule, when an assessment of the prices is made, it is established according to the ratio person/hour. In contrast to other methods or heuristic assessment techniques in which the factors time, equipment, installations, staff, etc. boost the cost of an analysis of the components of the multimedia system, with the method that is suggested, the assessment costs are noticeably reduced, since it does not require the use of a lab, nor a high number of staff. Starting from the profile of a professional presented by Francisco Cipolla-Ficarra for the heuristic assessments, it is possible to head it towards the new requirements in view of the industrial sector [11]. Table 2. Knowledge for heuristic evaluation and design in industrial multimedia systems Design Architecture Museum Industry Physical Marketing

Electronic Chemical Computer Graphic Computer Sciences Social Sciences Mass Media

Information Management Machines and Tools Telecommunications Photography and Video Composed Material Technologies of Manufacture

7 Conclusions In the current work initial guidelines have been presented in order to make websites for the promotion of industrial tourism. The graphical aspect of the interface, applying the concepts of communication and engineering , can facilitate the process of tourist promotion. Excellent results are achieved through an interdisciplinary team, supported by a new figure in the heuristic assessment of industrial hypermedia systems. Obviously, it is not advisable to resort to outsourcing to try to solve the technical questions of the industrial environment. It is better to set up a team where the different professionals of the formal and factual sciences interact. This new professional requires specialists with knowledge and experience in the context of hypermedia systems, where the intersection of the formal sciences and the factual sciences is located. Teamwork enables the results obtained in previous assessments with other experts of the assessment sector to be verified. Moreover, the assessment can be carried out by the various departments which take part in the creation of the multimedia system such as the design, programming, implementation, etc. communicating the failures that are revealed in an immediate way. The quality of the software and the design of the interfaces can power the fruition of the industrial exhibits, aside from the geographical location where they are located.

20

M. Cipolla-Ficarra and F. Cipolla-Ficarra

References 1. 2. 3. 4. 5. 6.

7. 8. 9. 10. 11.

Musée du Louvre DVD-ROM: Montparnasse Multimédia, Paris (2004) Italia Design CD-ROM: Editel, Milano (1995) Musée d’Orsay CD-ROM: BMG Interactive, Paris (1996) Eason, K.: Handbook of Human-Computer Interaction. North-Holland, Amsterdam (1997) Nielsen, J.: Multimedia and Hypertext. Academic Press, San Diego (1990) Cipolla-Ficarra, F.: Homepage and Communications: Quality Metrics. In: The Eight International Conference on Distributed Multimedia Systems, San Francisco, pp. 202–212 (2002) Kit, E.: Software Testing in the Real World: Improving the Process. Addison Wesley, New York (1995) Cobley, P., Jansz, L.: Introducing Semiotics. Icon Books, Cambridge (1999) Laurel, B.: The Art of Human-Computer Interface Design. Addison-Wesley, Massachusetts (1990) Spohrer, J., et al.: Steps Toward a Science of Service Systems. IEEE Computer 40(1), 71–77 (2007) Cipolla-Ficarra, F.: Evaluation of Multimedia Components. In: International Conference on Multimedia Computing and Systems. IEEE Computer Society, Ottawa, 557–564 (1997)

Guidelines of Quality for Industrial Online Tourism

21

Appendix 1: Guidelines Components –alphabetical order − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − −

Adaptability of menus. Appropriate international use of colour, typography, texture, effects of 2D/3D, etc. Available browsing methods. Balance between white space versus navigation. Clarity and brevity in labels, titles, legend buttons, etc. Coherence of text and illustration. Complexity or facility for the access to the catalogue and prices: passwords, biometrics systems, etc. Customer satisfaction: instruments for evaluation. Design innovation in active and passive media’s. Durability of the technical information: one year or more. E-mail addresses to ask for information. Errors treatment. Facility of querying and fast string search. Feed-back levels. Homogeneity between the main topic and secondary topics. Index of the contents of the site in each page. Inference about visitors: qualified or inexpert. Level content: professional, technical, general, etc. Localization and internationalisation context. Maps. Metaphor familiar and layout attractive. Much or little details of the products and/or services. Multilingual documents. News. Number of visitors. Personalization: types of customer/user information. Portal content: high tech, medium tech or low tech. Portal structural: Modern –i.e., Flash or Classical –i.e., Html. Possibility to receive online or paper material – Virtual leaflet. Precision of the referential links. Quality and univocal meaning for photo, graphic, illustration, animation, video, etc. Quality of details: high, medium, low, very low. Security of payment online. Segmentation of content: by industrial and job function, by surfing mode, by interest, etc. Site available in different languages. Synchronisation between active and passive media’s. Terminology: denotative or connotative. Transparent dialog box. Typology of access structures. Users' comments. Virtual tours: webcam, panoramic pictures 180 – 360°, animations 2D/3D, VR, etc. Visibility of the principals commands. Website: visibility online and web position. Web credibility.

Provide Context-Aware Advertisements with Interactivity Qin Gao1, Pei-Luen Patrick Rau1, and Gavriel Salvendy1,2 1 2

Department of Industrial Engineering, Tsinghua University, Beijing 100084, China Department of Industrial Engineering, Purdue University, Indiana, 47907-2023, US [email protected], [email protected], [email protected]

Abstract. Interactivity and context-awareness are two salient features of mobile advertising, but their actual influences on advertising effectiveness have not been studied. This study attempts to understand expectations and concerns of Chinese mobile users hold for context aware advertising, and to investigate the influences of interactivity and context-awareness. A web-based survey and an experiment were conducted for the two purposes respectively. Results suggested that Chinese users generally hold a modestly positive attitude towards context aware advertisements; advertisements adapt to physical environment, geographic context, and purchasing history are desired; young people are more open to context aware advertisements. Their major concerns were irrelevant messages and information flood. The interaction between perceived interactivity and context awareness has a big effect on advertising effectiveness. For high interactivity advertisements, contextual offers of the advertisements result in more responses and better advertisement attitude. But for low interactivity advertisements, contextual offers incur more negative attitude and lead to lower purchase intention. Keywords: mobile advertising, context-aware computing, interactivity, user experience.

1 Introduction The fusion of mobile technologies and daily communications is recognized as a new chance for advertising with high potential. Several trends have strengthened this belief. First, the penetration rate of mobile devices is huge. In China, the number of cell phone subscribers reached 383 million at the end of October in 2005, averagely 29 cell phones for 100 Chinese . Second, the migration from 2G to 3G is expected to be accompanied by mobile communications shifting the focus from voice to data services. Also the enhanced features of newly developed models allow rich content presentation on mobile devices. Mobile communications are characterized by high accessibility, personal/intimate nature, location-aware potential, and quick response capability of the device and the ubiquity of the wireless network. Practitioners have been advocating for mobile advertising for years, claiming that the context-aware potential and the easy interactivity J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 22 – 31, 2007. © Springer-Verlag Berlin Heidelberg 2007

Provide Context-Aware Advertisements with Interactivity

23

at anywhere of the device would make it possible to provide the information at the point of need and the possibility for the customer to respond immediately. Customers would be satisfied with the tailored information coming at the right time. Until now, however, the positive influences of interactive and context aware mobile advertisements are often believed to be rather than proved to be. Moreover, there are also concerns with privacy intrusions and spam possibility, especially for context aware advertisements. To understand how favorably customers will respond to experiences of receiving context aware mobile advertisements and what their major concerns is a challenge for mobile advertisers. This study, as an attempt to fill this void, is designed to understand expectations and concerns of Chinese mobile users hold for context aware advertising, and to investigate the influences of interactivity and context-awareness on the effectiveness of mobile advertisements.

2 Interactivity Interactivity became of research interest for decades, but variations still exist in the conceptualization of interactivity. McMillan and Hwang suggested that these definitions can be categorized into three groups based on the primary focus of the author [2]. Interactivity can be defined as properties of a certain medium, either general characteristics such as user control, reciprocal communication, and etc. [3, 4]; it can also be defined specific features such as chat room, registration form, and etc. [5]. The third school bases the concept of interactivity on process related variables [6, 7], arguing that interactivity should not be measured by analyzing the process or counting the features, but by observing how users perceive the interactivity during the communication [2, 8, 9] This perspective is more consistent with the human-centered principle, and is adopted by the current study. Perception of interactivity was found a more predictive factor for the effectiveness of internet advertisements [2, 9-11]. Mobile communications provide greater opportunities for interactive communications at anytime and anywhere, whereas the inferior device capacity, the slow network, and the awkward input restricts the bandwidth for information exchange in mobile communications, and consequently influences the form and quality of interactions. The possible influence of the interactivity of mobile advertisements on the advertising effectiveness remains unknown.

3 Context-Aware Advertising A system is context-aware if it uses context to provide relevant information and/or services to the user, where relevancy depends on the user’s task [12]. Mobile communications are known for the context-aware potential brought by the capacity of devices to identify the context within which the individual is operating 13], but in different studies the scale of the context varies. Many people refer context solely to location information, but Hritova and O’Hare argued that to describe the environmental factors which may affect the customer’s response, computing environment, physical context and history context should be included besides location information [13]. Dey and Abowdidentified three context-aware application behaviours: presentation of information and services to

24

Q. Gao, P.-L.P. Rau, and G. Salvendy

a user; automatic execution of a service for a user; and tagging of context to information to support later retrieval [14]. By utilizing context information in various ways, marketers attempt to provide content at the point of need, which may, for example, influence impulse purchases [15].

4 Research Questions This study was designed to address two questions: 1. What are the general expectations and privacy concerns of Chinese customers towards context aware advertising? 2. What are the actual influences of interactivity and context-awareness and their interaction on the effectiveness of mobile advertising?

5 Methodology 5.1 Survey A web-based survey was conducted to collect data of general considerations of Chinese users toward mobile advertisements, focusing on interactivity and context awareness. 264 mobile phone users, aged 17-45, were recruited in the survey held during November, 2005. The sample comprises a wide range of age, education, income, experiences of using cell phones, monthly fee for cell phones, and monthly fee for mobile data services, with an equal representation of genders, but the majority of respondents are characterized as young people with high level of education and middle level of income. Regarding experiences with mobile advertisements, most respondents (90.9%) reported that they have received short message advertisements, while only about 5% of respondents have experiences with multimedia message (MMS) advertisements, and less then 10% of respondents have experienced mobile Web advertisements or mobile game advertisements. Following a brief introduction to mobile advertisements, questions regarding three aspects were given: • How favorably/unfavorably will mobile users consider different context aware advertisements? • How do mobile users consider contextual offering of advertisements in the whole? • What are their fears and concerns when using such services? 5.2 Experiment Experiment Set-Up and Stimuli The experiment was conducted on a business street near to Tsinghua University, where the participant performed a walk along a predetermined path as shown in Figure 1. The place was chosen due to the high familiarity participants had with this street. All participants were from Tsinghua University and they all know the street well.

Provide Context-Aware Advertisements with Interactivity

25

To avoid possible operation problems with a new device, we ask our participants to use their own cell phones. They were asked to go to four shops as marked on the figure. In each shop, they looked for specific information to answer questions written on their task sheets. Questions were designed to exemplify typical tasks a customer would conduct when shopping around, for example, to check the price of a specific product, or to look for newly discounts. Participants sent the answer to the experiment moderator before they advanced to the next question as required, and their location was identified by the sent-back message. Short message advertisements were sent to participants at these locations by another experiment assistant. Participants in the context-aware group would receive advertisements related to their current location, as shown in Fig. 1. The other group received the same message but at locations irrelevant with the content of the message. Seven experimental short message advertisements were sent to each participant. They were developed into two versions: interactive or not interactive. The difference lied in whether there were feedback options and (/or) clickable links for the participant or not, as shown in Table 5. Participants in the interactivity group were able to reply the message, or call back for further interaction, or visit the provider’s website, or do all above. However, participants in low interactivity condition can only read the message. Experiment Design and Variables A 2*2 factorial experiment design was used to test the three hypotheses derived from the second research question. The first independent variable was the interactivity of the mobile advertisements, interactive or not-interactive. There are three key components which influence users’ perception of interactivity: two-way communication, user control, and speed. Two-way communication was manipulated as either there were options for the user to reply the message and (or) call back or not. User control

Start point: Tsinghua University

Ad 1: Xs-exes long sleeve knit

Lotus Department Store

JFCX Pastry Shop

Ad 4: Sorry Ericsson M600C Ad 5: Siemens SG75

Ad 6: Harry Potter and the Deathly Hallows, by J. K. Rowling Ad 7: Resume, by Jimmie

WDK Electronics End point: O2Sun Bookstore

Ad 2: Strawberry cheese cup cake Ad 3: Movie Doogal

Fig. 1. Message received by the context-aware group on the walking route

26

Q. Gao, P.-L.P. Rau, and G. Salvendy

was manipulated as either there were clickable links to the product-related website or not. Speed, the last component, was not manipulated, since it was not applicable for the group having no feedback options or clickable links. The second independent variable, context-awareness, was manipulated as follows: Participants in the context relevant group received advertisements of products which were either available in participants’ current context or related to the scenario, for example, an advertisement of the discounted product at the current store, or an advertisement of a newly coming movie at a pastry shop where movie magazines were also served. In the context irrelevant group, however, the same message were sent but at locations bearing no obvious relationships with the content of the message. To examine whether the manipulation of user options resulted in different levels of perceived interactivity, a with 15-item, seven-point Likert scale, derived from the instrument measuring perceived interactivity developed by Gao et al. (2006) , was used as a check of the manipulation. Procedure Each participant was tested individually. One participant was first asked to come to the IT Usability Lab at Tsinghua University to complete a general information questionnaire concerning his/her demographic information, experience with the internet and with mobile services, attitude towards mobile advertisements in general and involvement with four types of products which would be advertised in the experiment. An introduction of the purpose and tasks of the experiment was then given. The participant was asked to go to four stores near the lab (they were a bookstore, a department store, a digital product store, and a bakery, respectively; the furthest location was no more than 1.5 kilometer away from the lab), completing three tasks at each place, as listed on a task sheet. At each place, three short messages were sent to the participant. Participants were supposed to go to a department store, a digital product market, a bookstore, and a pastry shop to check for some information written on their notes. They could enjoy some refreshments at the pastry shop with the 10 Yuan (RMB) they got from the moderator before they started the journey. Participants were then given a print-out instruction of tasks to start their short trip. The order of tasks was the same for all participants. Upon completion of all tasks, the participant was instructed to return to the usability lab. A free-recall test and an advertisement recognition test were then given without prior informing the participant. Then the participant finished measures of perceived interactivity, attitude towards advertisements, brand attitude, as well as purchase intention. Finally 40 Yuan (RMB) was given as an incentive of participation.

6 Results and Discussion 6.1 Survey Results The survey of favorability of different contextual offerings of mobile advertisements provided the result as shown in Fig. 2. Among all, advertisements related to users’ physical environment were considered the most favorable, and the second is advertisements related to a store a user is just passing by (with permission). The least favorable type of contextual offerings was advertisements related to users’ visit history at a specific website.

Provide Context-Aware Advertisements with Interactivity

27

Fig. 2. Users’ favor regarding different types of context aware advertisements

As shown in Table 1, mobile users showed a modestly positive attitude towards context-aware advertisements in general. 48% considered such advertisements “overall liking”, 48% considered them “good”, 50% considered them “interesting”, and more encouragingly, 62% considered them “useful”. Irritation is a controversial issue, however, as reflected by the static that 32% found them “irritation”, 32% “not irritation”, and the rest 36% gave a neutral response. Table 1. Attitude towards context aware advertisements in general Items Overall disliking Bad Uninteresting Useless Irritating

No. % No. % No. % No. % No. %

1 10 0.04 6 0.02 11 0.04 10 0.04 20 0.08

2 11 0.04 6 0.02 6 0.02 10 0.04 17 0.06

3 24 0.09 21 0.08 34 0.13 21 0.08 47 0.18

4 89 0.34 102 0.39 79 0.30 61 0.23 96 0.36

5 78 0.30 80 0.30 83 0.31 85 0.32 52 0.20

6 38 0.14 38 0.14 38 0.14 65 0.25 19 0.07

7 14 0.05 11 0.04 13 0.05 12 0.05 13 0.05

Overall liking Good Interesting Useful Not irritating

According to respondents’ satisfaction ratings, they were classified into “positive respondents” (people whose satisfaction ratings were higher than 4) and “negative respondents” (people whose satisfaction ratings were lower than 4) for each type of context aware advertisements. Significant differences were found in age between positive and negative respondents for the first four types of context-aware advertisements, indicating that young people are more enthusiastic for advertisements related to their geographical, physical and temporary context. Table 2 presented major concerns Chinese mobile users hold for context-aware mobile advertisements. More than 80% respondents concerned with irrelevant junk message and information flood. 71.59% feared the normal usage of the device being interrupted. 60.23% people had privacy concerns on context aware mobile advertising. Being tracked all day long, however, is not as severe as the others, as shown by 43.56% agreement.

28

Q. Gao, P.-L.P. Rau, and G. Salvendy Table 2. Users’ concerns regarding context aware mobile advertising

Concerns regarding context aware advertising I might be tracked all day long. My privacy might be intruded. Too many advertisements might jump into my cell phones. Junk message. My normal usage of the device might be interrupted.

Responses 115 159 215

Percentage (%) 43.56 60.23 81.44

212 189

80.30 71.59

6.2 Experiment Results The main effects of interactivity and context awareness on advertising effectiveness were examined using ANOVA (for tests of advertisement memory, brand attitude, and purchase intention) and ANCOVA (for the test of advertisement attitude, with attitude towards mobile advertisements in general as the covariate). The results showed neither the level of interactivity nor context awareness had a significant influence on any measures of advertising effectiveness. However, an examination of the interaction between the interactivity and the context awareness revealed that there was a significant interactivity * context awareness interaction on advertisement attitude (F(1,36) = 9.52, p = 0.004), brand attitude (F(1,36) = 8.21, p = 0.007), and purchase intention (F(1,36) = 9.96, p = 0.003), but the interaction showed only marginal significance on advertisement memory (F(1,36) = 2.99, p = 0.09). Table 3. Data for Testing the Effect of Context Awareness for Low Interactivity Advertisements

Variables Advertisement memory (%) Advertisement attitude Brand attitude Purchase intention

Context irrelevant Mean STD

Context relevant Mean STD

%Diff

F(1,18)

pvalue

46.2

38.8

-16.1

1.85

0.19

5.07 5.27 5

11.7 0.93 0.88 0.88

4.36 4.32 3.99

12.7 0.88 0.83 0.80

-13.9 -18.0 -20.2

3.03 6.18 7.31

(1)

0.10 0.02 0.01

(1): attitude towards mobile advertisements in general were included as the covariates, the degree of freedom of the F test is (1,17), LSM (irrelevant)=5.06, LSM (relevant)=4.36. Then the effect of context awareness on advertising effectiveness was tested by fixing the level of interactivity. As shown in Table 3, for low interactivity advertisements, significant differences were found with brand attitude (F(1,18) = 6.18, p = 0.02) and purchase intention (F(1,18) = 7.31, p = 0.01). Contextual offerings engendered worse brand attitude and lower purchase intention. Despite of this, for high interactivity advertisements, contextual offerings advertisement to significantly more responses (Z=2.74, p=0.006 Meanirrelevant=0.1, Meanrelevant=1.5) and better advertisement attitude (F=6.08 p=0.02 Meanirrelevant=4.25, Meanrelevant= 5.41), as shown in Table 4.

Provide Context-Aware Advertisements with Interactivity

29

Table 4. Data for Testing the Effect of Context Awareness for High Interactivity Advertise ments

Variables Response to ad Advertisemen t memory (%) Advertisemen t attitude Brand attitude Purchase intention

Context irrelevant Mean STD

Context relevant Mean

STD

%Diff

F(1,18)

pvalue

1400

Z= 2.74(1)

0.006

1.15

0.30

0.1

0.32

1.5

1.27

42.5

12.5

47.9

9.1

4.25

1.19

5.41

0.86

23.6

6.08(2)

0.02

4.50

1.18

5.21

0.73

15.8

2.61

0.12

4.08

0.99

4.78

0.74

17.2

3.17

0.09

12.6

(1): Wilcoxon two-sample test; (2): attitude towards mobile advertisements in general were included as the covariates, the degree of freedom of the F test is (1,17), LSM (irrelevant)= 4.32, LSM (relevant)=5.34. With interactive advertisements, a contextual offer do increase user response effectively and result in more affectional attitude. This is consistent with the notion that the point of purchase is an ideal time to communicate with customers [16]. The study of contextual perceived usefulness also suggested that context-based marketing communication is key to the success of mobile commerce [17]. Especially, the response to advertisements sent in a context aware manner was much better than that sent in a context irrelevant manner, which indicates that the relevance between the advertisements and the surrounding context increases users’ involvement with the message, and the convenience to check the advertisement information or provision may encourage users to respond. However, in the case of advertisements not providing response options, users’ attitude towards the brand and the consequent purchase intention were impaired by context-relevant advertisements rather than improved. A possible reason might be that the message advertisements in this experiment were sent without participants’ solicitation. The low interactivity advertisements only “broadcast” information of products and services. The feeling of being intruded by such advertisements might become even stronger when users detect that their cell phones are spammed just because they pass by a store accidentally. As cautioned by those pioneer practitioners such as Vindigo, SkyGo, and DoubleClick, location based services and advertisements should be provided with great care so as not to impair users’ privacy since handsets are very personal devices. In a qualitative study of user experiences with location sensitive mobile applications, context aware notifications also gained numerous doubts and criticism due to possible spamming with unnecessary advertisements and information flood [18]. Mobile users don’t like the idea that “I should be spammed just because I passed the department store accidentally”. Obtaining users’ permission and using proper filtering are important in this case.

30

Q. Gao, P.-L.P. Rau, and G. Salvendy

7 Conclusions Context awareness and interactivity play an important role in mobile advertising industry. This study was an early attempt to understand the expectation and concerns of Chinese mobile users towards context aware advertising, and to study how interactivity and context awareness influence advertising effectiveness in mobile advertising. The results indicated that mobile advertisements adapt to physical context, geographical context, and purchasing history are favorable, but Chinese users only show a neutral attitude towards advertisements based on web visit history. Irrelevant messages and information flood are major concerns of Chinese mobile users holding for context aware advertising. Positive influences of context awareness only take place with interactive advertisements. Therefore, when the goal of a mobile campaign is to generate responses, the context in which the response options are given to the users are of importance and must be taken into considerations of the mobile marketer. On the contrary, with “broadcasting” advertisements, users’ attitude towards the brand and the consequent purchase intention were impaired by context-relevant advertisements rather than improved. Permission from the audience should be taken as a necessary condition if marketers are going to provide such advertisements to handset users near vicinity of a certain place. Due to the experimental nature of the study, findings reported here may be limited in several aspects: • The experiments only last for about 2 hours for each participant, and we cannot say what reactions mobile users would give over a longer time period; • The influences of moderators (user involvement with products and attitude towards mobile advertisements in general) were excluded from the result by treating these variables as covariates in this study. More effective approaches to control such variables could be used in future studies. Further study can explore further the antecedents and consequences of perceived interactivity in mobile advertising. The cognitive processing of context-aware advertisements is also interesting. Furthermore, from the experiment result it seems that context-awareness of the advertisement has an influence on perceived interactivity, but to prove the influence, empirical studies are needed.

References 1. Chinese ministry of information technology: Annual revenue report of Wireless communication industry (2006) 2. McMillan, S.J., Hwang, J.-.: Measures of perceived Interactivity: An Exploration of Communication, User Control, and Time in Shaping Perceptions of Interactivity. Journal of Advertising 31(3), 41–54 (2002) 3. Coyle, J.R., Thorson, E.: The Effects of Progressive levels of Interactivity and Vivideness in Web Marketing Sites. Journal of Advertising 30(3), 65–78 (2001) 4. Steuer, J.: Defining Virtual Reality: Dimensions Determining Telepresence. Journal of Communication 42(4), 73–93 (1992)

Provide Context-Aware Advertisements with Interactivity

31

5. McMillan, S.J.: Interactivity is in the eye of the beholder: Function, perception, involvement, and attitude toward the web site, pp. 71-78 (2000) 6. Heeter, C.: Implications of new interactive technologies for conceptualizing communication, in Media Use in the Information Age. In: Salvaggio, J.L., Bryant, J. (eds.) Emerging Patterns of Adoption and Computer Use, Lawrence Erlbaum Associates, Hillsdale, NJ (1989) 7. Rafaeli, S., Sudweeks, F.: Networked Interactivity. Journal of Computer Mediated Communication [online], 2 No.4 (1997) 8. Lee, T.M.: The impact of perceptions of interactivity on customer trust and transaction intentions in mobile commerce. Journal of Electronic Commerce Research 6, 165–180 (2005) 9. Wu, G.: Perceived Interactivity and Attitude toward Web Sites. In: Proceedings of the American Academy of Advertising, pp. 254–262 (1999) 10. Cho, C.H., Leckenby, J.D.: Interactivity as a measure of advertising effectiveness: Antecedents and Consequences of Interactivity in Web Advertising (1999) 11. Yoo, C.Y., Stout, P.A.: Factors Affecting User’s Interactivity with the Web site and the Consequences of User’s Interactivity. In: Conference of the American Academy of Advertising (2001). 12. Kaasinen, E.: User needs for location-aware mobile services. Personal and Ubiquitous Computing 7, 70–79 (2003) 13. Hristova, N., O’Hare, G.A.: Ad-me: Wireless Advertising Adapted to the User Location, Device and Emotions. In: Proceedings of the Proceedings of the 37th Annual Hawaii International Conference on System Sciences (HICSS’04) vol. 9(9), IEEE Computer Society (2004) 14. Dey, A. and Abowd, G.: Towards a Better Understanding of Context and ContextAwareness. In: Proceedings of the 1st international symposium on Handheld and Ubiquitous Computing, Karlsruhe, Germany (1999) 15. Kannan, P., Chang, A., Whinston, A.: Wireless Commerce: Marketing Issues and Possibilities. In: Proceedings of the 34th Annual Hawaii International Conference on System Sciences ( HICSS-34) 9 No. 9, IEEE Computer Society (2001) 16. Shimp, T.A.: Advertising Promotion. Dryden Press, Orlando (2000) 17. ThaeMin, L., JongKun, J.: Contextual Perceived Usefulness? Toward an Understanding of Mobile Commerce Acceptance. In: Proceedings of the International Conference on Mobile Business (ICMB’05), IEEE Computer Society (2005) 18. Haekkilae, J., kkil, Maentyjaervi, J.: User experiences on combining location sensitive mobile phone applications and multimedia messaging. In: Proceedings of the 3rd international conference on Mobile and ubiquitous multimedia. ACM Press, College Park, Maryland (2004)

Skills Matter: A Tale of the Anxious Online Shopper Yi Maggie Guo and Barbara D. Klein Department of Management Studies, School of Management, University of Michigan – Dearborn, 19000 Hubbard Dr. Fairlane Center South, Dearborn, Michigan, USA 48126 {magyiguo, bdKlein}@umd.umich.edu

Abstract. In this study we investigate the effects of user skill and task challenge on online shoppers’ experiences. We use a model suggested by flow theory in which shoppers are grouped into four categories (flow, anxiety, boredom, and apathy) based on their perceptions of task challenge and their skill in performing an online shopping task. Results show that anxious shoppers have lower perceptions of the dimensions of flow, believe online shopping websites are less useful, and believe they are less likely to use the system in the future compared to the other three groups of online shoppers. Keywords: Flow, Internet, online shopping.

1 Introduction Designers of online shopping systems hope to create systems that online shoppers will find engaging and that will lead to the purchase of items. However, compared to brick-and-mortar stores, online shopping systems present shoppers with new challenges and demand different types of skills. If there is a mismatch between the challenges posed by an online shopping system and the skills of an online shopper, the purchasing behavior of the shopper may be affected. In the extreme case, the shopper may simply stop using the online shopping system. This paper aims to increase our understanding of the effect of these mismatches between task challenges and user skills in the context of online shopping systems. The theory of flow is used to examine the effect of these challenge-skill mismatches on user perceptions and beliefs about the online shopping experience.

2 Theoretical Framework The theory of flow provides a theoretical lens for understanding human behavior in a variety of task contexts. Csikszentmihalyi [1] proposed that humans can enter into a state of “flow” in which they are in a state of intense concentration and experience a shift in their perception of control over the activity. They also feel a mergence of their conscious awareness and the activity in which they are engaged. Csikszentmihalyi [1] also argued that people in a state of flow perceive time differently than normal, with J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 32 – 41, 2007. © Springer-Verlag Berlin Heidelberg 2007

Skills Matter: A Tale of the Anxious Online Shopper

33

time generally seeming to fly by while the person is engaged in the activity. They also experience a “loss” of self and forget their everyday concerns temporarily [1]. Csikszentmihalyi [1] argues that human performance in a wide variety of tasks is enhanced when a person enters into a state of flow. People experience flow while engaged in a variety of activities. Since the theory was developed, it has been used in a variety of fields including chess, rock climbing, and surgery [1]. Since the late 1980s, information technology researchers have used the theory of flow to explain the usage of spreadsheet software [2], email [3], and the Internet [4]. The theory of flow suggests that in order for flow to occur, the task should have a clear goal and a quick, unambiguous feedback mechanism [1]. These two factors are apparent in tasks such as chess and basketball. A third precursor to flow, which we will focus on in this paper, is said to be a perceived balance of challenge and skill [1]. If challenges exceed skills, people feel overwhelmed and anxious. On the other hand, if the activity is too easy, people get bored. Results of prior studies suggest that both challenge and skill must exceed a threshold in order for flow to occur, otherwise the person experiences apathy toward the task [1], [5]. A four-channel model of flow was developed to highlight the fit between perceptions of task challenge and the level of skill brought to the task (see Figure 1) [5].

High

Challenge

Anxiety

Flow

Apathy

Boredom

Low Low

Skill

High

Fig. 1. 4-Channel Model of Flow

Csikszentmihalyi [6] also introduced a set of dimensions of flow which can be used to assess whether a person is in flow while performing a task. These dimensions include focused concentration (C) [1], merging of activity and awareness (M) [7], perceived control (CON) [8], transformation of time (TT) [7], and transcendence of self (TS) [7]. A sixth factor recognizes that tasks performed while in a state of flow are intrinsically rewarding or autotelic (AE) [7]. Prior studies have demonstrated that both emotional and cognitive outcomes are associated with flow states. Since flow is an engrossing, highly enjoyable experience, it is only natural for people to develop a positive emotion when in flow [9], [10], [11]. Pleasure is an aspect of emotional response that has been found to be associated with flow in prior studies [9]. In addition to pleasure, perceived usefulness of a technology and intention to use a technology are two cognitive outcomes associated with flow in prior studies. Cognitive absorption, a concept similar to flow, was found to positively

34

Y.M. Guo and B.D. Klein

influence the perceived usefulness of web technology [4]. Shopping enjoyment, as well as perceived usefulness, has a significant relationship with the intention to return to the website in the future [12].

3 Research Questions and Hypotheses The central research question examined in this study is whether there are differences in the perceptions and beliefs of online shoppers in states of flow, anxiety, boredom, and apathy. The presence of differences in these four groups would suggest that the online shopping experience can be improved through changes to system design aimed at changing the level of task challenge or through user training aimed at improving skills. Hypotheses 1-6 examine the differences in the six dimensions of flow (focused concentration, merging of activity and awareness, perceived control, transformation of time, transcendence of self, and autotelic experience) among the four groups of online shoppers. A sample hypothesis is shown below. The hypotheses for the other dimensions follow the same form. Hypothesis 1: There will be differences in focused concentration (C) among online shoppers in states of flow, anxiety, boredom, and apathy. Hypotheses 7-9 examine the differences in the three outcomes of flow (pleasure, perceived usefulness, and behavioral intention) among the four groups of online shoppers. A sample hypothesis is shown below. The hypotheses for the other outcomes of flow follow the same form. Hypothesis 7: There will be differences in pleasure (P) among online shoppers in states of flow, anxiety, boredom, and apathy.

4 Methodology A study was conducted to collect data on key flow and outcome measures. Subjects with various majors were recruited from undergraduate courses in a major American university. Upon agreeing to participate in the study, subjects were randomly assigned to one of eight pre-selected commercial websites. A package of documents and instructions was given to subjects when they arrived. Subjects were directed to go to the website assigned to them and find something they were interested in purchasing. A seven-point Likert scale questionnaire measuring the dimensions of flow, outcomes of flow, perceived challenge of the task, and perceived skills appeared six to eight minutes after subjects started shopping. Subjects were instructed to complete the questionnaire at once when it appeared on the screen. Once they completed the questionnaire, subjects were asked to continue shopping. All of the experimental manipulations and measures were conducted online, and all measures were captured by computers. Although goal clarity and feedback mechanism are conceptualized as precursors to flow in the theory, these two factors were held constant in the experimental design.

Skills Matter: A Tale of the Anxious Online Shopper

35

Flow State Scale [13], a validated instrument based on the theory of flow, was used to measure focused concentration (C), merging of activity and awareness (M), control (CON), transformation of time (TT), transcendence of self (TS), and autotelic experience (AE). Additional measures based on previously validated instruments were used to measure pleasure (P) [14], perceived usefulness (PU) [12], and behavioral intentions (BI) [12], [15], [16]. Perceived challenge (PC) and perceived skill (PS) were each measured with five items adapted from prior research [11].

5 Results A total of 354 subjects, including 211 female and 143 male subjects, participated in the study. The average age of the subjects was 21.2 years. The analysis of the data began with an assessment of the reliability of the measures used in the study. One item was dropped from the perceived challenge and the perceived skill constructs to improve the measurement properties of the measures. All of the study constructs exhibited acceptable reliability with Cronbach’s α ranging from .73 to .93. We next assigned subjects to one of the four groups of online shoppers (i.e., flow, anxiety, boredom, and apathy) based on the relationship between perceived challenge (PC) and perceived skill (PS). An examination of a scatter chart of perceived challenge and perceived skill revealed that online shopping was not considered a challenging task and that subjects believed they had considerable skills for the task. This is a common perception, but using a sample of college students made it more so. We used a median split to segment subjects into the four groups. If perceived challenge or perceived skill exceeded or was equal to its median, it was coded as 1; otherwise it was coded as 0. Results are presented in Table 1. Table 1. Results of Segmentation

PC Value >=2.6 >=2.6 =PID?

Yes

Application: 15. Derive Key, (with P3-Secret) 16. Decrypt/encrypt

P3-Secret

ElPaso Army Hospital Data

No

Reject Service

Secure Server:

4. PID & Data

Madigan Army Hospital Data

Secure Server

Internal

Secure Server

Data Server VA @ Indy (Kernel)

Key

Z,P(x) 10. Polynomial Generator

(K1)

Notations: P3-Secret: Personal Permanent Portable SP3: Server P3-Secret SS: Secure Server Red: New Modules

Fig. 1. ACP based Secure Medical Information System

suited for MIS systems. Due to the power of our new ACP mechanism (see Section 3.2), different functions/services required in the MIS can be implemented efficiently and flexibly. We have implemented our ACP mechanism into a real MIS. The result has been recently featured by the news [11]. We present the result and describe the architecture of the implemented secure MIS system here. Some parameters and data we have cited are either normalized or shifted so as to maintain the patients' data privacy as put forth in the agreement with the VA and DoD Sponsors, but we still retain sufficient context and clear examples for the purpose of illustration. Moreover, besides securing MIS systems, the ACP mechanism can be applied to various other collaborative computing applications that require all those security needs, such as financial systems, information forensics, government systems, distributed signal processing, etc. 3.1 System Architecture Figure 1 shows the architecture of the ACP based secure MIS system. Three main security components are added into the existing MIS system: ACP Key Management Server (KMS), Secure Server (SS), and User Application (UA). The UA module is integrated in the existing client application, the SS is configured to stand at front of each existing database/application server and the KMS offers comprehensive key management services (It is worthy to note that the integration did not change anything in the existing systems). The three components work together to provide comprehensive security services involved in MISs. Let us begin with the simple case that a user (e.g., a doctor) accesses a patient data and how access control and confidentiality are enforced, and then dig into more interesting but complicated functions and how does the ACP mechanism support them.

Dependability and Security in Medical Information System

553

When a doctor accesses a patient's data (assume that the doctor has logged in and been authenticated) (Step 1,3 in Figure 1), the request with the patient's basic information is sent to the SS (Step 2) which is standing in front of the database server. The SS passes the request to the database server and obtains the patient's record including patient ID (PID) and other required fields (Step 4). The SS then passes the user's UID and the patient's PID to the KMS (Step 5). The KMS then checks the access control hierarchy against (UID, PID). Because the hierarchy specifies which groups of users can access which groups of patient records in which granularity, the KMS can validate whether and how the user can access the patient (Step 9). If yes, the KMS generates a random key and the ACP polynomial A(x) according to the hierarchy and hides the key in P(x ) (Step 10) and sends (z, P(x)) back to the SS (Step 11, See Eq.( 1) and Eq.(2) in Section 3.2 for the definition and formula of A(x) and P(x)). The SS then derives the key from P(x) (Step 12) and encrypts the data (Step 13) and sends the data along with (z, P(x )) to the user (Step 14). The user can derive the key from P(x) (Step 15) and the decrypts the data (Step 16). Note: (1) the SS is treated as a (super)user and belongs to the highest node of one organization's hierarchy. (2) the ACP mechanism does not constrain data types. Patient data could be in the form of text, images, video, audio, multimedia, etc. Our security scheme can seamlessly encrypt all of them and guarantee their confidentiality and integrity. 3.2 Access Control Polynomial (ACP) Access Control Polynomial (ACP) is the core component for key management and secure information exchange/sharing among a group of users. We briefly discuss it here. Assume that every valid user (doctor, physician, nurse, etc.) in the system is assigned a Permanent Personal Secret (PPS), denoted by SIDi for user U i . Assume q is a large prime which forms a finite field

Fq and acts as the system modulus.

Whenever there is a group of users participating in a health care service, the Key Management Server (KMS) constructs a polynomial A(x ) in finite field Fq [ x ] as:

A( x) = ∏ ( x − f ( SIDi , z ))

(1)

i∈ψ

whereψ denotes the group under consideration and assigned to the members in ψ .

SIDi are group members' PPSs

f ( x, y ) is a public one-way hash function and z is a

Fq . A( x) is called an Access Control Polynomial (ACP). From Eq.(1), it is apparent that A( x ) is equated to 0 when x is substituted with f ( SIDi , z ) by a valid user ψ ; otherwise, A( x) is a random value.

random integer from

The KMS selects a random group key K for group and computes the polynomial:

P ( x) = A( x) + K Finally, the KMS publicizes

( z , P ( x)) .

(2)

554

X. Zou et al.

From this public information, any group member

U i can get the key by:

K = P( f ( SIDi , z ) )

(3)

U i computes f ( SIDi , z ) first and then substitutes into P ( x) . For any other member U r excluded by ψ , P ( f ( SIDr , z )) yields a random value from which U r cannot get K. This key management mechanism guarantees that only a user whose SIDi is included in A( x ) can extract the key from P ( x ) . Here

With this scheme, dynamic groups can be easily managed to accept and revoke users. If a new user U t needs to be added, the KMS creates a new SIDt and assigns it to

U t . Then, the KMS includes ( x − f ( SIDt , z )) in the formation of A( x) as: A' ( x) = ∏ ( x − f ( SIDi , z ))( x − f ( SIDt , z ))

(4)

i∈ψ

A' ( x) is used to mask key K by computing P' ( x) = A' ( x) + K . Then ( z , P' ( x)) is sent to U t . U t can use SIDt to derive the key from Eq.(3).

U t needs to be revoked from the group, the KMS just selects a new random z ' and recomputes A' ( x ) by excluding the corresponding ( x − f ( SIDt , z ' )) . Then, the KMS selects a new group key K ' , computes P ' ( x) = A' ( x) + K ' , and multicasts ( z ' , P ' ( x)) . Now, the deleted user U t cannot extract K ' from P ' ( x ) . If a current group member

3.3 Security Functions Supported by ACP All the security functions, i.e. Secure Group Communication (SGC)/Secure Dynamic Conferencing (SDC), Differential Access Control (DIF-AC), Hierarchical Access Control (HAC), prevail in the MIS and are supported by our ACP mechanism. The central component in the MIS system is hierarchical access relation between users (e.g., doctors, nurses, hospital clerks, insurance personnel, etc.) and patients. This is a typical HAC scenario which has been correctly implemented and enforced by our ACP mechanism. For example, in Figure 2, suppose a physician p2 in node P2 wants to access patient data

d8 in node P8 . The KMS will generate a random key K.

The patient data will be encrypted using K before the data is sent back to the physician. The K will be hidden in the polynomial P ( x ) = A( x) + K where

A( x) = ( x − f (CID8 , z ))( x − f (CID5 , z ))( x − f (CID4 , z ))( x − f (CID2 , z )) and CIDi is the secret node ID for node i and it is only known to the users of node i. Thus, physician p2 can obtain K first and then decrypt the patient data and use it.

Dependability and Security in Medical Information System

P1

P2

P3

P6

555

P4

P7

P5

P8

P9

Fig. 2. An access hierarchy in MIS The Differential Access Control (DIF-AC) also appears in MIS. One basic format of DIF-AC is that one patient has multiple doctors such as his family physician and multiple specialists who share the patient' information; on the other hand, a family physician can have multiple patients. Another format of DIF-AC is the following many-to-many relation: for any patient in a node, his data can be accessed by all the users in this node and the ancestral nodes; correspondingly, a user in a node can access the patient data of his node as well as all his descendant nodes. Yet, some other DIF-AC format is that a doctor transmits some notifications to his patients or a hospital announces some events to certain patients. No matter which kind of DIF-AC, our ACP scheme can support and enforce it correctly. SGC/SDC scenarios occur often in MIS. (Secure Group Communication (SGC) refers to a setting in which a group of members can communicate (or share the information) among group members, in a way that outsiders are unable to understand the communication (or the information) even when they are able to intercept the communication (or steal the information). Secure Dynamic Conferencing (SDC) refers to a scenario where any random subset of the given universe of users can form a secure communication (sub)group.) One case is that a patient's data is located in several databases. The data need to be collected from all these databases and encrypted for transmission. Thus, the SSs in front of these databases form a temporary conference and the KMS generates a key for this transmission and hides the key in the P(x ) which is formed from these SSs. Another typical scenario is tele-medicine. Our ACP based SGC scheme enables tele-medicine to be highly secure with great power. Apart from its supports to SGC/SDC, DIF-AC, and HAC, the ACP mechanism also provides several important advantages for MIS such as seamless integration of different data sources, fine-tuned access control in flexible granularity, and portable login and accessing to patient data from anywhere. 3.4 Tele-Medicine Service Tele-medicine is the utmost goal of MIS systems. In tele-medicine, physicians and specialists from geographically distant locations collaborate to investigate some common disease such as SARS, diagnose a patient illness, and participate in or instruct a tough surgery. Most of them need to share the real-time image signals/streams and audio/video information. The size of data transmitted through the Internet could be huge (such as images). Therefore, the communication channels

556

X. Zou et al.

would be unstable due to the long distance. The above ACP based security mechanism is generic and independent from data content, formats, or volumes; therefore, it can be applied to secure tele-medicine. As for reliability of tele-medicine, we will introduce our new grid-computing based reliability mechanism which can provide comprehensive reliability services for MIS in Section 4. Here we propose using an effective technology to improve the reliability for data transmission: checkpoint. During the transmission, checkpoints are set up. At each checkpoint, the sender and receiver exchange signals to make sure the data that have been received before this point are correct. The signal sent by the receiver can include the check information. The sender will compare the signal from the receiver with the real values. If matched, the information before the checkpoint has been correctly received. Thus, the data being transmitted between two checkpoints is temporarily cached in the memory. Otherwise, some errors occurred during the communication, and the cached information between the two checkpoints will be resent. When it passes the check, the temporary cached data can be deleted and the cache storage can be used for the next section of data. As a result, this scheme can reach a very high reliability and is very efficient compared to the prior scheme which includes resending the large file again when the receiver cannot read it (perhaps just due to the loss/error of a small section of information). In addition, the overhead of this scheme is small ( ]]> .

Fig. 5. The Tailoring Taglib and the XML file DTD

An additional tool, the AMPI tool (Figure 3), can be used by an interface specialist to manipulate the XML files. It enables a Web service to change, include, remove and search code according to the XML structure. Each operation is logged to future audit and to control the changes. The framework also provides a structure to make decisions based on the action plan. The “switch” and “case” tags are comparable to the conventional “switch/case” programming structure, but in this case the decision is based on the action plan. Figure 6 presents their definitions, the “switch” contains an attribute to specify the interface component and the “case” contains the action over this component. For example: means that the action has effects over a textarea; means that if the action is acBIGTEXTAREA then the following code is chosen . The “tailoring” and “switch/case” tags can be also combined to include interpreted dynamic code on the case blocks.

A Semiotic-Based Framework for the Development of Tailorable Web Applications

867

Fig. 6. The Switch and Case Taglib

In order to evaluate the basic architecture and the TF, we have applied them in a proof of concept to develop of a Web application in the context of an e-Government project that we are conducting [4]. The TF was applied in a minimalist Web site design where the citizens could participate of the project giving opinions and including questions about issues related to the main public areas such as: education, healthcare, tributes, sport and transport. We have explored three scenarios: the access by people with subnormal vision, the access by people with difficulty to use the mouse, and a business requirement change that results in the necessity to audit the site. In the first and second scenarios the changes occur in the interface, the TF change the “.css” files according to the action plan. In the third scenario a Web service that does logging is activated due to the action plan results; this feature is implemented by other architecture components. Although this Web site was not big enough to stress all the architectural features, it pointed out some aspects to be addressed in the practical application in real use: performance and scalability issues, security issues, and the complexity of specifying and maintaining the norms specification. Nevertheless, the main objective of the architecture, to provide flexibility, was achieved.

5 Conclusion Flexibility and adaptability are key qualities to provide Web applications suitable for the individual as well as for groups of people. By using tailorable applications, specialists on social inclusion, domain specialists and the users themselves could adapt and maintain the Web interface according to their needs. In this paper we presented a normative approach to the construction of tailorable Web applications. By using Organisational Semiotics theory and methods, norms are used to represent pragmatics and social aspects of the system usage, in addition to the syntactic and semantic aspects. The Tailoring Framework aims to improve the productivity in constructing Web based applications. The TF is an extension of the Struts Framework and use a Web service based infrastructure to define Norms and to link them to user interface components. Through new JSP tags the developer can define the flexible components encapsulating the access to the basic infrastructure. Examples of each Tag usage were provided in order to facilitate the general understanding. As further work we propose the implementation of new tags involving other semiotics concepts aiming to improve the interface flexibility. Scalability and security

868

R. Bonacin, M.C.C. Baranauskas, and T. Medeiros dos Santos

aspects will be analysed in the future investigations with more complexes real world case studies. We are also considering to adapt the TF to new Java technologies such as: Java Server Faces and Struts 2.0 technologies.

References 1. Apache Struts Framework, ( last access November 14, 2006). http://struts.apache.org/ 2. Bonacin, R., Baranauskas, M.C.C., Liu, K.: Interface Design for the Changing Organisation: an organisational semiotics Approach. In: Liu, K. (ed.) In Virtual, Distributed and Flexible Organisations, vol. 15, pp. 236–256. Kluwer Academic Publishers, Boston, MA (2004) 3. Bonacin, R., Baranauskas, M.C.C.: An Organisational Semiotics Approach Towards Tailorable Interfaces. In: Proceedings of 11th Human-Computer Interaction International, vol. 3, CD-ROM (2005) 4. Bonacin, R., Simoni, C.A.C., Melo, A.M., Baranauskas, M.C.C.: Organisational Semiotics: Guiding a Service-Oriented Architecture for e-Government. In: Baranauskas, M.C.C., Liu, K. (eds.): Proceedings of International Conference on Organisational Semiotics, pp. 47–58 (2006) 5. Curino, C., Quintarelli, E., Tanca, L.: Ontology-based information tailoring. In: Proceedings of the 22nd International Conference on Data Engineering Workshops (2006) 6. Gibson, J.: The Ecological Approach to Visual Perception. Houghton Mifflin, New York (1979) 7. Kahler, H., Morch, A., Stiemerling, O., Wulf, V.: Special Issue on Tailorable Systems and Cooperative Work, Computer Supported Cooperative Work, vol. 19. Kluwer Academic Publishers, Boston, MA (2000) 8. Liu, K.: Semiotics in information systems engineering. Cambridge University Press, New York, NY, USA (2000) 9. Liu, K., e Dix, A.: Norm governed agents in CSCW. In: Proceedings of First International Workshop on Computational Semiotics. University of De Vince, Paris (1997) 10. Newcomer, E., Lomow, G.: Understanding SOA with Web Services. Addison Wesley, London, UK (2004) 11. Peirce, C.S.: Collected Papers, pp. 1931–1958. Harvard University Press, Cambridge, Mass (1958) 12. Repenning, A.: The Pragmatic Web: Customizable Web Applications, In: Workshop on End-User Development, Computer/Human Interaction (2003) 13. Singh, M.P.: The Pragmatic Web. IEEE Internet Computing (May - June 2002) 14. Simoni, C.A.C., Baranauskas, M.C.C, Bonacin, R.: User Autonomy in Requirements Changing Supported by Organizational Semiotics and Tailoring. In: Manolopoulos, Y., Filipe, J., Constantopoulos, P., Cordeiro, J. (eds.): Proceedings of 8th ICEIS - International Conference on Enterprise Information Systems, pp. 493–496 (2006) 15. Stamper, R., Liu, K., Hafkamp, M., Ades, Y.: Understanding the roles of signs and norms in organizations – a semiotic approach to information system design. Behaviour & Information Technology 19(1), 15–27 (2000) 16. W3C. Semantic Web Activity ( last access January 22, 2007) http://www.w3.org/2001/sw/ 17. W3C. World Wide Web Consortium ( last access January 22, 2007)http://www.w3.org/

Travelling Stories: Mobile Applications for Storytellers Anxo Cereijo Roibás1, Nina Sabnani2, and Riccardo Sala3 1

SCMIS,University of Brighton (UK) National Institute of Design (India) 3 Dare (UK)

2

Abstract. This paper discusses the outcomes of a research carried out in collaboration with the BT Mobility Research Centre and the National Institute of Design in India, aimed of understanding appropriate applications for the use of mobile phones as leisure multimedia devices for nomadic users such as commuters and travellers. This work intended to go beyond the use of mobile devices to broadcast of TV or download of music, video clips, paying special attention to the contextual usage of this media and trying to solve some unsolved issues for these interfaces as the low sociability, creativity, contextual sensitivity and interaction that so far they enable. It prospects a use of mobile interactive multimedia systems in future communication scenarios in which users can create and share self-authored & contextual digital content. Websites such as YouTube, AOL and Yahoo providing access to personal videos that have been taken using webcams, video cameras or mobile phones, evidence an emerging trend where users become authors of multimedia content. This selfauthored content production is finding application in different areas: information (travel, finance, mortgages, cooking, culture, health, etc), entertainment (sports, gossips, performance, etc), government, commerce, etc. For example, BeenThere and TheWorldisnotFlat are user generated travel sites where people can share tips about places to go on holiday. Moreover, some major newspapers like The Guardian, use this content in their Travel section. Furthermore, other more structured websites link the videos to specific places – using, for example, Google maps - enabling users to locate the videos in a map, relating the self-authored content to a specific context. Another interesting example of self-authored content is http://www.wefeelfine.org, which is an ‘exploration of human emotion on a global scale’, or in other words, a navigation among different people’s feelings (self-authored texts, sounds, pictures or videos) and emotions in the past few hours. These feelings are organized by the users into six formal movements titled: Madness, Murmurs, Montage, Mobs, Metrics, and Mounds.User centered design methodologies that take effectively into account peripatetic users interacting in their real contexts are crucial in order to identify realistic scenarios and applications for pervasive interactive multimedia systems that provide positive user experiences. This article supports the statement that handhelds due to intrinsic attributes such as friendly multimedia production tools (video, pictures and text mainly), ubiquitous presence, communication capabilities and nimbleness to dialog with surrounding platforms such as iTV, PCs, PDAs, in-car-navigators and smart-house deployments, are highly plausible tools to support users’ creation and distribution of self-authored multimedia content in pervasive communication scenarios. This paper explores the futures J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 869–880, 2007. © Springer-Verlag Berlin Heidelberg 2007

870

A. Cereijo Roibás, N. Sabnani, and R. Sala of pervasive interactive multimedia systems, and in particular the user experience related to the generation and publishing, broadcasting and narrowcasting of self-authored multimedia content through mobile devices. For this scope, it analyzes some traditional storytelling methods and tools, specially the kavaad that is still in use in Rajasthan, India - in order to understand mechanisms can be the most suitable for storytelling self-production and sharing.

1 Introduction 1.1 Mobile and Pervasive TV Becoming interactive, TV is replacing traditional 'passive' TV platform (Spigel et al., 1992) through the increase of active participation by the viewers, substantially influencing people's experience with television and their TV-related social behaviour (Lee et al., 1995). As shown in table 1, several network operators in Europe, the USA, Japan, Korea and Canada are starting to broadcast TV on handhelds. Table 1. Commercial and Trial Mobile TV Worldwide

This is commonly defined as mobile TV (Fig. 1, Source: Yankee Group, 2005).

Fig. 1. TV broadcasting on a handheld

Users’ adoption of powerful handhelds with multimedia features and an increasing interoperability between platforms is resulting not only in expanding the iTV consumption beyond the domestic context, but also in supporting a ubiquitous TV presence. We can define this ‘almost everywhere TV’ as ‘pervasive TV’. While mobile TV regards handhelds as interfaces to receive and interact with TV content, pervasive TV considers a whole system of interfaces (TV, PC, mobile phones, Public Digital Displays, etc) that can be chosen by the user according to his specific context (Cereijo et al., 2005). A good example of TV ubiquity can be already found in the 2000 UK edition of the reality TV show Big Brother. As shown in Figure 2 this program has been made accessible through four interfaces (TV plus iTV, mobile phones and the Web). Users could access information about what was happening in the ‘house’ from

Travelling Stories: Mobile Applications for Storytellers

871

any location and at any time with the most practicable interface. More interestingly, each interface allowed different applications (e.g. forum and archives in the Web, voting, additional textual information and videos on iTV and alerts and news on mobile phones).

Fig. 2. Big Brother UK on interactive TV, mobile phones and the Web

Each interface (PC, iTV, mobile phone, PDA, car navigator, etc.) has its own specific characteristics from an HCI (screen size, resolution, etc.) and a technical (memory, transfer. info speed, processing capability, etc) points of view. HCI designers need to optimise the usability of such interfaces by identifying service formats that are most suitable and distinctive interaction patterns (Weiss, 2002). There are several reasons to believe that the future of pervasive TV won’t consist of merely broadcast of traditional TV content on mobile devices. The first one is related to the intrinsic physical diversity between both interfaces (TV and handhelds) making them unsuitable for the same way of delivering of content. The second regards the context of use: TV is traditionally used in a domestic private environment (Spigel et al., 1992) and usually involves a social sharing (Morley et al., 1986) while mobile phones are mainly used in public environments, entail an individual experience (Perry et al., 1001) and generally the viewing time is much shorter. Moreover, unlike TV, handhelds are regularly used in different situations and with different purposes and they are likely to be used as an auxiliary tool to assist users' in a main activity (in this sense, mobile content could be related to the specific context of the user - context awareness) (Harper, 2003). Finally, there are also operability differences: TV (including interactive TV) is considered a passive or low interactive medium while handhelds entail a high interactivity and connectivity. Therefore, broadcasting of TV programs on handhelds is likely to be as deluding as interactive TV was. In other words, pervasive iTV will be something else and will have to do with issues such as socialibility, context awareness, creativity, interactivity, convergence (iTV, mobile phones, in-car-navigators & Internet) and connectivity (one to one and one to many). These dissimilarities influence the way of interacting with them and therefore imply distinct interaction patterns and content as well as different service formats and features.

2 Creating Future Scenarios Due to the high uncertainness about technological trends and users’ future needs and expectations, the analysing the user experience in future scenarios of mobile and ubiquitous interactive multimedia for leisure is not an easy task. In fact, rapid changes

872

A. Cereijo Roibás, N. Sabnani, and R. Sala

in users’ habits and technological advances have generated enormous uncertainties and call for innovative research and development methodologies (Cereijo et al., 2006). This research is structured in 2 main parts that ran in parallel: the first one aimed to unfold what users would expect from interactive multimedia systems for leisure while on the move through their mobile phones. We’ve denominated these interactive multimedia systems as Mobile iTV in order to use an easily comprehensible concept for the users; however, we made clear that this notion should be abstracted from the traditional model of TV broadcasting. The second part intended to understand how crucial issues for this media – especially in nomadic situations - such as sociability and context sensibility could be solved with mobile phones. To do this an ethnographic research (consisting in observation and interviews) about how storytellers in Rajasthan used interactive boxes (Kavaad) to gather people together, entertain and share experiences related to their specific context. 2.1 Part One: Unfolding Users’ Expectations This phase has been carried out in collaboration with the BT Mobility Research Centre in the UK. Its methodology consisted of initial focus group sessions with representatives of the users (commuters and travellers have been selected as target group1). Each workshop involved around 13 participants living in South England and aimed to get the users view about trends on multimedia mobile applications, TV at home and on the move, new forms of content for mobile TV, advanced interaction possibilities and finally, possible interconnections between handhelds and other devices. This activity has been combined with a theoretical investigation of existing technologies and successful interactive user experiences in other areas (e.g. games). This phase also included ethnographic research using Cultural Probes, questionnaires and naturalistic observation (photo/video recording in-the-field and data analysis). While focus groups and analysis of study-cases were good sources of functional and data requirements; Cultural Probes and questionnaires provided good information about users’ requirements and finally in-the-field observation has been a very valuable technique to identify environmental and usability requirements. Cultural Probes aimed to get inspirational responses to understand beliefs, desires, aesthetic preferences and cultural concerns of users without observing them directly. This technique that was initially used by Gaver to find new features in community design (Gaver et al., 1999) has been recently exported to HCI (Hulkko et al., 2004). Six selected users have been given a pack with the probes material under condition of returning it back completed after two weeks. Each pack (Fig. 3) included four main items with the following indications: 1

It has been assumed that commuters and travellers are a particularly relevant and interesting population segment for investigating novel processes and forms of interaction with mobile multimedia content. The growing interest and dedication to mobility and mobile life among commuters and travellers are strikingly shown in the rapidly increasing share of resources used for mobile life. Use of public and private transportation and walking in urban mobility has been increasing steadily, as has the amount of money spent for mobility and telecommunications (Pooley et al. 2005).

Travelling Stories: Mobile Applications for Storytellers

873

• Maps: World (‘where would you imagine having a daydream?’); City (‘Where would you like to go but you can’t?’); House (‘Where would you like to be alone?’, ‘where would you like to meet people?’); Family, friends and colleagues relationships (‘show frequency and nature of contacts’). • Questionnaires: a set of eleven postcards aimed to provide a very informal and open approach, encouraging instinctive and casual replies about the users’ vision on the topic and were distributed to twelve target users. • Camera: ‘take a picture of an image/video you’d like to take with your mobile’. • Media Diary: ‘record TV, cinema and radio use (what, when, where, with whom)’. • Photo album + colour pencils: ‘collect things, images and stories of your week; make sketches’.

Fig. 3. Cultural Probes packs

Review of Results of Part One Unsurprisingly, this research uncovered a scarce users’ appeal in having broadcasting of traditional TV (or iTV) formats on their mobile phones (except some exceptions such as brief life updates of a decisive football match or extraordinary news). Mobile and pervasive iTV will have to do with issues such as socialibility (Lull J., 1980), context awareness2, creativity, interactivity (Palen et al. 2000), convergence (iTV, mobile phones, in-car-navigators & Internet) and connectivity (one to one and one to many). Therefore the concept of mobile and pervasive iTV will likely have more to do with the emerging of mobile communities that are a sort of 'DIY producers' of multimedia content: they will create multimedia content in specific contexts and with precise purposes and share it with others. Moreover, the questionnaires, observations and focus groups revealed two main categories of users in terms of sending multimedia messages (photo/video with or without text & sound): − Spontaneous or impulsive user (e.g. when travelling, during an exciting night out, when sighting an interesting thing, place or performance, or just to update on domestic issues such as children, new partner, etc). The addressees are the members of the user’s restricted social personal circle: family, friends & colleagues. − Reiterative or structured user (e.g. mob blogs). The addressees belong to a broader social circle such as enlarged communities. 2

A recent definition of context-awareness is due to Dey & Abowd (1999) who defined it as “any information that can be used to characterise the situation of an entity, where an entity can be a person, place, physical or computational object” and “the use of context to provide task-relevant information and/or services to a user, wherever they may be”. Context is becoming increasingly important in handheld and ubiquitous computing, where the user's context often changes rapidly (Pascoe et al. 1999).

874

A. Cereijo Roibás, N. Sabnani, and R. Sala

Also identified were users’ preferences when receiving multimedia content on their handset from people, places or things: ‘If on the move, better if related to my context’. Context awareness provides customized information that can be defined as the right information in the right place and in the right time. This relates to the work of Abowd and Mynatt (Abowd and Mynatt, 2000) who apply a set of five questions to obtain what they call a good minimal set of necessary context. These are who, what, where, when, and why questions related the context of the system. The cultural probes showed clearly the desire of users to access TV on their handhelds with two main purposes: as an enhanced democratic tool (e.g. voting on public issues or having ‘5 minutes of fame’) and to leave their ‘signature’ along the way (e.g. by putting down personal digital content on public digital boards). Therefore, the concept of mobile or pervasive iTV will probably be related to the emerging of mobile communities that are a sort of 'DIY producers' of multimedia content: users that create content in multimedia formats and share it with others. 2.2

Part Two: Encouraging Sociability in Context

This phase of the research has been carried out in collaboration with the Indian Institute of Design in India and involved ethnographic methodologies such as observation and interviews with storytellers and kaavad makers in the Rajasthan region and aimed to analyze how storytelling making and sharing (or socialization) processes can be stimulated by means of an interactive tool (the kaavad) and understanding at the same time the dynamics of this process in order to apply this concept to catalyze contextual sociability in context among travellers and commuters using ordinary mobile devices. Storytelling Tradition in Rajasthan Telling stories was and still is an effective way of communicating and getting people together. Once upon a time when there was no television, no radio and no cinema there were stories and storytellers. These stories were both, verbal and non-verbal. The cave paintings tell the story of what was important to our ancestors and then much later, when events needed to be remembered, record-keepers were especially assigned the task of memorizing and narrating the achievements of the community. Soon, the record keepers turned storytellers to hold an attentive audience mesmerized. Some took this up as a profession and began to travel. These storytellers used to travel from place to place, telling stories in which they wove fantasy and information about people in strange and magical places, triggering the imagination and curiosity of their listeners. Music and dancing were integral part of the telling and often included chanting and playing of instruments. In India, telling and listening to stories was traditionally considered very sacred. (Miller, 1994) Even today, amongst believers, it is very auspicious and sacred to hear and tell stories from the great epics. Such a tradition of storytelling is still alive in Rajasthan, India. Rajasthan has a rich tradition of storytelling and some forms are better known than others are. Most popular is the Kathaputli or string puppets, where the heads are carved from wood and painted. The body is usually made of stitched cloth and the hidden puppeteer manipulates them with strings attached to the heads of the puppets (Ahuja, 1994). The performance usually includes puppets moving to music, narration and sound effects. The stories are small vignettes and do not necessarily connect one

Travelling Stories: Mobile Applications for Storytellers

875

to the other. The other two forms of storytelling are more religious and belong to the Phad and Kaavad tradition. (Jain, 1998) Both are traveling shrines, however, the Phad is a painted story scroll about local heroes and the Kaavad is a three-dimensional shrine with painted stories from the great epics of mainstream Hinduism (see Fig. 4). The Kaavad tradition dates back to the 16th century AD when the vernacular languages made it possible for the epic legends of Ramayan and Mahabharat to be told in local dialect and language. This tradition may have evolved out of a need to integrate with mainstream Hinduism and facilitate worship at the doorstep when travel may have been difficult for most people.

Fig. 4. The researcher, the storyteller and the kaavad maker

Fig. 5. The kaavad closed and opened

The Kaavad Shrine The Kaavad is a traveling temple/shrine, which comes to the devotee rather than the devotee coming to the temple. It consists of several wooden panels, hinged together that are iconic of a temple (see Fig. 5). There is a sense of traveling inside a real temple as the panels open and close. A typical Kaavad has a text up front that tells its origins and explains how the storyteller will use the donations given by the audience. There is a secret panel at the bottom, which serves as the donation box at the end of the performance. The panels have painted images from the epics of Ramayan and Mahabharat. The stories on the panels may differ from one Kaavad to another but the inner sanctum sanctorum always has the painted idols of Ram, Sita/Janaki and Lakshman from the holy epic, Ramayan. (Bhanawat, 1975). There are several kinds of Kaavads used by different communities. The Telling The stories can be told in the day or night but never on an eclipse day. The Kaavad is worshipped with incense before the panels are opened. In the true epic mode, the storyteller introduces himself, his village and what story he is going to present. He also tells his audience about the benefits of listening to these stories. He squats on the ground, holding the Kaavad in his lap, lights the incense sticks, says a small prayer and begins the narration by pointing at the images with a peacock feather (see Fig 6). As the stories are told, the panels are opened and more stories are narrated. Eventually all panels are opened and the devotees have an audience with the gods. Anyone may come and listen to the stories. It is open to all regardless of caste or creed. The storyteller gives an overview of the stories on the panel, introducing all the characters. After making the general introduction, he returns to each image and tells the story in detail. Anyone from the audience can stop him and ask for explanations. If the audience wishes to move onto another story, he may skip some and move to the

876

A. Cereijo Roibás, N. Sabnani, and R. Sala

next panel. As several stories are associated with the same character, he may use the same image to tell a different story each time. At the end of the telling, the devotees offer food, flowers, and gifts and are blessed in return. This blessing ensures them a place in heaven and immortality. The audience offers food and donation to the storyteller and he wraps the Kaavad in its special square cloth ( Bhanawat, 1975)

Fig. 6. The storyteller interacting with the Kavaad

Fig. 7. e-Kavaad for educational purposes

This convergence of media and performance from the oral tradition has some parallels in the technology led, Multimedia of today, where various experts collaborate and bring together media like text, image, sound and film onto a single platform; to teach, entertain, inform and inspire. The concerns of mobility and access have been addressed in the past by the makers and tellers of the Kaavad tradition and can offer insights into similar issues today as well issues of interactivity and bringing the community together. Review of Results of Part Two The idea of a mobile temple has also inspired the use of the E- Kaavad as a travelling school. Just the way, the temple came to the devotees, the contemporary Kaavad interface allows primary school children to access stories and information by touching an image and making it come alive in an animated form. The first prototypes of e-Kaavad developed in this research aimed to use the current technologies and new media to bring quality education to all, without distinction, without any barriers of language, religion or economic status. However, in a further development of the Kavaad concept led to conceive an interactive system where a nomadic teller could share his/her experiences by guiding the narrative of the story in a spontaneous way. This system consists of a shared platform that enables the teller to manage in real-time the storyboarding of his/her story. This storyboard making implies contextualizing the story (e.g. to the location, time and audience) by deciding the audio-visual content he/she wants to share in synergy with the multimedia communication ritual he/she uses to interact in real time with the audience by (e.g. talking, pointing on the visual content he/she is showing, gesticulating, singing and dancing). Contextualization is supported by a direct feedback with the audience. This nomadic system is conceptually different to user generated travel sites such as travel blogs websites where there is no physical (or virtual) co-presence between the teller and the audience and the content is not contextualized.

Travelling Stories: Mobile Applications for Storytellers

877

3 Prospected Applications From all the several personas that have been created, the following one (Table 2) represents better the major typologies of users of these prospected pervasive multimedia systems. Table 2. Summarized description of the main Persona Raoul is a single 28 year old lawyer living in East London and frequently traveling abroad for work, is interested in archaeology, traveling and practicing extreme sports.

Different future scenarios for pervasive multimedia systems have been proposed, considering various typologies of users. In this phase we used dramatizations and role playing in order to achieve a high engagement of the users during the creation of the scenarios. The scenarios were created in collaboration with 10 representatives of both target groups of users, who thereafter confirmed their reliability. From all this scenarios, the following one has been selected as more appropriate for this context. Scenario Raoul is spending two weeks travelling in Northern India combining his archaeological passion with some trekking in the Himalaya. As his schedule foresees to visit many locations every two days he moves to a new destination which implies finding a new guest house and meeting new people. Raoul loves sharing his experiences with other travellers and people from the village. For this purpose he uses the m-Kavaad application on his mobile phone to aggregate people around to illustrate a multimedia story of his travel and generate discussion. The narrative process involves a spontaneous storyboarding in which the teller decides which content show to his public dragging it into the shared application. This performance might include singing, gesticulating, dancing and other corporal and oral expressions. The m-Kavaad application has a context aware feature that informs users in the proximity about other travellers with stories to tell and matching profiles around. M-Kavaad The concept behind the mobile Kavaad goes beyond a mere mobile-container of information but intends to be a ‘shared album’ for multimedia content. It consists on a shared Bluetooth enabled whiteboard application for mobile phones using that helps the storyteller to share photos, video, audio and text for example regarding his travel experiences with other users. The storyteller might use a pen to drag the content he/she wants to share with the other users and to indicate some detail in his/her screen and the other users can see in their own devices the same content and a sign (e.g. dot or arrow) indicates the specific spot the storyteller is pointing with his/her pen (see Fig. 9).

878

A. Cereijo Roibás, N. Sabnani, and R. Sala

Fig. 8. Raoul uses his m-Kavaad to share his travel experiences

Fig. 9. m-Kavaad application

4 Future Directions The next phase avails of enactments and simulations - instead of mediated data collection as in the previous stages - and foresees the creation of proof of concept mock-ups and development of user experiments that will bring to light the relevance and feasibility of the scenarios, applications and forms of content previously identified. In this phase there are being developed (and empirically evaluated) some prototypes of applications for the concept of m-Kavaad. Three validation sessions have been organized consisting in in-situ theatre performances following the scripts of the scenarios. The plays were performed in public environments by some of the users while the others could comment on what they were watching. The advantage of these in-situ enactments is that they provided valuable information about some contextual factors that had not been identified in the research process. Finally, thanks to explorative prototypes - horizontal proof of concept mock-ups -, relevant and plausible applications that have been identified during the research are been assessed – making use of the practice of pretending - in the field. These very early prototypes incite experimentation, are easy to use and adopt, encourage discussion between users and designers and have a very low cost. However, due to their low-Fi appearance they result unconvincing raising criticism by the users. Moreover, they focus excessively on functionality if not tested in the real usage contexts. The m-Kavaad application shows how such open and pervasive interactive multimedia systems might provide an exceptional virtual platform that might foster and enhance the development of new communities of creative users that can share content and collaborate with leisure purposes during nomadic activities. As a future development of this topic we can foresee advanced mobile shared applications where users cannot only share self-authored content and moods but also create multimedia content together. In this sense it is worth mentioning Davenport’s view of the topic: ‘Since the earliest days of cinema, artists and technologists have dreamt of a future in which everyone could create and share their vision of the world. With the evolution of ubiquitous mobile networks and the enhanced mobile handset as creative device, we are on the cusp of realizing improvisational media fabrics as an active expression in our daily lives’3. 3

Glorianna Davenport, Principal Research Associate at the MIT Media Lab in Cereijo Roibas A. (2003), Ubiquitous media at the intersection: iTV meets Mobile Communications, Panel at the Proceedings of HCI 2003 Conference. Bath.

Travelling Stories: Mobile Applications for Storytellers

879

5 Conclusion This research tries to recognize the mutual influence between technology and society. Just as technology shapes society, we also need to investigate how society shapes technology. This particularly holds with a social technology that needs to be integrated into household routines. In making predictions about new technology we need to explore the critical disconnections between the ways in which such technologies are produced and the ways in which they are consumed, naturalized and rejected (Cereijo and Sala, 2004), (Fischer, 1992), (Lee and Lee, 1995). In order to address complex issues such as understanding, emotion, security, trust and privacy, the data gathering techniques presented in this paper focused on users rather than on their tasks or objectives with the analyzed interfaces. This research showed how the physical and social contexts have a strong impact in the users’ attitudes towards mobile interactive multimedia applications: the context influences in a positive or negative way the users’ emotions and feelings towards the interaction process, persuading or discouraging its use (Kjeldskov et al. 2004). For example, during the in-the-field assessment of the proof of concept mock-ups some users found unsafe recording video with their mobile phones in a crowded street, as they were very concerned about thefts. The user centered approaches illustrated by this article are based on ethnomethods and user studies of new and emerging behaviour and needs, focusing on several multi-disciplinary and cross-disciplinary domains of pervasive interactive multimedia systems such as sociability, context awareness, creativity, interactivity, interoperability and connectivity (one to one and one to many). The physical and social context might have a strong impact in the PQoE (Perceived Quality of Experience) of the analyzed interfaces: it influences in a positive or negative way the users’ emotions and feelings towards the interaction process, persuading or discouraging its use. This field research uncovered a scarce users’ appeal in having broadcasting of traditional TV formats on their mobile phones (except some exceptions such as brief life updates of a decisive football match or extraordinary news). Therefore the concept of mobile and pervasive multimedia systems will likely have more to do with the emerging of mobile communities that are a sort of 'DIY producers' of multimedia content: they will create multimedia content in specific contexts and with precise purposes and share it with others. This project explored the storytelling tradition in Rajasthan looking at identifying the storytelling processes, tools and methods used to convey the message to the audience in a synchronous way. Thereafter it aimed to understand whether these mechanisms and traditional interaction modalities could be ‘transferred’ to the area of HCI with handhelds to enable users to share experiences, or in other words, to socialize. In this way new communities will find themselves in new communication contexts and in new expressive situations: they will be able to create their own ‘movies’ and share them with other users, places (real and virtual environments) and objects (intelligent objects and other digital-physical hybrids). This expression of users’ creativity needs to be corroborated by interfaces that support some form of social interaction (Preece et al. 2003). Open, diffuse and pervasive interactive multimedia systems can provide an exceptional virtual platform that might foster and enhance the development of new communities of creative users that can share moods, content and collaborate with different purposes such as work, entertainment or government.

880

A. Cereijo Roibás, N. Sabnani, and R. Sala

References 1. Abowd, G.D., Mynatt, E.D.: Charting Past, Present and Future Research in Ubiquitous Computing. In: ACM Transactions on Computer-Human Interaction, vol. 7(1), pp. 29–58. ACM Press, New York (2000) 2. Ahuja, D.R.: Folklore of Rajasthan. National Book Trust India, Delhi (1994) 3. Bhanawat, M.: Phad, kaavad, kilangi. Lok Kala Mandal, Udaipur ( 1975) 4. Campbell, J.: The Hero with a Thousand Faces. Fontana Press, London (1993) 5. Cereijo Roibas, A., Johnson, S.: Unfolding the User Experience in New Scenarios of Pervasive Interactive TV. In: Proceedings of CHI 2006, Interact. Inform. Inspire, Montreal, Canada (April 22-27, 2006) 6. Cereijo Roibas, A., Sala, R.: Main HCI issues for the design of interfaces for ubiquitous interactive multimedia broadcast. In: Interactions Magazine, ACM, March-April 2004 issue, pp. 51–53. New York (2004) 7. Dey, A.K., Abowd, G.D.:Towards a better understanding of context and contextawareness. GVU Technical Report GIT-GVU-99-22, College of Computing, Georgia Institute of Technology. (1999) ftp://ftp.cc.gatech.edu/pub/gvu/tr/1999/99-22.pdf 8. Fischer, C.S.: America Calling: A Social History of the Telephone, Berkeley, CA, U of CA Press, Berkeley (1992) 9. Gaver, W.W., Dunne, A., Pacenti, E.: Cultural Probes, Interactions. vol. vi(1), pp. 21–29 (1999) 10. Harper, R.: People versus Information: The Evolution of Mobile Technology. In: Chittaro, L. (ed.) Human Computer Interaction with Mobile Devices, pp. 1–15. Springer, Berlin (2003) 11. Hulkko, S., Mattelmäki, T., Virtanen, K., Keinonen, T.: Mobile Probes. In: proc Third Nordic Conference on Human-Computer Interaction, Tampere, Finland, October 23-27, 2004, pp. 43–51. ACM Press, New York (2004) 12. Jain, J. (ed.): Picture Showmen. Marg Publications, Mumbai (1998) 13. Lee, B., Lee, R.S.: How and why people watch tv: implications for the future of interactive television. Journal of advertising research, vol. 35(6) (1995) 14. Lull, J.: The social uses of television. In Human Communication Research 6(3), (1980) 15. Miller, J.C.: The twenty-four Brothers and Lord Devnarayan: The story and performance of a folk epic of Rajasthan, India. University of Pennsylvania (1994) 16. Morley, D.: Family Television. Cultural Power and Domestic Leisure. Comedia, London (1986) 17. Palen, L., Salzman, M., Youngs, E.: Going Wireless: Behavior of Practices of New Mobile Phone Users. In: proc. CSCW 2000, pp. 201–210 (2000) 18. Pascoe, J., Ryan, N.S., Morse, D.R.: Issues in developing context-aware computing. In: Gellersen, H.-W. (ed.) HUC 1999. LNCS, vol. 1707, pp. 208–221. Springer, Heidelberg (1999) 19. Perry, M., O’Hara, K., Sellen, A., Harper, R., Brown, B.A.T.: Dealing with mobility: understanding access anytime, anywhere. ACM Transactions on Computer-Human Interaction (ToCHI) 4(8), 1–25 (2001) 20. Pooley, C.G, Turnbull, J., Adams, M.: A Mobile Century? Changes in Everyday Mobility in Britain in the Twentieth Century, Ashgate, Aldershot, Hampshire (2005) 21. Richman, P.: Many Ramayanas. Oxford University Press, Delhi (1992) 22. Spigel, L.: Make Room for TV: Television and the Family Ideal in Postwar America. Chicago, U Chicago P (1992) 23. Weiss, S.: Handheld Usability. Wiley, New York (2002)

Analysis of Internet Users’ Interests Based on Windows GUI Messages Jinhyuk Choi1, Geehyuk Lee1, and Yonghoon Um2 1 School of Engineering Information and Communications University 119 Munjiro, Yuseong-gu, Daejeon, 305-732, South Korea 2 IBM Ubiquitous Computing Laboratory, The MMAA Building, 467-12 Dogok-dong, Gangnam-gu, Seoul, 135-700, South Korea {demon,geehyuk}@icu.ac.kr, [email protected]

Abstract. It is important to unobtrusively observe users’ interactions on web browsers for the measurement of users’ interest. Moreover, the observation should be performed on the client side in real time, because the contents of web pages are dynamically subject to change in the current Internet environment. In this paper, we suggest a simple method of inferring a user’s interest for web contents by monitoring the amount of processed GUI messages while the user is reading a web page. We developed a software module that runs behind the Internet Explorer and monitors the number of processed GUI messages, the viewing time and the size of a web page. We found that the number of processed GUI messages, when properly normalized by the size of a web page, is close correlated with a user’s interest for web contents. Keywords: User interests, web browsing patterns, GUI messages.

1 Introduction Tools to assess users’ interests about a web page in real time are important for user modeling and for design of personalized services. Such tools may run on the clientside or on the server-side, but we are more interested in client-side tools because they do not restrict information sources to a specific web site and also because they can derive timely information from dynamic web sites whose contents may change frequently. One of the important requirements of a good client-side tool is that it should preserve the web browsing environment as much as possible, which in turn means that it should not depend too much on the implementation details of diverse web browsers. We supposed that one of such tools for monitoring users’ interests may be built simply based on the amount of processed Windows GUI messages while users are reading a web page. In order to verify the idea, we built a software module, called the Browser Monitoring Module (BMM), which runs behind the Internet Explorer and counts Windows GUI messages. Our experimental results showed that the amount of message traffic, though it may sound simplistic, is indeed an effective indicator of J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 881–888, 2007. © Springer-Verlag Berlin Heidelberg 2007

882

J. Choi, G. Lee, and Y. Um

users’ interests about a web page. We are planning to use the proposed method and the BMM in our future designs of various user-adaptive services, for example, personalized web browsers, personalized search engines, and recommendation systems. This paper is organized as follows. In section 2, we review some of previously proposed methods. In section 3, the concept of our method is presented briefly. In section 4, we describe our experimental procedure. In section 5, the evaluation results are shown and a concluding section follows.

2 Related Works Many studies have implicitly measured web users’ interest. Server-side analyses have shown good performance and have been successfully applied to consumer analyses of commercial web sites. The users’ interest at the server-side can be analyzed more easily because relevant information can be found in log files maintained by the server. For example, from the log file, the users’ login/logout time, the web pages that users have visited, users’ IP addresses, and so on can be obtained. However, the server-side analysis has critical limitations – only users on a specific site can be analyzed and the contents of a server are not sufficient to construct a general user model. On the other hand, through a client-side analysis, the users’ interest can be analyzed from various sites and a user model can be constructed using a wealth of information. Therefore, many researchers have focused on client-side analyses. However, this form of analysis also brings forth certain challenges, largely because there are no standardized methods to determine what users’ activities are relevant to users’ interests. Finding relevant activities is important in that an explicit user feedback method (e.g.: think aloud protocol, post interview) cannot be applied in natural web browsing environments. In order to predict users’ interest implicitly at the client side, a modified web browser was built [4]. This browser monitors the number of mouse clicks, mouse movement, the scrolling amount, and the elapse time on a page. From this method, it was found that a user’s interest for contents of a web page is correlated not with a unit activity but a combination of several activities. However, with regard to measuring the amount of scrolling, the authors counted only the number of mouse clicks on the scrollbars and measured the duration of scrollbar usage. However, users may also use up/down keys or a mouse wheel to scroll the windows. In [3], the authors measured the number of command-state changes to detect scroll activity and assessed statusbar-text changes for mouse activity. Changes may occur according to the activities but they may also arise by performing a different activity. In [6], the authors collected the results of several activities for analyzing user behaviors but the methods of detecting the activities were not described in detail. In [5], the authors built Weblogger, a tool to detect several activities on Internet Explorer. However, the detection of all events from a browser is not an easy task. In [9], the method of using the length of time a user views a document in his/her web browser as implicit feedback was investigated and their conclusion was that there is no significant relationship between display time

Analysis of Internet Users’ Interests Based on Windows GUI Messages

883

and document preference. However, they only addressed display time. In [7], the authors used bookmarking as a relevant activity, but this approach is inadequate for dynamic web sites whose contents may change frequently. Moreover, because most users may visit a web site repeatedly to obtain new content that is updated frequently, the number of visits is not a good measurement of interest for content. In numerous studies, the activities during task-oriented web browsing, such as using search engines, information seeking, and problem solving, have been analyzed. For example, in [10], the authors analyzed various user activities and document properties, but their focus was limited to organize some links for class material as high school teacher. Users browse the web not only for searching important information but also for entertainment or distraction. In other words, one of the purposes of web browsing is merely to seek enjoyment.

3 Concept of the Proposed Method We analyzed several methods that have been proposed thus far in order to identify some requirements. First, web user analyses should be conducted at the client side and in a real time manner. In addition, it is necessary to find simple but effective methods for detecting user activities that can be used as a measure of interest with minimal unnatural change to the web-browsing environment. The last aspect is that evaluation of the proposed method should be conducted with natural tasks such that users read web pages without any specific goals in mind. To meet these requirements, BMM detects Windows GUI messages while users are reading web pages and thus it is possible to measure user activities in real time without any interruption to the users. We also evaluated the proposed method under a natural web browsing environment in which users could read web pages of various topics in a desired manner.

4 Activity Detection BMM uses two main objects to detect a user’s activities: an event detector and a message hook. We developed a DLL component and register the component to the Windows registry system in order to obtain an event interface of Internet Explorer. Whenever Internet Explorer is executed, it searches the Windows registry system and gives the event interface to our component. The event detector object uses the event interface through which we can monitor processed messages of Internet Explorer while users are reading web pages. Among several messages that can be obtained from the component, we recorded the visited URL information. We also measured the viewing time using a system clock. In addition to the event detector object, we developed a message hook object to count GUI messages while users are reading web pages. Because these two objects are installed behind Internet Explorer, it is not necessary to change the web browsing environment and users can use the browser as they normally would.

884

J. Choi, G. Lee, and Y. Um

Modules Event detector Message hook

Behaviors to be detected Record visited URL and viewing time Mouse, keyboard and other GUI messages Fig. 1. The structure of BMM

4.1 Windows GUI Messages For Internet Explorer, some GUI messages are processed when a user changes the size of a window, scrolls a window, moves a mouse cursor, reloads the content of a web page, and so on. For example, WM_PAINT messages are processed mostly when a user scrolls the client area of Internet Explorer and WM_MOUSEMOVE messages are processed whenever a user tries to point an object on the client area with his/her mouse or even simply moves the position of the mouse cursor. These activities for invoking GUI messages are good indicators of users’ interests about the contents of a web page. Thus, we analyzed 3 GUI messages - WM_PAINT, WM_MOUSEWHEEL, and WM_MOUSEMOVE - and formulated the following hypotheses. 1. The number of processed GUI messages is relatively higher on web pages that contain interesting contents. 2. The amount of information in a web page affects the number of processed GUI messages. 4.2 Experimental Method Under the above assumptions, we conducted experiments to verify the positive relationship between the amount of processed GUI messages and users’ interest for the content. First, we collected 120 web pages offering information on various topics – Politics, Economics, Education, Engineering, Entertainment, Science, Health, and Sports – with varying content size. 25 subjects read each page in their own desired manner. To obtain appropriate data, the subjects were not told that some activities would be measured while they read the web pages. During the experiments, user activities while reading a web page and some measurable data were recorded in a log file for future analysis. In addition, whenever a subject finished reading a web page, a small window appeared wherein the subject recorded his/her interest and preference level for the contents of the page. There are 5 levels of interest, and subjects recorded their interest for the contents of a web page accordingly. Due to some malfunctions of the BMM in the users’ browsing environment and failures to properly obtain user feedback, 5 users’ log files were excluded. Therefore, we analyzed 20 users’ log files.

Analysis of Internet Users’ Interests Based on Windows GUI Messages

885

Fig. 2. An example log file produced by the Browser Monitoring Module

In figure 2, an example of a log file is shown. BMM records several data – the visited URL, the number of typed keys, the number of GUI messages, file size of the web page, viewing time, and the interest level as obtained from the user’s feedback. Each low represents log data of a web page that user visited. Among these data, the number of GUI messages, the file size of the web page, viewing time, and user feedback were analyzed in this experiment.

5 Evaluation The main objective of the experiment was to determine whether there is a positive relationship between the number of processed GUI messages, which is normalized by the information size, and the users’ interest level. We measured the amount of users’ interaction on a web page as follows.

Ui =

wi , fi

Vi =

ti ( i : index of web page) fi

(1)

wi is the normalized value of the number of processed GUI messages on the i -th web page, ti is the normalized value of the elapsed time on the i -th web page, and f i is the file size of the i -th web page (amount of information). Because In Eq. (1),

the absolute number of each user’s processed messages varies according to the user’s habit, we normalized each user’s wi and ti using min-max normalization and took the average of all users’

U i and Vi according to each interest level. From figure 3

and table 1, it is observed that the users caused the browser to process more GUI messages for interesting web pages. For example, the value of U i at interest level 5 is much higher than that of level 1 in all cases. We verified that the results are statistically significant on the basis of a one-way ANOVA test (p-value < 0.01). The pattern is most distinct in the case of WM_PAINT message – the value of U i increases according to the interest level. We also observed that there is a strong positive relationship between the viewing time and interest level, but there remains debate as to whether viewing time can be used as a measure of interest. [9]

886

J. Choi, G. Lee, and Y. Um -5

-5

7

x 10

4 3.5

6

3 U - WM-MOUSEWHEEL

5 U - WM-MOVEMENT

x 10

4

3

2

2.5 2 1.5 1

1

0.5

0

1

2

3 Interest level

4

0

5

1

2

(a) 7

6

5

5 Viewing time

U - WM-PAINT

5

4

5

-5

x 10

6

4

3

3

2

1

1

1

2

3 Interest level

4

5

(c)

x 10

4

2

0

4

(b)

-5

7

3 Interest level

0

1

2

3 Interest level

(d)

Fig. 3. Analysis of experimental result: (a) U-value for WM_MOUSEMOVE vs. Interest level, (b) U-value for WM_MOUSEWHEEL vs. Interest level, (c) U-value for WM_PAINT vs. Interest level, (d) Viewing time average vs. Interest level Table 1. Results of statistical significance analysis. The p-value shows us that there is a strong relationship between the number of each message and interest level. Message WM_MOSEMOVE message WM_MOUSEWHEEL message WM-PAINT message Viewing time

p-value 0.0006 0.000036 0.000013 0.00000032

6 Conclusion and Future Works Thus far, numerous researchers have attempted to obtain users’ preferences or interests implicitly for the contents of web pages by observing their interactions on a browser. This information could then later be used for information filtering, recommendation applications, and adaptive user interfaces. However, due to the limitations of interaction channels between users and computers – i.e., only a mouse and keyboard - predicting users’ interests is not an easy task.

Analysis of Internet Users’ Interests Based on Windows GUI Messages

887

In this paper, we used the number of processed GUI messages to predict users’ interest, as the messages are processed whenever users perform certain activities while reading web pages. It was found that the proposed method is simple and easy to develop while still being adequately effective. The results of our experiments showed that if a user engages in more activities that make the system process more GUI messages while reading a web page, even in the event that the page offers relatively little information, it can be inferred that the page contains interesting or preferable content. This provides an important guideline to follow, because finding preferable web pages is the first step of user modeling procedures and personalization services. In this work, we presented preliminary results, and we are currently working on a more complete analysis of the results. Here, we considered only text-based web pages for ease of defining the amount of information on a web page. However, web pages contain an abundance of multimedia objects such as pictures and videos. Subsequent experiments should consider the use of a well-defined measure for the amount of information in such web pages. Also, because users’ interaction is believed to be influenced by the layout of a web page, this aspect will also be taken into consideration in future works.

References 1. MSDN Online. Microsoft Developer’s Network. http://msdn.microsoft.com 2. Kelly, D., Teevan, J.: Implicit feedback for inferring user preference: a bibliography. ACM SIGIR Forum 37, 18–28 (2003) 3. Goecks, J., Shavlik, J.: Learning users’ interests by unobtrusively observing their normal behavior. In: Proceedings of International Conference on Intelligent User Interfaces, pp. 129–132 (2000) 4. Claypool, M., Le, P., Waseda, M., Brown, D.: Implicit interest indicators. In: Proceedings of International Conference on Intelligent User Interfaces, pp. 14–17 (2001) 5. Reeder, R.W., Pirolli, P., Card, S.K.: WebLogger: A data collections tools for Web-use studies. UIR Technical report UIR-R-2000-06, Xerox PARC (2000) 6. Hijikata, Y.: Implicit user profiling for on demand relevance feedback. In: Proceedings of International Conference on Intelligent User Interfaces, pp. 198–205 (2004) 7. Seo, Y., Zhang, B.: Learning user’s preferences by analyzing Web-browsing behaviors. In: Proceedings of the fourth international conference on Autonomous agents, pp. 381–387 (2000) 8. Fu, X., Budzik, J., Hammond, K. J.: Mining navigation history for recommendation. In: Proceedings of International Conference on Intelligent User Interfaces, pp. 106–112 (2000) 9. Kelly D., Belkin, N. J.: Display time as implicit feedback: understanding task effects. In: Proceedings of the 27th annual international ACM SIGIR conference on Research and development in information retrieval, pp. 377–384 (2004) 10. Badi, R., Bae, S., Moore, J.M., Meintanis, K., Marshall, C.C.: Recognizing User Interest and Document Value from Reading and Organizing Activities in Document Triage. In: Proceedings of International Conference on Intelligent User Interfaces, pp. 218–225 (2006) 11. Sharma H., Jansen B. J.: Automated Evaluation of Search Engine Performance via Implicit User Feedback. In: Proceedings of the 28th annual international ACM SIGIR conference on Research and development in information retrieval, pp. 649–650 (2005)

888

J. Choi, G. Lee, and Y. Um

12. Nakamichi, N., Sakai, M., Shima, K., Matsumoto, K.: Detecting Low Usability Web Pages using Quantitative Data of Users’ Behavior. In: Proceeding of the 28th international conference on Software engineering, pp. 569–576 (2006) 13. Carreira, R., Crato, J.M., Goncalves, D., Jorge, J.A.: Evaluating adaptive user profiles for news classification. In: Proceedings of International Conference on Intelligent User Interfaces, pp. 206–212 (2004) 14. Lieberman, H.: Letizia: An Agent That Assists Web Browsing. In: Proceedings of the Fourteenth International Joint Conference on Artificial Intelligence, pp. 924–929 (1995) 15. Chen, L., Sycara, K.: WebMate: A Personal Agent for Browsing and Searching. In:Proceedings of the 2nd International Conference on Autonomous Agents and Multi Agent Systems, pp. 132–139 (1998) 16. Joachims, T., Freitag, D., Mitchell, T.: WebWatcher: A Tour Guide for the World Wide Web. In: Proceedings of the Fifteenth International Joint Conference on Artificial Intelligence, pp. 770–777 (1997)

Deconstructing Web Experience: More Than Just Usability and Good Design Chandra Harrison and Helen Petrie Human Computer Interaction Group, Department of Computer Science, University of York, Heslington, York YO10 5DD, United Kingdom {chandra.harrison,helen.petire}@cs.york.ac.uk

Abstract. To design for positive web experience designers require an understanding of what factors elicit specific affective or cognitive responses. To gather user perceptions, which may offer insight for future objective exploration, an online survey asked five open ended questions regarding user perception of design aspects associated with a range of affective and cognitive states; specifically what stimulates web user’s curiosity and creativity, what challenges their abilities, what surprises them and what alters their sense of time. The majority of responses can be categorized as either usability or design factors, but 17.4% cannot and warrant further exploration to determine their utility for designing optimal user experience. Results also indicate that for each of the states there are trends in the responses, suggesting that certain aspects of web interface design elicit specific reactions. These aspects should be explored and applied if designers want to encourage pleasurable website experience. Keywords: Usability, emotional design, user experience.

1 Introduction Recent human-computer interaction (HCI) research focus has moved away from addressing just usability towards adding positive interaction aspects and enhancing emotional and aesthetic design of interfaces and websites to improve satisfaction [9]. However, some researchers recognized the role of affective and/or cognitive states in optimizing experience before it became a major research trend. More than 30 years ago Csikszentmihalyi [7] suggested that certain factors were required to create an optimal psychological experience. He suggested that if nine factors were present a person experienced ‘flow’ or optimal experience. These nine factors are clear goals, immediate feedback, skills suited to the challenge presented, merging of action and awareness, concentration, sense of control, loss of self-consciousness, altered sense of time and an autotelic response. Most of these nine factors relate to cognitive states or thoughts, but being in flow creates an optimal emotional experience. The concept of flow has endured and been widely tested in a variety of domains over the years including web interaction [4]. Some years later Malone [17] proposed three heuristics that he believed would encourage enjoyment in HCI. He suggested that curiosity, challenge and fantasy were required to create an enjoyable experience. Malone hypothesized that people who experience immediate pleasure and joy, positive affect J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 889–898, 2007. © Springer-Verlag Berlin Heidelberg 2007

890

C. Harrison and H. Petrie

in other words, from computer games are more likely to use computers than others who have not had a rewarding experience during their interaction. While he tested his heuristics in gaming, Malone suggested that the heuristics could also be applicable in other areas of HCI. However, Malone’s factors have not received as much popular attention as Csikszentmihalyi’s concept of flow, despite his original findings. In recent years more research has been undertaken in the area of user experience [10]. Research has explored the relationship between perceived beauty and usability of consumer electronics which indicates that the aesthetic appeal of a product does influence the perceived usability [26]. In addition, guidelines are available for usability, accessibility and design to enhance the user experience [20]; [6]. One failing of guidelines is that they do not include specific examples of aspects covered [13]. It is still unclear whether some of these guidelines are valid and whether they cover all aspects of user experience. As well as research and guidelines, abstract models of what will assist with emotional experience do exist [12]. Other researchers have also created more complex models in an effort to explain emotional experience [9]. Applying good usability and design practice and knowing what specific aspects of design elicit which affective or cognitive responses help designers create an overall positive experience. Despite the guidelines, models and research, users are still experiencing frustration and stress in HCI [2] and other negative responses during interaction because of poorly designed interfaces [18]; [24]; [21]. While there is evidence that websites are not adhering to the guidelines and models available, specific examples of what elicits the affective responses discussed in the literature, both old and new, are lacking. In addition, large data sets of what specific design aspects elicit certain reactions from users are not available and are needed to explore user experience further. Subjective user data provides useful insight into areas which may warrant further exploration. While gathering objective data about what aspects elicit these responses is ideal, this is a longer term goal. 1.1 Creating Positive Affect or Thoughts in Utilitarian Interaction Methods used to measure affect or emotion, pleasure and frustration are varied [16]. While objective measures such as eye-tracking, facial expression and biometrics can be used, it is necessary to first understand what should be measured. Unfortunately, affect/emotion still raises debate in the literature. There is little agreement on even the major emotions and how to measure them [22]. While agreement cannot be reached on precise terminology, one area where there does seem to be agreement is that there is a difference between transient affect states and enduring emotional experience [1]. For our purposes, Bentley et al’s terminology suggesting that affect is a transient state and emotion is the more enduring condition is adequate. They also suggest that transient affective response during HCI influences overall emotional experience. Therefore, to ensure a positive emotional experience it is important to optimize the affective responses during interaction. While some web interaction is voluntary, inherently pleasurable and is less likely to cause stress, other interaction is compulsory either because of work requirements or because alternative routes to the information, goods or services are not available. According to Hassenzahl [9] those websites we choose to use because of the inherent

Deconstructing Web Experience: More Than Just Usability and Good Design

891

pleasure are ‘hedonic’, or related only to pleasure pursuits, websites such as those for social interaction and games for example. Websites that we are forced to use at work or have to use to find information or services because of difficulty accessing the information elsewhere (i.e. out of hours), would then be referred to as ‘utilitarian’. Government websites, for example, have to be used by people who may have no alternative to find the information they are seeking. Because people have to use these websites, the usability, design and accessibility of the site may not influence whether the person uses it and because of poor interaction design negative affect or thoughts may result, possibly leading to stress. Continued or intense negative reactions to any stimuli can cause undue stress which is undesirable for physical and mental well-being [23]; [25]; [27]. In an effort to recognize the impact of the cumulative effect of stress the UK Health and Safety Executive introduced legislation to encourage more healthy interaction with technology [11]. If negative affect or thoughts during utilitarian interaction, such as using websites, does cause negative emotional experience, this could potentially lead to stress and result in physical consequences [23]. Therefore, enhancing the user experience for utilitarian websites is important to ensure quality of life. Considerable research attention has been paid to hedonic interaction such as games [5]; [17]. Other aspects such as trust in eCommerce interaction [15] and ways to facilitate user acceptance of online shopping have also been explored [19]. However, little research has been done on exploring utilitarian websites. Because utilitarian websites are those we are forced to interact with and those that are more likely to cause stress than hedonic websites and because of the push by government to reduce work and technology related stress, they make a suitable domain for testing. While affective responses such as computer anxiety [2] have been explored, little research has been done to gather user perception about affective or cognitive states that might lead to stress, such as what stimulates user’s curiosity and creativity, challenges their abilities, surprises and alters their sense of time. If trends are apparent between respondents and within the states, and responses can be categorized, they could perhaps be applied in the design of more enjoyable web interfaces and would provide insight for further objective exploration. Emotional experience is important as part of holistic HCI and involves both good usability and satisfaction. Designers need to consider not only optimum usability and design [14] but must also still strive to achieve satisfaction. So what does stimulate web user’s curiosity and creativity? What challenges their abilities? What surprises them and alters their sense of time? To explore what users associate with these states and to explore whether there are trends in perception of aspects of design that elicit these affects five open ended questions were asked in an online survey. Therefore, the research questions for this study were: 1. Are there trends in perceptions of aspects of web design that facilitate affective or cognitive responses relating to curiosity, creativity, challenge, surprise and an altered sense of time? 2. Are all of the aspects reported by respondents covered by usability or design guidelines?

892

C. Harrison and H. Petrie

2 Method 2.1 Design Five free text response questions were developed for an online survey asking respondents to describe their perceptions of what causes Curiosity, Creativity, Challenge, Surprise and an Altered Sense of Time during their personal website interaction (see Table 1). The questions relating to Challenge and Altered Sense of Time use factors from Csikszentmihalyi’s [7] concept of flow. While the dimensions of flow have been validated in other studies [8] many of them would require explanation for respondents to understand or some form of Experience Sampling Method (ESM) relating to a specific task, to measure them. Because no specific task was required in this study, several of the nine factors of flow were unsuitable in this situation. However, as an altered sense of time appears to be understood by users [4], the Altered Sense of Time question was included. Challenge is involved in both Malone’s heuristics and flow and is suggested as enhancing positive experience in both. The Surprise and Creativity questions were developed to address Malone’s [17] fantasy heuristic, with the question designed to explore perceptions of what causes Curiosity addressing another of his heuristics. An online survey was chosen for data gathering because of the relative ease with which large numbers of responses could be gathered in a short time period. It was believed that during an online survey respondents would be in their own environment, have choice about whether to answer the questions, were not under observation, and had to answer the questions from an abstract point of view. General abstract responses were asked for rather than specific incidents to gain a broad indication of user experience to explore rather than quantify. Table 1. Questions for each of the states used in the online survey State Curiosity

Creativity Challenge Surprise Altered Time

Question What things about a website make you want to look at information other than what you were looking for originally, in other words what stimulates your curiosity and makes you actively explore a website? What things about a website give you innovative thoughts, make you feel creative or make you experiment? What things about a website challenge your abilities? What things about a website surprise you thrill you or give you unexpected pleasure? What things about a website make you lose track of time?

2.2 Participants The 141 respondents to the online survey included 58 males and 83 females with an average age of 37 years (range of 18-65). Responses were received from around the world including United Kingdom (n=56), America (Canada and US) (n=56), Asia and Oceania (n=20) and Europe and Scandinavia (n=9). An assessment of respondent’s computer experience involved a simple seven point Likert scale (1 limited experience – 7 extensive experience). While the majority (86%) of respondents having a reasonably high level of computer experience. Fourteen percent had moderate or limited computer experience, offering a novice

Deconstructing Web Experience: More Than Just Usability and Good Design

893

perception. While the level of computer experience is skewed towards the higher end this is to be expected given that it was an online survey. 2.3 Procedure A URL link to the online survey was circulated via email within the author’s university and to professional and personal email lists of both authors in an effort to gain a spread of respondents, not just those involved in HCI. The link was in turn passed on to other groups internationally resulting in a wide variety of responses from sources known and unknown to the authors. QuestionPro was chosen from several other online survey tools to host the survey due to its ease of use, data presentation and efficiency of uploading the survey and downloading the results. The survey was launched online with student sponsorship from QuestionPro. The survey was available online for one month. 2.4 Data Analysis Responses to each question varied from one word to 125 words for a single question, often with more than one topic mentioned. For example, for the Curiosity question one response was ‘a nice layout, a page that has a lot of other information and has easy navigation makes me want to explore more’. This response was categorized in layout, related information and navigation. Categorization involved an initial content analysis by the first author to determine suitable categories for each of the questions. Because it was unclear what type of responses would be received from such a diverse sample, content analysis was used to categorize the responses in an effort to develop a future grounded theory, allowing conceptual interpretation to emerge from the data rather than being driven by existing theories [3]. Thirty categories were originally formulated due to repeated occurrences of similar statements by respondents. A second thorough coding followed to ensure consistency both within and between the questions. The second author analyzed a sample of the responses to assess inter-rater reliability. Due to some people implying one factor was a negative and others as positive (i.e. navigation could be cited as both poor and good in different questions or by different respondents), each of the 30 categories were separated into positive and negative, resulting in a total of 60 possible categories. Often the responses were not explicit whether they were positive or negative. However, where it was obvious the coding was deemed a negative function of its category. Across the five questions, several respondents stated that they were not looking for aspects to stimulate their Curiosity or Creativity, Challenge or Surprise them or to Alter Their Sense of Time. Comments included "I am very task orientated," "I use the internet for specific purposes and rarely if ever wander," "I really don't want these things from a website!" These responses were categorized as ‘task only’ while those that provided no answer or suggested it was not applicable were categorized as Nil. Of the original 60 categories, 26 were used by at least 5% of participants (i.e. 7 out of 141) for at least one individual question (see Table 2 for explanations of the categories). Many of those which did not reach this level were either the positive or the negative of a category that did reach the 5% response level, indicating that some categories were specific to a positive or a negative experience rather than both. The

894

C. Harrison and H. Petrie

categories that did not reach the 5% response level were eliminated from the analysis and are not reported. Because of the number of different categories, they were grouped to explore whether there were aspects other than those that could easily be categorized as usability and design. The six groups are: 1. 2. 3. 4. 5. 6.

Usability – aspects included in usability and accessibility guidelines Design – aspects relating to the content, aesthetics and layout Personal Interests– aspects relating to specific interests of respondents Commercial Considerations – aspects broadly relating to economics Experience – aspects not otherwise covered specifically Other – nil and task only responses

3 Results and Discussion 3.1 Curiosity, Creativity, Challenge, Surprise, Altered Sense of Time, The majority of respondents (82%) identified at least one aspect of web interaction that they associated with each of the five affective/cognitive states. There were a total of 705 response opportunities (141 respondents by five questions). However, many of the individual responses contained multiple concepts, resulting in a total of 905 categorizations. Table 2 illustrates the break down of group and category by the percentage and frequency of responses per question. Curiosity – The most frequently cited aspect of web interaction that stimulated curiosity was links to related information (71 responses from 141 opportunities = 50%). One respondent commented that they found "links with a bit of info, i.e. a summary of what is behind the link" stimulated their curiosity. Curiosity accounted for the least amount of ‘experience’ categories, with only the reputation of the website stimulating their curiosity. The majority of categories related to usability, design and personal interests. The prevalence of personal interests suggests that pursuing personalization of websites may encourage users to actively explore. This question also had the least nil responses suggesting a greater understanding of the concept or the question. Creativity – The content of a site was the most frequent response for this question (24 responses). The greatest variety of ‘experience’ categories were used for this question, suggesting that stimulating creativity can be achieved through incorporating good design and experience elements rather than just focusing on usability. Personal Interests also played a major role in the responses to this question (23 responses). Challenge – It was clear from the almost unanimous negative response that having skills challenged was not necessarily perceived as a positive factor. Technical aspects of the web interface presented the most challenge (21 responses). For example, one respondent commented that websites "…that require specific applications I don't have or don't want to download," were what challenged them. However, for this question several respondents stated that while nothing necessarily challenged them there were factors that ‘frustrated’ or ‘tried my patience’. While this result suggests that people see challenge as a negative thing rather than a positive, perhaps the problem is that the

Deconstructing Web Experience: More Than Just Usability and Good Design

895

respondent’s skills did not match the challenge presented [7]. Alternatively the wording of the question may have led people to answer in a negative way. Surprise – This question resulted in the most overall responses in the experience group (total of 41 responses). For example one respondent commented that "I am thrilled to see a daily saying, something positive!" This suggests that if surprise is a desired response, then including fun, social, and novel elements may facilitate this responses. Unexpected content (28 responses) and mention of bargains or ‘freebies’ (18 responses) were the most often cited aspects to cause surprise. The question was worded to elicit positive responses therefore it is not surprising that there were few negative comments. Altered sense of time – The quality of the content (61 responses) and linking to information related to the original goal (37 responses) were the two most often cited aspects that made respondents lose track of time. Few respondents suggested that losing track of time was a negative experience. From the comments it is clear respondents experienced losing track of time, but the reasons appear to be related to personal interests and the quality and relatedness of the content. The above results suggested that there are trends in identifiable aspects of web design that facilitate a positive affective response. What is also clear from the initial analysis of this data is that there are aspects that are not covered implicitly by usability or design/aesthetics guidelines or practice. 3.2 Usability, Design or ‘Experience’ Table 3 shows the breakdown of all responses in the six groups. It can be seen that the Design group accounted for the largest number of responses (24.4%), but Usability, Personal Interests and Experience each accounted for almost 20% of responses. The Experience aspects included items such as quirky features of the design (use of humor, trivia, miss-spellings, content out of context). For example, one response included "games, or lots of entertaining things to do". Others had a strong focus on social aspects and learning, for example "sites that teach you new ways to have fun and learn." While design/aesthetics and usability issues have been studied extensively in the past the results of this study suggest it is not just these aspects that are important in web interaction and positive user experience. Malone’s heuristics imply that elements such as fun are necessary to keep people using computers and this seems to be supported by this study. The frequent occurrence of interaction, fun, social and learning categories, not specifically related to usability or design, found in the results of this study suggest that affective responses such as curiosity and fantasy may in fact be valuable aspects for designers to consider. While quirky elements and unexpected content was highlighted, respondents were clear that related information that they found engaging or stimulated their creativity or curiosity needed to be relevant. For example, one respondent commented that “…links that actually add to the experience that I might not have found on my own.” However, comments such as “… (links that) try to divert you to some stupid rubbish…” indicated that irrelevant information had a negative effect. The substantial percentage of responses categorized in Personal Interests group (18.7%) indicates that web interaction is a highly personal and individual experience.

Simple Navigation (Navigation) (Technical) Colour Graphics Headings Design Content Layout (Layout) Inspiring Design Interests Related Info (Advertising) Bargains Fun (Patience) Experience Novelty Interaction Social Learning Problem Solving Task Only Nil

Usable

Other

Commercial Considerations Experience

Personal Interests

Design

Category

Group

Usability, accessibility, mention of ease of use or simplicity Consistent, clear labels, understandable, active links Inconsistent links, dead links Safety, security, jargon, skills, form filling, log in, Good contrast, good use of colour, favourite Clear - images, graphics, diagrams, pictures, flash Understandable, interesting headings, links or categories Aesthetics, nice looking, attractive Well written, concise, current, good quantity Organised, readable, uncluttered uncrowded Large blocks of text, cluttered Design ideas to use in personal interests Hobbies, sports, mention of a specific type of interest More detail, extra information on topic searched for Too many ads, obtrusive ads, popups Discounts, samples, offers, products, giveaways Humour, quirky, oddities, trivia, daily inspiration Frustration Reputation of website, mention of emotion Innovation, unique, original Games, quizzes, flash animation, involvement, feedback Blogs, email, chat rooms, personalisation, responsibility Inspiring content, learn new things Figuring out different ways to find information Focused only on the task No answer, not sure, unsure, N/A etc

Explanation

9

12

71

10 20 21 15 22 37

7 14 15 11 16 26

50

f 21 21

% 15 15

Curious

5 32

5 9 7 11

6

11 16

5 17 5

6

%

7 45

7 12 10 15

9

16 23

7 24 7

8

f

Creative

33

5

9

5

8

14 15

%

47

7

12

7

11

19 21

f

Challenge

Questions

27

10 8

7 6

19

8

18 15

8 14

9 28

14

f 14

6

13 11

6 10

6 20

10

% 10

Surprise

7

14

9

26

43

9 11

%

10

20

12

37

61

12 16

f

Time

Table 2. Groups and categories of responses with explanations (negative aspects in brackets), percentage and frequency of respondents answering for each question. The most frequent response for each question is highlighted.

896 C. Harrison and H. Petrie

Deconstructing Web Experience: More Than Just Usability and Good Design

897

Table 3. Percentage of total responses across five questions by group Group Usability Design Personal Interests Commercial Considerations Experience Other

% of total responses 21.8 24.4 18.7 2.8 17.4 15

Total frequencies 197 221 169 25 157 136 905

4 Conclusion The results of this analysis suggest that there are certain aspects of design that are associated with certain affective/cognitive response by users. The strong trends in the Personal Interests and Experience categories suggest that there are some aspects that are important that are not entirely covered in the guidelines. Personalization of websites and being able to learn and have fun and interact are suggested as ways to stimulate curiosity, creativity, surprise and exploration. The negative focus on usability faults causing challenge requires researchers and practitioners to not lose sight of the importance of fundamental elements of design. Reducing the jargon, increasing feelings of safety and security may be necessary. There does appear to be trends across the 141 respondents regarding their perception of what stimulates creativity and curiosity, what challenges abilities and provides positive surprises to users as well as what makes users lose track of time. This suggests that focusing on these aspects in utilitarian websites might add a sense of fun or pleasure, assist in enhancing the user’s experience associated with interacting with websites and reduce stress. While respondents were not asked to report on utilitarian web interaction specifically, these results suggest that user experience could be enhanced in utilitarian web interaction as well as hedonic interaction. The results from this analysis also suggest that while some aspects that are associated with the five states tested here are covered by existing design and usability guidelines, there are some ‘experience’ aspects that require further exploration to assess their utility. Interaction, learning, fun and social aspects of web interaction need to be explored.

References 1. Bentley, T., Johnston, L., von Baggo, K.: Evaluation using cued-recall debrief to elicit information about a user’s affective experiences. In: Proceedings of OZCHI 2005, Canberra, Australia. CHISIG (2005) 2. Ceyhan, E.: Computer anxiety of teacher trainees in the framework of personality variables. Computers in Human Behaviour 22, 207–220 (2006) 3. Charmaz, K.: Constructing grounded theory: A practical guide through qualitative analysis. Sage, London (2006)

898

C. Harrison and H. Petrie

4. Chen, H.: Flow on the net–detecting Web users positive affects and their flow states. Computers in Human Behavior 22, 221–233 (2006) 5. Cheng, K., Cairns, P.: Behaviour, realism and immersion in games. In: CHI 2005. ACM Conf. on Human Factors in Computing Systems, pp. 1272–1275. ACM Press, (2005) 6. Chisholm, W., Vanderheiden, G., Jacobs, I.: Web Content Accessibility Guidelines 1.0 (website accessed April 2006), http://www.w3.org/TR/WCAG10/ 7. Csikszentmihalyi, M.: Beyond boredom and anxiety. Jossey Bass, San Francisco (1975) 8. Csikszentmihalyi, M.: Creativity: Flow and the psychology of discover and invention. Harper Collins, New York (1996) 9. Hassenzahl, M.: The thing and I: Understanding the relationship between user and product. In: Blythe, M.A., Monk, A.F., Overbeeke, K., Wright, P.C. (eds.): Funology: From Usability to Enjoyment, Kluwer, Netherlands (2003) 10. Hassenzahl, M., Tractinsky, N.: User experience: A research agenda. Behaviour and Information Technology 25(2), 91–97 (2006) 11. Health and Safety Executive (HSE). Work related stress. Website accessed (January 2006) http://www.hse.gov.uk/stress/ 12. Jordan, P.W.: Designing pleasurable products. Taylor & Francis, London (2000) 13. Koyani, S.J., Bailey, R.W., Nall, J. R.: Research-based Web Usability and Design Guidelines, (2003), ( last accessed 2006-04) http://usability.gov/pdfs/guidelines.html 14. Lindgaard, G., Dudek, C.: What is this evasive beast we call user satisfaction. Interacting with Computers 15, 429–452 (2003) 15. Kaluscha, E.: The importance of initial consumer trust in B2C in electronic commerce: A structural equation modelling approach. Unpublished PhD Thesis. Universität Klagenfurt (2005) 16. Mahlke, S., Minge, M., Thüring, M.: Measuring multiple components of emotion in interactive contexts. In: Proceedings of CHI 2006, Montreal, Canada, pp.1061–1066 (2006) 17. Malone, T.W.: Heuristics for designing enjoyable user interfaces: lessons from computer games, ACM, pp. 63–68. (1981) 18. McCarthy, J., Wright, P.: Technology as experience. MIT Press, Cambridge, M.A (2004) 19. Moore, G.A.: Crossing the chasm: marketing and selling high-tech products to mainstream customers. revised edn. Harper Business, New York (1999) 20. Nielsen, J.: Usability Engineering. Academic Press, San Diego (1993) 21. Norman, D.A.: Emotional Design: Why We Love (or Hate) Everyday Things. Basic Books, NY (2005) 22. Picard, R.W.: Affective computing. MIT Press, Cambridge, M.A. (1997) 23. Selye, H.: History and present status of the stress concept. In: Monat, A., Lazarus, R. (eds.) Stress and coping: An anthology, pp. 21–35. Columbia University Press, New York (1991) 24. Shneiderman, B.: Leonardo’s laptop: Human needs and the new computing technologies. MIT Press, Cambridge, M.A. (2002) 25. Smith, J.H., Block, R.T., Smith, M.J.: Cumulative social trauma and human stress disorders in the age of technology. In: Derjani, A. (ed.) The American Society for Ergonomic Systems Engineering Conference. Madison, Wisconsin (1996) 26. Tractinsky, N., Katz, A., Ikar, D.: What is beautiful is usable. Interacting with Computers 13, 127–145 (2000) 27. Wilson, G.F, Eggemeier, F.T.: Psychophsiological assessment of workload in multi-task environments. In: Damos, D. (ed.) Multiple task performance, pp. 329–360. Taylor and Francis, London (1991)

Report on Project to Identify Excellent Local Government Web Sites: An Attempt to Educate Web Systems Developers on the Importance of Human-Centered Design Processes Seiji Hayakawa, Haruhiko Urokohara, Toshikazu Shinohara, Naoto Aizawa, and Yayoi Kitamura Steering Committee, Authentication Department Human-centered Design Organization (HCD-Net)1, c/o NOVAS, Fujita INZX Bldg. 9F, 9-5 Shinsen-cho, Shibuya-ku, Tokyo 150-0045, Japan

Abstract. In 2006, the Human-centered Design Organization (HCD-Net) began advocating the utility of the HCD process for web systems developers. In this project, we focused on the “move-in/move-out” procedures services of local-government web sites and identified sites of excellent usability. This document reports on the history and findings of the project, and suggests tasks for the future. Keywords: usability evaluations, the human-centered design process, local-government web sites, marking system, usability education.

1 Preface 1.1 Outline of Project The Human-centered Design Organization (HCD-Net) is a nonprofit organization established in March 2005 to educate developers in all fields on the human-centered design (HCD) process. During its first year, it focused primarily on integration engineers, but in 2006 developed a project to educate web systems developers about the effectiveness of the HCD process. The reason for doing so was the increasing frequency of news reports like the following: “A total of 13,000 government administration systems have been developed under the ‘e-Japan’ program. These include 166 systems to handle major procedures that generate in excess of 100,000 documents per year. For more than 80% of these 166 systems, utilization is at less than 1%.” [1] There are any number of examples of systems being developed with poor usability because of the lack of recognition in Japan for the human-centered design process. 1

The Human-centered Design Organization, is a nonprofit organization given special status under the Law to Promote Specified Nonprofit Activities.

J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 899–905, 2007. © Springer-Verlag Berlin Heidelberg 2007

900

S. Hayakawa et al.

HCD-Net therefore decided to highlight the effectiveness of the HCD process by conducting a project to identify local-government web sites of excellent usability. By applying the HCD process, local governments can create web sites that residents can use with confidence and efficiency to achieve their purposes and generate high degrees of satisfaction. Doing so will enable residents to enjoy the public services provided by local governments in the manner expected. In the next section, we provide a brief overview of the history of the project. This material was originally published in “Local Government Channel +”[2] February Edition (p. 14-19), a journal for local governments that has a circulation of 2,600 and is distributed monthly to 1,820 municipal governments around Japan and was therefore deemed an excellent vehicle for communicating with web systems developers in local government organizations. 1.2 Selection of the “move-in/move-out information” Task2 This project limited its evaluation to “move-in/move-out” services. Moving is something that many residents have experienced at least once, and it entails contact with a wide range of public information over the course of a fairly restricted period of time: outgoing/incoming registration, basic-living infrastructure, procedures for schools and elderly services, to name a few. Providing residents with the information they need for these procedures accurately and without stress is the “quality of use” that public web sites need to achieve. Web sites must be beneficial, one-stop services that residents can be expected to use, and that is why we consider it possible to compare them on the comprehensiveness and comprehensibility of their information.

2 Usability Evaluation Methodology 2.1 Defining Web-Site Usability As a medium, web sites have been used as a means of providing public information and advertising analogous to what is provided by paper media, as the term “web page” suggests. Gradually, however, greater emphasis is being placed on their utility as a tool for providing the services and information of the site operator—in other words, as a means of achieving useful purposes by those who use the site. There are therefore two different aspects of web sites that must be understood. The first is that of a collection of documents that comprise the multiple web pages used to communicate information. The second is the assumption of two-way communication made possible by screen movement and software-based functions that accomplish specific purposes. Because of this, from the very early stages it was recognized the web sites require improvements not only in the quality of the products and services they furnish in completed form, but also in the usability of the site in providing those products and services. Managing constant improvements in usability is an important component within the web site administrative cycle because by nature web sites must communicate and update information over short time cycles and be designed to respond dynamically 2

“Task” in this report refers to a series of actions taken by a user to achieve an objective.

Report on Project to Identify Excellent Local Government Web Sites

901

according to the context in which the user finds himself. The HCD process is a method of positioning different usability techniques within the design and development process, and as such offers insights that are essential underpinnings to web-site structure.

©2007 Mizuki Nishihira

Fig. 1. Persona: Hyper-Busy Businessman

©2007 Mizuki Nishihira

Fig. 2. Persona: Newlyweds who have just purchased a new home

©2007 Mizuki Nishihira

Fig. 3. Persona: Family with children

2.2 Using Personas to Evaluate Usability An effective way to evaluate usability is to posit use by a “persona,” who represents a standardized user for evaluation purposes. The personas used for this evaluation were extracted and created from the following process, which was used in the “Project to Create Electronic Support Service Concepts: Moving into Sapporo” (one of the projects in the Ministry of Education, Culture, Sports, Science and Technology’s “Sapporo Region Intelligent Knowledge Cluster: Sapporo IT Carrozzeria” program). 1. Marketing survey based on Sapporo population movement data 2. Resident opinion survey 3. Interviews with users in each segment covered

902

S. Hayakawa et al.

4. Identification of requirements for Sapporo’s New Resident Electronic Registration Service 5. Drafting of Sapporo New Resident Electronic Registration Service concepts Using the information gained through the first three steps of the study, up to the interviews, we identified the key move-in events for each segment, which led to the definition of the following user personas: 1. 2. 3. 4. 5.

Hyper-busy, single businessman Newlyweds who have just purchased a new home Family with children transferring in Senior citizen retiring after recovery from a serious illness Foreign national coming to Sapporo for the first time

Participants in the workshop (see next section) learned evaluation methods based on these personas (except for v) and evaluated usability on that basis. Evaluations were made from individual perspectives because evaluation criteria will differ from persona to persona. For example, the “hyper-busy, single businessman” (Figure 2) would need to check the following information when moving in: how to transfer his resident card, moving-out/moving-in procedures, transfers of electricity/water/gas accounts, matters specific to the location (where to purchase kerosene in Sapporo, for example), transfers of telephone and Internet contracts, availability of parking, and information about the new neighborhood. Procedural requirements would include: deciding on housing, selecting and contracting with a moving company, and filing moving-out forms with current municipality. These procedures require additional, peripheral information that goes beyond the local government itself. Usability was evaluated by pretending to be the persona and experiencing firsthand whether information could be accessed and procedures taken via the web site.

Fig. 4. Abbreviated checklist

Fig. 5. Usability checklist

Report on Project to Identify Excellent Local Government Web Sites

903

2.3 Evaluator Training Workshop This project was coordinated with an educational project designed to communicate the importance of usability evaluations. Evaluators were therefore drawn from the public at large. Applicants participated in a group workshop where they learned the basics of evaluating web site usability. Those demonstrating an understanding of the evaluation criteria were qualified3 as initial evaluators and asked to evaluate local-government web sites. The workshop began by confirming the nature of web sites and how they function as media to provide participants with an overview of the basic aspects of web-site quality. Specific teaching points included: “Information Architecture” (the ease with which the information embodied in web-site documents can be understood), “Interactivity” (the ease with which screen movement and software functions can be understood) and “Accessibility” (the ease with which the site can be used by the disabled, the elderly and people operating in different environments). The workshop also confirmed techniques to be used in evaluating usability, trends in web-site usability evaluations, and HCD methodologies that can be practiced by web-site administrators to achieve usability. The persona methodology has the advantage of enabling issues and problems from the user’s perspective to be identified at the early stages of web-site creation. In addition to these core effects, the persona method is also an effective technique for evaluating usability from a variety of perspectives and represents an easy tool that web site operators can use to incorporate user perspectives into their design.

3 Evaluation Methodology 3.1 Selection and Initial Evaluation of Sites (October 4 through December 15, 2006) HCD experts selected 100 local-government web sites that demonstrated care in site design from among the 305 web sites operated by municipalities with populations of at least 120,000. Members of the HCD-Net Standardization and Authorization Department’s Steering Committee then used an abbreviated checklist (Figure 4) to identify 30 sites that contain move-in/move-out procedure information and achieved a certain level of ease of use. Thirty evaluators were selected from among participants in the “Local-Government Web Site Usability Evaluator Seminar” held by the HCD-Net Education Department (October 4, 17) and were asked to grade sites based on a usability checklist (Figure 5). The 10 sites with the best general scores4 for each persona were then identified. 3

Participants in the workshop who qualified as initial evaluators received a “Certificate of Completion of the Usability Evaluation Project for Local-Government Web Site Move-In/Move-Out Task Services” from HCD-Net. HCD-Net plans to give this certificate a degree of standing as it develops its usability certification program. 4 Scores were tabulated by standardizing evaluation criteria. The HCD-Net Standardization and Authorization Department plans to provide individual explanations and consulting for local governments who desire detailed information on results.

904

S. Hayakawa et al.

3.2 Final Evaluation The HCD-Net Board of Directors and Standardization and Authentication Department Steering Committee members performed final evaluations in which seven excellent sites were selected. The methodology for the final evaluations was to verify the findings of the initial evaluation. Expert reviews5 were performed for the Top 10 sites identified in the initial valuation, the Top 3 were approved and four excellent sites were selected, primarily from among the sites with excellent results for individual personas.

4 Evaluation Findings [No. 1 overall: Shinagawa-Ku, Tokyo] (Figure 6) http://www.city.shinagawa.tokyo.jp/ The site has consistent visual design and functionality and provides one-stop access to information. The life-stage menu is constantly displayed, which assists understanding. However, there was some confusion because “move-in/move-out” was included in the top menu under both the function-based “procedures” entrance and the objective-based “life stage” menu. From the perspective of senior citizens, font sizes can be changed on screen and the icons (entrances) for “move-in/move-out” and “senior citizens” were located close together so that just these two needed to be checked in order to ascertain whether most information was available. From the businessman’s perspective, the simple structure made the entrance points easy to grasp. Unfortunately, while the site did contain information about “mini” class cars and motorcycles, there was no information about ordinary-sized vehicles (proof of parking requirements etc.).

Fig. 6. Top Page for Shinagawa-Ku, Tokyo 5

Teams of at least five evaluators checked sites. Evaluators were chosen for their practical experience in user-centered development and usability evaluations and for their expertise in cognitive psychology etc.

Report on Project to Identify Excellent Local Government Web Sites

905

5 Future Enhancements This project focused on the “move-in/move-out” procedures that are used by large numbers of people. “Move-in/move-out” procedures contain a wide range of information and are one of the leading services provided by local governments. Given the importance of assessing local-government services on whether they can be used as expected, HCD-Net will continue to evaluate the usability of the “move-in/move-out” procedures on web sites. One task for the future is to improve the precision of initial screening. To do this, we are studying the potential for creating task scenarios and performing evaluations in line with these scenarios. We would also like to include the processes by which web sites are created in our evaluation criteria. Our goal is to provide convincing usability evaluations that will help to improve the quality of public services. The objective of this project was to communicate the importance of human-centered design processes to web system developers. We therefore conducted interviews regarding development processes with the local governments that produced the sites noted for excellence above. We found a very interesting relationship between local-government efforts and the HCD process. We will provide a summary of the project by reporting on the efforts of individual local governments during the session.

References 1. Shimbun, N.: Morning Edition, p. 1. (November 8, 2005) 2. Mitsubishi Research Institute, Inc. (2007) (Japanese), (accessed 2007-03-08). (online), available from http://www.mri.co.jp/REGION/ ?

A Survey of Factors Influencing People’s Perception of Information Security Ding-Long Huang, Pei-Luen Patrick Rau, and Gavriel Salvendy Department of Industrial Engineering Tsinghua University, Beijing 100084, China [email protected]

Abstract. Information security is a great concern to computer users, which is not only a technical problem, but also related to human factors. The objective of this study is to investigate the factors that can influence people’s perception of different threats to information security. In the survey study, 602 respondents were asked to evaluate one of 21 common threats to information security with regard to its position on each of the 20 threat-related items. An exploratory factor analysis was then conducted, and a six-factor structure modeling people’s perception of different threats to information security was derived. The relations between the factors and the perceived overall danger of threats were also tested by multiple regression analyses. Keywords: Perception; Information Security; Survey; Factor Analysis.

1 Objective and Significance Information security (InfoSec) is a great concern to computer and Internet users, who are suffering from a variety of threats to information security. Every day there are millions of accidents caused by virus, hackers, spam, spyware, zombie networks and many other threats to information security [1]. Those incidents are having serious effects on the economy and society [2], by bringing about infringements of human rights, financial damage to corporations, and the failure of the entire information system. Moreover, those threats to information security can influence IT users’ behavior and cognition. Information security involves both technology and people, and it is becoming increasingly evident that “the human factor is the Achilles heel of information security” [3]. Numerous sophisticated security methods have been developed, but the situation of information security is getting worse and worse [4, 5]. No matter how well designed, security methods rely on individuals to implement and use them. These methods may not accomplish their intended objectives if they are not used properly [6]. Moreover, Many people hesitate or refuse to adopt IT appliances because of worrying about the security problems [7], and that whether people willing to adopt an IT appliance depends not only on its “real security level”, but also on its “perceived security”. J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 906–915, 2007. © Springer-Verlag Berlin Heidelberg 2007

A Survey of Factors Influencing People’s Perception of Information Security

907

It has been realized that information security is not just a technology problem [8], and in recent years many researches on human factors in information security have been carried out. However, few researches have been made to study people’s perception of information security. Perception is a main part of human intelligence and a key component to understand human behavior [9]. It is “the mechanism with which a person evaluates inputs from the external environment, which, in turn, determines his/her behavioral response”[10]. IT users respond to different kinds of threats according to their perceptions of information security. Overestimates of risk can stifle people’s will to adopt IT applications, while underestimate of risk can wrongly encourage people to take insecure behaviors. Therefore, what they perceive, why they perceive it that way, and how they will subsequently behave is a matter of great import to the study of human factors in information security. This research seek to understand how to better model people’s perception of information security and what are the factors that can influence people’s perception of information security.

2 Background Literature 2.1 Information Security and Threats Nowadays, people seldom question the benefits of using computers and Internet for communication and doing business. However, the problem of information security is becoming more and more an essential issue. For common computer and Internet users, information security may means “being able to work with computers without being attacked by virus”, “being able to conduct on-line business without worrying that the credit card numbers will be stolen”, “being able to read e-mails without receiving Spam”, or “being able to talk with friends through instant messaging software without worrying that the information will be wiretapped”. According to an official definition, information Security is the protection of information and the systems and hardware that use, store and transmit that information [11]. From the technical view, information security is to protect the availability, accuracy, authenticity, confidentiality, integrity, utility and possession of the information [12]. We are suffering from a variety of threats to information security. A threat to information security can be seen as everything indicating a danger of information being attacked, destroyed or modified [13]. The 2006 Computer Security Institute/Federal Bureau of Investigation (CSI/FBI) survey on computer crime and security survey [5] found that 72% of respondents (primarily corporations and government agencies) detected computer security incidents within the last 12 month. The top four types of attack were virus, laptop/mobile theft, insider abuse of net access and unauthorized access to information, which accounted for more than 74% of financial losses. The number of threats to information security is almost uncountable and is on the increase. To better understand the numerous threats to information security, a model

908

D.-L. Huang, P.-L.P. Rau, and G. Salvendy

categorizing the threats into 12 general categories has been developed [12, 14]. A list of 21 common threats to information security and their categories in terms of this model was shown in Table 1. Table 1. Common threats to information security Categories 1. Acts of human error or failure 2. Compromises to intellectual property 3. Deliberate acts of espionage or trespass

4. 5. 6.

Deliberate acts of information extortion Deliberate acts of sabotage or vandalism Deliberate acts of theft

7.

Deliberate software attacks

8.

Forces of nature

9.

Deviations in quality of service

10. 11. 12.

Technical hardware failures or errors Technical software failures or errors Technological obsolescence

Threats Operation accidents Piratical software Hacker Passwords attack Information wiretapping Users’ online behaviors being recorded Data extortion Denial-of-service(Dos) Computer theft Phishing Virus Worms Trojan horse Backdoor Zombie PC SPAM Nature disaster (such as fire, earthquake, and lightning) Deviation in quality of service from service providers Hardware failure Scampish software Software bugs

2.2 Perception of Hazards and Perception of InfoSec Perception of hazard has been a focus of interest of researchers for some decades. How do people perceive different hazards? What factors can influence people’s risk acceptance? Many researches on risk perception have been carried out, dating back to Starr’ 1969 paper which showed that risk acceptance was related not only to technical estimates of risk and benefits but also to a subjective dimension such as voluntariness [15]. Starr’s work gave rise to much interest in the question of how people perceive risks. The psychometric paradigm was developed and became the most influential model in the field of risk analysis [16-18]. The aim of the psychometric paradigm is to unveil the factors that determine risk perception [19]. An important paper published in 1978 by Fischhoff et al. [16] indicated that people’s perception of different hazards can be influenced by certain factors. This paper compiled 9 items from the literature, including voluntariness of risk, immediacy of effect, known to exposed, known to science, control over risk, newness, chronic-catastrophic, common-dread and severity of consequences. The subjects were asked to rate a number of hazards on each of the items. The resultant Items×Hazards matrix was factor analyzed, and a 2 factor

A Survey of Factors Influencing People’s Perception of Information Security

909

(labeled as ‘technological risk’ and ‘severity’) solution was achieved. A following study [17] was extended to a boarder set of 18 risk-related items, and found a 3 factors (labeled as ‘dread’, ‘familiarity’ and ‘number of people exposed’) model. Many studies on risk perception utilizing psychometric paradigm was then carried out. More and more items and their influence on risk perception have been investigated. According to Covello [20-22], psychological researches have identified 47 known items that influence people’s perception of risk. The study of risk perception have been implied in many areas, such as nuclear engineering [23, 24], epidemics [25], automobile safety [26] and construction safety [27]. In the field of information technology, some efforts have also been made. Jackson et al. [28] reviewed the social science literature on the public perception of risk and extended their discussion to perceptions of crime in cyberspace. Vyskoc and Fibikova [29] conducted a survey about how IT users’ perceive information security. The results shown that users not always thought in the same way as the security specialists do. Yenisey, Ozok and Salvendy [30] investigated users’ feeling of security in e-commerce and developed guidelines for perceived security in e-commerce.

3 Method This research firstly derived 20 items from literatures review and investigated their influence on people’s perception of information security. Table 2. Survey derived items that may influence users’ perception of information security Items 1. Familiarity 2. Severity of consequences 3. Voluntariness 4. Catastrophic potential 5. 6.

Understanding Personal exposure

7. 8. 9. 10. 11. 12.

Observability Ease of reduction Preventive control Control of severity Immediacy of effect Known to those exposed

13. Newness 14. Media attention 15. 16. 17. 18. 19. 20.

Accident history Reversibility Duration of impacts Scope of impacts Predictability Possibility

Description in the survey I am familiar with this threat. The consequences of this threat are serious. I am exposed to the risk of this threat voluntarily. The outcomes of this threat are grouped in time and space (instead of being scattered in time and space). I understand this threat. This threat harm personal benefits (instead of public benefits) This threat is observable. The effect of this threat can be reduced easily. This threat can be prevented The severity of the effect of this threat can be controlled The effects of this threat are shown immediately. The existence of this threat can be known to those exposed to it. This threat is new or novel one. This threat is paid much attention by media (such as newspaper, TV and Websites). I have never been exposed to this threat. The effect of this threat can be reversed. The duration of impacts of this threat is long. The scope of impacts of this threat is wide. This threat can be predicted. The possibility that I will be exposed to this threat is large.

910

D.-L. Huang, P.-L.P. Rau, and G. Salvendy

These 20 items and their descriptions in the survey were listed in Table 2. An interview was then administered to 20 university students (who had plenty experience of using computer and Internet) with an in-depth, open-ended questionnaire. Their opinions about the possible issues that can influence perception of information security were elicited. Their answers validated the 20 items list in Table 2. A questionnaire was then constructed and a survey was conducted. Each respondent in this survey was asked to evaluate one of the threats listed in Table 1. The threat was selected from Table 1 randomly, and a brief introduction to this threat would be given. The respondents evaluated the threat with regard to (a) its position on each of the 20 items list in Table 2; (b) its perceived ‘overall risky level’. The sevenpoint Likert scale anchored with opposing adjective phrases was used. Demographic information including respondents’ age, gender, education level, occupation, experience with computer and Internet, knowledge on computer and security methods they had adopted were collected. The target population of this survey was people who had experience of using computer and Internet. The survey was posted on the website of IT Usability Laboratory in Tsinghua University, China, which could be accessed by all Internet users. Number of respondents participating in the survey study was 646. After deletion of invalid data, the total sample size was 602. There were 381 males (63%) and 221 females (37%) ranging in age from 17 to 51 (mean=24.3, SD=5.17).

4 Results 4.1 Overview The mean and standard deviation of the score for each of the 20 items arranged from lowest score to the highest score were listed in Table 3. The ‘severity of consequence’ item had the highest mean score. It indicated that most respondents considered those threats to information security as having serious consequences. The ‘personal exposure’ item had the lowest mean score. It showed that the respondents consider that those threats harm not only their personal benefits, but also public benefits. Table 3. Descriptive statistic of survey responses (N=602) Items Personal Exposure Newness Catastrophic Potential Voluntariness Understanding Reversibility Immediacy of Effect Control of Severity Media attention Known to Those Exposed

Mean 2.7 3.1 3.4 3.5 3.6 3.7 3.7 3.9 3.9 3.9

SD 1.60 1.52 1.55 1.61 1.76 1.44 1.55 1.68 1.62 1.62

Items Familiarity Predictability Duration of Impacts Possibility Observability Preventive Control Accident History Scope of Impacts Ease of Reduction Severity of Consequence

Mean 4.0 4.2 4.4 4.5 4.5 4.6 4.8 5.0 5.1 5.5

SD 1.76 1.45 1.41 1.49 1.56 1.67 1.80 1.36 1.51 1.38

A Survey of Factors Influencing People’s Perception of Information Security

911

4.2 Factor Analysis To explore the latent structure of the 20 items and identify the factors that can influence people’s perception of information security, an exploratory factor analysis was conducted on the raw data. Before performing a factor analysis, the KaiserMeyer-Olkin Measure of sampling adequacy (KMO) test was employed to check whether the 20 items are suitable to run factor analysis. The KMO test result (value=0.641) is substantially higher than the acceptable exploratory research norm of 0.5 established by Nunnally [31]. Six factors were extracted using a “screen test”; all of them evolved with the eigenvalues that were greater than 1 and account for 51.6 percent of the total variance. The factors were rotated by EQUAMAX methos. The results of factor analysis were shown in Table 4, with the highest entry in each row shadowed. Table 4. Factor analysis of items influencing perception of InfoSec Items Familiarity Understanding Control of Severity Newness* Duration of Impacts Scope of Impacts Media attention Personal Exposure Voluntariness Severity of Consequence* Preventive Control Observability Ease of Reduction Reversibility Predictability Catastrophic Potential* Accident History Possibility Immediacy of Effect Known to Those Exposed Variance explained by each factor % Variance explained by each factor

1 0.719 0.730 0.546 0.430 -0.031 0.016 0.043 -0.083 0.181 0.008 0.149 0.088 0.157 -0.128 0.037 -0.005 0.088 -0.038 -0.012 0.243

2 0.045 0.033 -0.025 -0.091 0.790 0.794 0.440 -0.008 0.052 -0.152 -0.167 0.018 -0.001 0.103 0.355 -0.144 -0.089 0.251 -0.086 -0.003

Factors 3 4 0.016 -0.093 0.004 0.056 0.305 0.314 -0.349 0.039 -0.105 -0.104 -0.215 0.053 0.243 0.096 0.591 -0.178 0.608 0.137 0.755 0.04 0.546 -0.176 0.510 0.005 0.629 0.001 0.485 0.265 0.548 0.104 0.422 -0.166 -0.094 -0.034 0.032 0.061 -0.036 -0.014 0.093 0.140

5 0.261 -0.173 -0.027 0.214 -0.023 0.132 0.060 0.085 -0.054 -0.093 -0.278 0.123 0.059 -0.172 0.085 0.097 0.816 0.724 -0.076 0.044

6 0.108 0.091 0.150 -0.291 0.002 0.046 -0.073 0.232 -0.092 0.048 -0.185 0.441 0.103 0.073 -0.008 -0.404 0.023 -0.072 0.787 0.651

Communality estimates 0.607 0.575 0.514 0.447 0.647 0.700 0.273 0.449 0.436 0.606 0.491 0.478 0.435 0.367 0.446 0.398 0.692 0.598 0.634 0.514

1.716

1.750

1.781

1.878

1.525

1.656

10.307

8.6

8.7

8.9

9.4

7.6

8.3

51.6

* Scores of the items marked with star have been reversed.

As seen in Table 4, the result of factor analysis clearly clustered the related items together. Factor 1 contains four items: ‘familiarity’, ‘understanding’, ‘control of severity’ and ‘newness’. These items are related to people’s knowledge of the threats to information security. People usually know little about the new threats, and they also feel unfamiliar with them, hard to understand them and hard to control the severity of them. Therefore this factor is named as ‘knowledge’ Factor 2 comprises items of ‘duration of impacts’, ‘scope of impacts’ and ‘media attention’. These items are related to the impacts of the threats. People are concerned

912

D.-L. Huang, P.-L.P. Rau, and G. Salvendy

about how long the duration is, how wide the scope is, and how much media attention are paid to the threats. Thus, this factor is named as ‘impact’. Factor 3 contains items of ‘personal exposure’, ‘voluntariness’ and ‘severity of consequence’. This factor reflects how people perceive the severity of the threats to information security. High scores on this factor were associated with threats which can bring serious consequences, harm public benefits, and make people exposed involuntarily. It seems appropriate to label this factor as ‘severity’ Factor 4 includes 6 items: ‘preventive control’, ‘observability’, ‘ease of reduction’, ‘reversibility’, ‘predictability’ and ‘catastrophic potential’. This factor indicates that to what extent people can control the threats. Whether the threats can be prevented, observed, reversed and predicted, whether their effect can be reduced, and whether the outcomes of them are scattered in time and space. Thus, factor 4 is labeled as ‘controllability’. Factor 5 contains items of ‘accident history’ and ‘possibility’. People may perceive relatively high possibility of being exposed to a threat if they have already been exposed to it in the past. So this factor is named as ‘possibility’. Factor 6 includes 2 items: ‘immediacy of effect’ and ‘Known to those exposed’. These items are related to the awareness of the threats. Those threats whose effects are shown immediately are usually easier to be known to those exposed to them. Therefore, this factor is labeled as ‘awareness’. The reliabilities of these factors were assessed by Cronbach’s alpha coefficient and the results showed that they were almost all over 0.5 and satisfied the exploratory research norm set by Nunnally[31], except one factor, ‘severity’, whose Cronbach’s alpha value was 0.484. The ‘severity’ factor contains 3 items: ‘personal exposure’, ‘voluntariness’ and ‘severity of consequence’. Looking into the raw data, a trend could be found that those threats perceived as harming public benefits were usually also perceived as making people exposed involuntarily and having serious consequences. However, the situation is special for the threat of ‘computer theft’, which was perceived as harming personal benefits, but still making people exposed involuntarily and having serious consequences. The special situation of ‘computer theft’ is understandable considering that it is the only “physical” threats within these 21 threats. So the Cronbach’s alpha of factor ‘severity’ was re-computed using the data of the other 20 threats, and the result was well exceed the 0.5 exploratory research norm set by Nunnally [31]. 4.3 Multiple Regressions Using the ‘overall dangerous level’ as the dependent variable, and the six factors extracted from factor analysis as the independent variables (factor scores for each case were computed with factor loading matrix), a stepwise multiple regression analysis was performed to investigate which factor best predict the overall perceived danger of threats to information security. The results were shown in Table 5. The factors of ‘Severity’, ‘Impact’, ‘Possibility’ and ‘Knowledge’ were found to have significant effect, which together accounted for 17.7% of the respondents’ overall perceived danger.

A Survey of Factors Influencing People’s Perception of Information Security

913

Table 5. Prediction of overall perceived danger: stepwise selection of factors Factors Severity Impact Possibility Knowledge

Coefficients 0.269 0.238 0.210 -0.084

F 46.764 44.394 41.735 32.790

p G(12%, 75%)

And then we broadcast candidate datasets to other nodes and get global support counts. Table 6 records the data exchanges of nodes. Table 6. The data exchanges of nodes local fre-candidate itemset AB

source

X.sup1

X.sup2

X.sup3

S1

5

4

4

BC

S1,S2

10

10

2

CD

S2

4

8

4

EF

S3

4

3

8

GH

S3

4

4

6

After this iteration we can find that only BC is global large in node S1, because BC.sup=(10+10+2)=22>10%*150=15, while AB.sup=(5+4+4)=13 C(14.6%, 88%) ． C => D(10.6%, 75%) ． E => F(10%, 75%)

1188

S. Xue and Z. Lu

But we also find in table 2,3, and 4 that the similarity between S1 and S2 is high while the similarity between S1 and S3 and the similarity between S2 and S3 are relatively low. The formula above also validates the conclusion. So we divide the three nodes into two groups: one is S1 and S2 and the other is S3, and then we get the association rules based on similarity by above FDM as following table 9. Table 9. New global association rules

1 2 3

． B => C(20%, 100%) ． C => D(12%, 75%) ． E => F(12%, 75%)

Compared with table 8 and table 9, we can find that rules have been increased on some extent and it certainly will provide us with a new method to mine exact information in every field.

5 Conclusion The data mining technique develops rapidly along with the popularization of large database and expanding of dataset. And association rules’ mining is one of most important part of data mining. With the development of distributed system and application of distributed database, association rules’ mining under distributed environment is becoming an important part in current application fields. Aiming at the heterogeneous database under distributed environment, in this paper we put forward a model for distributed association rules based on similarity as a new data-mining model. This model thinks enough the influence of support by heterogeneous datasets; at first we integrate the datasets based on their similarity, and then get the final rules by association rules mining.

Reference [1] Kargupta, H., Park, B., Hershbereger, D., Johnson, E.: Collective data mining: A new perspective toward distributed data mining. In: Kargupta, H., Chan, P. (eds.) Advanced in distributed data mining, pp. 133–184. AAAI / M1T (2000) [2] Yamanishi, K.: Distributed cooperative Bayesian learning strategies. Proceeding of COLT 97, pp. 250–262. ACM, New York (1997) [3] Ran, T., Ruihua, K., Ling, L.: The Design for Distributed Database of Logistic Information System. Journal of Wuhan University of Technology 28(8), 38–41 (2006) [4] Cheung, D.W., Ng, V.T., Fu, A.W., Fu, Y.J.: Efficient mining of association ru1es in distributed databases. IEEE Trans. On Knowledge and Data Engineering 8, 911–922 (1996) [5] Li, T., et al.: A new distributed data mining model based on similarity. In: Proceedings of the 2003 ACM symposium on Applied computing (2003) [6] Ming, Z.: Data mining. University of science and technology of china press (2002)

Research of Distributed Data Mining Association Rules Model Based on Similarity

1189

[7] Li, Z.: Research and realization of mining association rules in distributed system. Master paper of Da Lian Jiaotong University (2005) [8] Yue, W.: The method research of mining association rules in distributed environment. PHD paper of Chong Qing University (2003) [9] Hui-Chang, Z., Chang-Xia, M.: Two improved algorithms of association rules [J]. Computer application and software. vol.23(8) (August 2006)

Applications of Virtual Reality in Product Design Evaluation Jilin Ye1, Saurin Badiyani1, Vinesh Raja1, and Thomas Schlegel2 1

Warwick Manufacturing Group, International Manufacturing Centre, The University of Warwick, Coventry, CV4 7AL, United Kingdom {J.Ye, S.Badiyani, Vinesh.Raja}@warwick.ac.uk 2 Fraunhofer Institute for Industrial Engineering (IAO), Nobelstraße 12, D-70569 Stuttgart, Germany [email protected]

Abstract. Product evaluation throughout various stages of a design process is crucial to the final product's success, and this may be a costly, time consuming and logistically complex process. Currently there are limited product design evaluation tools available to provide better support to the whole product design and development process. Virtual reality (VR) has matured to become useful technology to support efficient and effective product design and development applications. This paper reports an investigation on applying VR technologies to computer aided product evaluation. The research work has concentrated on: i) investigating the potential of emerging VR based technologies such as three dimensional (3D) haptic interaction and 3D stereoscopic viewing, ii) integrating and implementing these VR based technologies into a computer aided product evaluation application and iii) exploring the efficiency and effectiveness of these VR based technologies in comparison with traditional techniques used during the product design evaluation process. Keywords: Product evaluation, Virtual reality technologies, Human computer interaction, Product design and development, Evaluation application.

1 Introduction Virtual reality (VR) is not a new invention. Scientific research has been undertaken in the field of VR for several decades, having recognised it as a very powerful tool for creating more natural and intuitive human computer interfaces. VR can be described as an interactive, computer-generated three dimensional (3D) environment with which users can interact using specialised peripherals such as electrical data gloves and haptic force feedback devices. As the VR technologies evolve, its applications become literally unlimited. It is assumed that VR technologies will reshape the interaction interfaces between user and computer technology by offering new approaches for the communication of information, the visualisation of processes and the creative expression of ideas. Product evaluation throughout various stages of a design process is often crucial to the final product's success, and this may be a costly, time consuming and logistically J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 1190–1199, 2007. © Springer-Verlag Berlin Heidelberg 2007

Applications of Virtual Reality in Product Design Evaluation

1191

complex process. Currently there are limited product design evaluation tools available to provide better support to the whole product design and development process. Therefore, extensive research is needed to investigate the efficient and effective product design evaluation tools based on up-to-date advanced computer technologies to provide better support to the whole product design and development process. The aim of this paper is to report an investigation of applying VR technologies to computer aided product design and development process in the context of product evaluation using product examples of handheld highly interactive products, such as mobile phones and personal digital assistants. A new VR based product design evaluation system has been developed which employs several VR interaction technologies to provide better support to the product design evaluation process. The paper is organised as follows. Section 2 reviews related work concerning the importance of product design evaluation and traditional and VR based techniques used for product design evaluation. Section 3 presents requirements for product design evaluation. Section 4 describes the proposed VR based product design evaluation system and its integration and implementation issues. Section 5 presents the results from the user trial tests. Section 6 draws conclusions and outlines future work and suggestions for research.

2 Related Work 2.1 The Importance of Product Design Evaluation Global competition and evolving high technologies are setting new challenges for every company with the aim to develop a new product with the short lead time, better quality and lower costs. Effective and efficient product design evaluation is critical to the success of a new product, especially in the telecommunication industry, like mobile phone and personal digital assistant manufacturers. It has been recognised that the product design and development cycle in the mobile phone product industry is extremely short. Nowadays most mobile phone products are designed and produced within months and this cycle time is constantly decreasing [1]. There are numerous factors that affect the success of the mobile phone product, for example, new functions, price and so on. The customer preference is identified as one of the most important factors for this purpose according to the research reported in [2]. The user’s knowledge, experience and understanding are valuable resources to mobile phone designers to take into consideration. Finding out what the customer really needs has become the focal point for most companies in order to achieve their products’ success on the market. Product design evaluation techniques are important in the retrieval of users’ perceptions on products’ appearance, functionality and usability, both to identify new requirements and to evaluate the product concepts. The evaluation may occur at all phases of the design and development process to make sure the product manifests the customer’s point of view from the initial concept to its production in the factory. Normally, the product design evaluation process will ask a specific group of users for a particular activity or query. The initial evaluation of the data retrieved will then be sent to designers for further analysis and consideration.

1192

J. Ye et al.

In order to gain customer satisfaction, it is also necessary to be able to define, design and evaluate all the features of a new product prior to the development of any physical prototypes, because at this stage of the design process, modifications are still quite cheap, compared with changes to a physical prototype or, even worse, the final product. As shown in Figure 1, the cost of fixing any mistake or making any change to a design increases dramatically as a product’s development carries on [3].

Fig. 1. Cost of fixing errors during the product design

2.2 Product Design Evaluation Techniques For the purpose of this research, two main categories of product design evaluation technique have been classified based on their interactive abilities and evaluation functionalities involved in the product design evaluation process: conventional evaluation techniques and VR based evaluation techniques. Conventional Product Design Evaluation Techniques. Conventional product design evaluation techniques include physical product prototype creation, virtual product model generation and utilising advanced computer based analysis and simulation tools. Regarding physical model making, there are two general approaches to produce physical models and prototypes: the building up and the carving away approach [4]. When building up, a prototype is created manually out of planar pieces cut out from paper, cardboard or thin styrene plastic. These pieces are taped or glued at the edges to form a rough 3D shape. More recently, a model can also be built up using advanced automation techniques such as rapid prototyping and manufacturing (RP&M). When carving away, the designer usually starts with a solid material such as blue foam or clay from which parts are cut away or reshaped by hand using different tools. On the other hand, the designer can also fabricate a prototype by advanced automation

Applications of Virtual Reality in Product Design Evaluation

1193

approaches like computer numerical control (CNC) machining. At present, physical prototype creation is still a costly and time consuming process. In virtual product model creation, two dimensional (2D) hand sketch and drawing are the early methods to outline product concepts for evaluation purpose. Later on, sketching using 2D computer aided design (CAD) systems provides an alternative approach to conventional manual paper based 2D drawing approach. Although 2D CAD remains a useful design tool for virtual model generation, mainstream CAD has now evolved from simple 2D drawing to complex 3D geometric modelling. It is evident that CAD systems have great potential for product design and development. However, the use of such systems for product design evaluation is limited. One reason for this might be that although the modelling within CAD has become 3D, sketching and displaying its results has remained 2D. Therefore, it is argued that new computer based 3D viewing tools should be developed so as to provide a true 3D capability for both the design process and evaluation process. There are many advanced computer based analysis tools which are useful in product design evaluation process, such as finite element analysis (FEA) and computational fluid dynamics (CFD). However, all these tools have the same limitations as current CAD systems do. They only provide 2D graphic and text information to support the product design evaluation process. VR Based Product Design Evaluation Techniques. Virtual reality (VR) has matured now to become useful technology for a wide range of applications in automotive, military, construction, medicine, oil & gas, etc. In the product design and development related domains such as CAD and computer aided manufacturing (CAM), VR technologies can support efficient and effective product design and development applications from the initial stage of the design process (e.g. conceptual design) to the entire product lifecycle activities (e.g. product lifecycle management). It is expected that in the future a VR based product design and development environment could provide both an improved visualisation of product by allowing the designer to co-exist in the same space as the product model therefore gaining a better appreciation of product geometry and aesthetics and an improved interaction with product design in terms of more intuitive model manipulation and functional experimentation that the designer could effectively interact with the product model directly in 3D rather than using the conventional 2D interactions like mouse movement and keyboard input operations. VR technologies bring new concepts of interaction paradigm with computers and new examination approaches to the product design and development process. For example, 3D display facilities and haptic force feedback devices simulate the natural visual and tactile senses which give more interpretation power to experiment and manipulate with the product model during the product design and development process. Although the VR hardware and software tools are still under development, many industries have already employed VR technologies for their specific applications. For example, A leading automobile manufacturing company employed VR systems which targeted very specific tasks within the product design and development cycle to make sure that the cars they have designed and developed are the ones that customers will want to buy and drive with quite satisfaction [5].

1194

J. Ye et al.

Here product design evaluation process using VR technologies are normally referred as VR based product design evaluation techniques. These VR technologies such as 3D stereoscopic viewing and haptic force feedback interaction could provide better support abilities when combing with CAD modelling, analysis and simulation methods for the product design evaluation process.

3 Requirement Definition for Product Design Evaluation Recent improvements in computer technologies provide advanced tools in the field of product design evaluation. In order to provide more responsive techniques for the product design evaluation process, there is a need for high level, understandable and effective evaluation requirements from practical case studies and/or user trials with usability testing. These identified evaluation requirements should not only establish more effective and usable evaluation mechanisms, but also reveal the real needs and expectations from users when they perform design evaluation work using computer based systems. The authors conducted a research study on improving human computer interaction (HCI) for computer aided conceptual design [6]. They found that an intuitive interface between human and computer is one which requires little training and offers a working style similar to that used by users when interacting with environments and objects in their day-to-day life. In other words, users interact with elements of their task by looking, hearing, holding, feeling and manipulating using as many of their natural skills and experience as appropriate, or can reasonably be expected to apply to a task. Haptics or force feedback technology opens the door to a new level of interactivity between users and computers. Prior to haptics, users only have the chance to interact with the computer through vision and sound. The sense of touch has been conspicuously absent in traditional computer interfaces like the keyboard and mouse. The introduction of haptic technology could bring significant changes to the way users interact with information and communicate ideas, by permitting users to touch and manipulate virtual computer-generated objects in a way that creates a compelling sense of tactile realness. These HCIs identified above also have huge potential when they are applied to the product design evaluation process as they will bring new concepts and new senses to it compared to the traditional product evaluation techniques. It is possible to define the aim of a new product design evaluation system as follows: To integrate VR based technologies into a design evaluation process for maximising its interactivity and efficiency; to provide better support to product design evaluation.

4 Product Design Evaluation System: Integration and Implementation The overall framework of the proposed product design evaluation system can be seen in Figure 2. Several VR based technologies are employed to provide better evaluation functions during the product design and development process. Each evaluation function’s design and implementation is discussed in detail in the following sub-sections.

Applications of Virtual Reality in Product Design Evaluation

1195

Fig. 2. Framework of the product design evaluation system

4.1 3D Model Creation The product model creation is the first step to conduct the product evaluation performance. The 3D product model used in this research should provide necessary information for 3D stereoscopic display and force feedback haptic interaction. There are many data formats available for 3D model representation to meet the above requirements such as different CAD model data formats, Virtual Reality Modeling Language (VRML), eXtended 3D (X3D) and Universal 3D (U3D). To simplify this research work, a 3D mobile phone model in VRML data format downloaded from the internet website [7] was used as the product sample for this study. The obtained VRML model is then converted into X3D data format in order to apply evaluation functions of haptic interaction and 3D stereoscopic view efficiently. The proposed evaluation system employs the H3D Application Program Interface (API) [8] (an open source haptic API) to implement these functions. This virtual sample model was also used in the user trial tests. 4.2 3D Stereoscopic Display Design and Implementation Most CAD systems display their models only on 2D planar screens. In order to understand 3D product models, a series of 2D views are needed simultaneously to interpret both inside and outside structures of the model. This normally requires additional brain work to reconstruct the complex structures of the model from different view channels and thus leads to much confusion in the product design evaluation process. To overcome this obstacle, a 3D stereoscopic display evaluation function has been created in the proposed product evaluation system by using the CrystalEyes wireless liquid crystal display (LCD) glasses from StereoGraphics [9] (now a part of Real D company) to enable users to analyse and manipulate the product model as a 3D whole. The LCD eyewear was activated by an infrared emitter that connected to the user’s

1196

J. Ye et al.

computer. Compared with fully immersive head mounted display systems, which often caused uncomfortable intrusive viewing problems, the lightweight stereoscopic LCD glasses provided a comfortable way to view virtual models in a 3D environment. This also enhanced the information exchange between the user and the digital virtual model during the product evaluation process. 4.3 Haptic Interaction Design and Implementation Haptic interaction brings profound changes to the way users interact with digital models by feeling, touching and manipulating them in a way that creates a compelling sense of tactile realness. Haptic interaction allows users to investigate and understand the structure of the virtual model in a more intuitive way because it mimics the way people interact with the physical model. In the proposed product design evaluation system, the haptic interaction evaluation function was implemented by employing the SensAble PHANToM Desktop haptic device along with the OpenHaptics [10] software development toolkit. When integrating a haptic interaction function into a product evaluation application, the most important issue is to render the virtual model both graphically and using haptics. It is generally accepted that the update of a visually rendered object must be done with a frequency of about 30 Hz to avoid flickering [11]. However, the object rendered using haptics must be updated at approximately 1000 Hz so as to make the rendered forces appear realistic. This obviously requires very efficient implementations of the rendering algorithms. In the proposed product design evaluation system, the OpenHaptics API from SensAble is used to solve the difficulty involved in the implementation of the haptic interaction design. In addition, the OpenGL API is employed to carry out the stereoscopic graphics display. Hence, this co-location task is implemented by having two different APIs, one for haptic rendering and one for stereoscopic graphics rendering. Using different APIs brings a great challenge to the proposed evaluation system, but it also provides much potential to further extend other functions for the evaluation process in the future. Since the H3D API facilitates application development of haptic and co-located hapto-visual applications, the proposed VR based product evaluation system was developed based on the H3D software development platform which uses the open standards X3D and OpenGL as well as advanced haptic technology from SensAble.

5 Results from User Trial Tests In order to test whether the proposed product design evaluation system could give better support for product evaluation through its VR based functions, thirty user trial tests were conducted by the authors. Though the proposed product evaluation system has limited evaluation functions, the evaluation performance of the mobile phone product example gave a full demonstration of using the evaluation functions available currently in the proposed product evaluation system described in this paper. As the result of the user trial tests, two new VR based evaluation functions employed in the proposed system received positive responses. Figure 3 shows the evaluation values gained on a Likert scale of 1 to 5 (where 1 = most un-satisfied and 5

Applications of Virtual Reality in Product Design Evaluation

1197

= most satisfied) along with the comparison with the traditional evaluation methods such as 2D model display. The two VR based evaluation functions (3D display and haptic interaction) had received higher values than the traditional 2D model display. High value for the stereoscopic display indicated that this VR technology has become more mature whereas the haptic interaction still needed further development and improvement in product evaluation applications. The proposed evaluation system also received higher values for several properties like the system is easy to use, the virtual model is quite similar to the real one, the overall performance and the reliable level of the system.

Fig. 3. Results of the user trial tests

The results from the user trial tests also indicated that 3D stereoscopic view of the product is a valuable benefit and being able to touch the product is an important function during the design evaluation process. The weakness of the proposed system was also identified by the users including inconvenience to use the PHANToM device stylus and there is limited evaluation functions available (e.g. 3D auditory rendering function).

6 Conclusions and Future Work In this paper, a new VR based product design evaluation system has been presented that fully integrates haptic interaction and stereoscopic display evaluation functions into one practical application. As an approach to new methods for the product design evaluation application, the proposed product evaluation system can provide better support capabilities for product evaluation through its new evaluation function integration and implementation. In particular, the proposed product evaluation system

1198

J. Ye et al.

provides the abilities to work quickly and more naturally by displaying the vitual model in 3D space instead of 2D planes with haptic force feedback interaction. New product evaluation systems are being developed which allow users to use their existing skills and experience while working in the digital environment. The potential of such technologies to allow an intuitive and natural interaction with virtual models has increased the drive towards computer support for the whole design and development process. The authors believe that the continued exploration of new evaluation technologies and their integration into product design applications will result in the future evolution of the next generation of product design evaluation systems. Although the current implementation of the proposed system provides better support for product evaluation process with its new VR based evaluation functions, it still can be improved in several respects in the future: • As mentioned above, other useful evaluation functions such as the 3D auditory evaluation function and the realistic colour and texture rendering will be developed and added to the proposed system so as to enrich the evaluation facilities as much as possible. • Besides the haptic interaction introduced in this paper which is based on the point – contact mechanism, full hand haptic interaction technologies will be investigated in order to provide more realistic force feedback interaction during the product evaluation process. • Further research will consider the integration of these new evaluation functions into the whole product design and development process. Acknowledgments. The University of Warwick and Fraunhofer IAO are partners of the EU-funded FP6 Innovative Production Machines and Systems (I*PROMS) Network of Excellence. http://www.iproms.org.

References 1. Ward, A.E., Angus, A.S.: Electronic Product Design. Chapman & Hall, London (1996) 2. Chuang, M.C., Chang, C.C., Hsu, S.H.: Perceptual Factors Underlying User Preferences toward Product Form of Mobile Phones. Int. J. of Industrial Ergonomics. 27(4), 247–258 (2001) 3. Tangram Technology Ltd.: Managing Costs – Materials and Design–Part 3, Located (accessed 21, Jul 2006) ( 2004) at http://www.tangram.co.uk/GI-Costs-3.html 4. Gribnau, M.W.: Two-handed Interaction in Computer Supported 3D Conceptual Modelling. PhD Thesis. The Delft University of Technology, the Netherlands (1999) 5. Stewart, P., Buttolo, P.: Putting People Power into Virtual Reality. Mechanical Engineering Design. The American Society of Mechanical Engineers, Located (1999) at http://www.memagazine.org/supparch/medesign/putting/putting.html 6. Ye, J., Campbell, R.I.: A New Virtual Reality Based Conceptual Design System. In: Proceedings of the EVEN International Conference on Virtual Engineering Applications and Product Development, pp. 52–63. Dublin, Ireland (2003) 7. (Accessed 22, November 2005) http://www.vr168.com

Applications of Virtual Reality in Product Design Evaluation 8. 9. 10. 11.

1199

(Accessed 25, November 2005) http://www.h3d.org (Accessed 01, December 2005) http://www.stereographics.com (Accessed 01, December 2005) http://www.sensable.com Bordegoni, M., De Angelis, F.: The Role of Haptic Devices for an Efficient Integration of Design, Simulation and Analysis. In: Brunet, P., et, al. (eds.) CAD Tools and Algorithms for Product Design, pp. 151–162. Springer, Berlin, London (2000)

Performance Evaluation for Automatic Protection Switching in a CDMA-TDD Wireless Communication System Li Zhang, Hiroyuki Okamura, and Tadashi Dohi Department of Information Engineering, Graduate School of Engineering, Hiroshima University, 1–4–1 Kagamiyama, Higashi-Hiroshima 739–8527, Japan {okamu,dohi}@rel.hiroshima-u.ac.jp

Abstract. The CDMA (Code Division Multiple Access) is classified into two mainstream technologies; CDMA2000 and CDMA-TDD (Code Division Multiple Access with Time Division Duplex mode). Specifically, in China, the CDMA-TDD mode has been used as the standard of 3G mobile network system. In this paper, we investigate the performance of automatic protection switching (APS) of CDMA-TDD wireless communication system, where APS is the familiar technology to guarantee highly reliable communication networks such as an asynchronous transfer mode network. We develop a stochastic model based on a continuoustime Markov chain for the CDMA-TDD wireless communication system with APS, and derive some related quantitative performance measures in terms of Quality of Service (QoS). Throughout numerical experiments, we examine how the design of APS and network affects the performance measures quantitatively.

1

Introduction

In recent wireless communication, almost all the second generation (2G) mobile systems have been switched to the third generation (3G) ones in all over the world. In fact, the international telecommunications union has identified a series of 3G technologies called the international mobile telecommunications–2000 (IMT-2000) [2]. The IMT-2000 family mainly consists of five technology standards which belong to CDMA and TDMA (Time Division Multiple Access). Further, the CDMA is classified into two mainstream technologies; CDMA2000 and CDMA-TDD (Code Division Multiple Access with Time Division Duplex mode). Specifically, in China, the CDMA-TDD mode has been used as the standard of 3G mobile network system. The CDMA-TDD is referred to as TD-SCDMA (Time Division - Synchronous Code Division Multiple Access), and has been extensively developed by Siemens and the China Academy of Telecommunications Technology. The main advantage of CDMA-TDD is that it can be implemented with less cost than the comparable other 3G systems. Because the CDMA-TDD is much more efficient in spectrum J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 1200–1209, 2007. c Springer-Verlag Berlin Heidelberg 2007 

Performance Evaluation for APS in a CDMA-TDD

1201

and is compatible with GSM (Global System for Mobile Communications) network elements [5], it enables us to use 3G services without installation of new infrastructure. On the other hand, the automatic protection switching (APS) is a control mechanism which automatically switches the ongoing talks on failed channels to idle channels. It can be extended to the system with multiple base repeaters, which may switch automatically all the ongoing calls on the failed base repeater to the working base repeater immediately. The APS has been widely used in asynchronous transfer mode network to enhance the network integrity. Nowadays the APS has been applied for improving the system availability in every wireless communications system. Ma et al. [4] and Sun et al. [7] considered TDMA wireless system [1] with and without APS, and performed the availability/performance evaluation by Stochastic Petri Net Package [6]. They derived quantitatively the system availability, the new call blocking probability and the handoff call dropping probability and referred to the effects of APS. In fact, it is quite important to evaluate the availability and performance of wireless systems with and without APS considering both traffic and control channels, not just focusing on performance of handoff and channel assignment [3]. In this paper, we investigate the performance of APS of CDMA-TDD wireless communication system. We develop a stochastic model based on a continuoustime Markov chain for the CDMA-TDD wireless communication system with APS, and derive some related quantitative performance measures in terms of Quality of Service (QoS). Similar to [4], [7], we give the stochastic Petri net representations of CDMA-TDD systems with and without APS. Throughout numerical experiments, we examine the dependence of the base repeaters and the expected number of services at a failure on traffic intensity and channel allocation, and refer to how the design of APS and network affects the performance measures quantitatively.

2 2.1

Modeling of CDMA-TDD System Technological Principles of CDMA-TDD

The CDMA-TDD combines two leading technologies; an advanced TDMA/TDD and an adaptive CDMA which can be operative in synchronous mode. First, we outline the basic technological principles of TDMA, TDD and CDMA. The TDMA is a digital technology that divides each frequency channel into multiple time slots and enables several subscribers to use transmission channels at the same time. In the CDMA-TDD wireless communication system, the TDMA uses a 5ms frame sub-divided into 7 time slots, which can be flexibly assigned to either a single user or several users who may require multiple time slots. The TDD principle permits the traffic to be uplinked (from the mobile terminal to the base station) and downlinked (from the base station to the mobile terminal) using different time slots within the same frame. Also, it allows both the uplink

1202

L. Zhang, H. Okamura, and T. Dohi

and downlink on the same frequency band and does not require pair bands. In other words, the uplink and downlink are transmitted in the same frequency channel but at different times. More precisely, it is possible to change the duplex switching point and to move the capacity from uplink to downlink, and vice verse for the capacity requirement from downlink to uplink. For this reason, it is seen that the TDMA utilizes spectrum optimally and allows both the symmetric and asymmetric data services. For symmetric services used during telephone and video calls (multimedia applications), the same amount of data is transmitted into both directions, and the time slots are split equally between the downlink and uplink. For asymmetric services used with Internet access (download), on the other hand, the high data volumes are transmitted from the base station to the terminal, and the more time slots are used for the downlink than the uplink. The FDD (Frequency Division Duplex) scheme, which is employed by the wide-band CDMA and CDMA2000, uses a pair of frequency bands for uplink and downlink, so that portions of the spectrum are occupied with asymmetric loads, but not used for data transfer. These idle resources cannot be utilized for any other services, leading to an inefficient use of the spectrum. In addition to the TDMA-TDD principle, the CDMA-TDD uses the CDMA to increase the capacity of the radio interface more and more, that is, the CDMA increases the traffic intensity in each cell by the simultaneous multiple-user access on the same radio channel. 2.2

Model Description

Consider a CDMA-TDD wireless communication system, where the base transceiver system of each cell has two base repeaters (BRs); BR1 and BR2. Each BR provides two kinds of services to mobile clients; service A and service B. The arrival rates of service requests are assumed to be constant, and are defined by λA,1 , λB,1 at BR1 and λA,2 , λB,2 at BR2. BR1 and BR2 process service A and service B independently with respective rates μA,1 , μA,2 , μB,1 and μB,2 per request. Different services may require different numbers of uplink and downlink channels. In this paper, we suppose that the service A uses mA uplink and nA downlink channels, and that the service B does mB uplink and nB downlink channels in both BRs. If there are not a sufficient number of channels at the arrival of a service, the service is refused from BRs. Each BR has total N channels to use either uplink or downlink, and its allocation for uplink and downlink is initially fixed as Cu for uplink and Cd for downlink. BR1 and BR2 break down with constant failure rates; γ1 and γ2 , respectively. The failed BR undergoes recovery operation, where the recovery operation is completed with constant repair rate η1 or η2 for each BR. 2.3

Recovery Operation

This paper deals with four recovery policies for the CDMA-TDD system. The first two policies concern the communication services at a system failure as follows:

Performance Evaluation for APS in a CDMA-TDD

1203

Policy I-A (Cancel): All of the ongoing services at a failed BR are cancelled. Policy I-B (Delay): All of the ongoing services at a failed BR wait until the completion of recovery operation. In addition to the above policies, we further introduce the fail-over operation by APS. Though APS can switch automatically the ongoing services at the failed channels to that at the idle channels, its mechanism can be extended to the system with multiple base repeaters [7], and enables us to switch automatically all ongoing services at the failed base repeater to the working base repeater. According to the functionality of APS, it assumed that the busy channels at the failed BR are migrated to other working BRs. If there are fully idle channels in the other BR, the migrated services are immediately available. Otherwise, in the case where the shortage of channels occurs, the migrated services may wait until the completion of recovery or may be cancelled in accordance with Policy I-A or Policy I-B. This paper considers the following two rules for the migration: Policy II-A (Service A): The APS migrates service A with high priority. After the migration of service A, the ongoing service B is switched to the working BR. Policy II-B (Service B): The APS migrates service B with high priority. After the migration of service B, the ongoing service A is switched to the working BR.

3 3.1

Analysis CTMC Model

We consider the continuous-time Markov chain (CTMC) model to represent the CDMA-TDD model. Let Ai (t) and Bi (t) be the numbers of service A and service B in BRi , respectively. Also let Si (t) denote a state of BRi . If BRi fails at time t, Si (t) = 0. On the other hand, if BRi is available, Si (t) = 1. Then the system state including both BRs at time t is given by {A1 (t), B1 (t), S1 (t), A2 (t), B2 (t), S2 (t)}. Define the system state by (a1 , b1 , s1 , a2 , b2 , s2 ), which indicates the values of A1 (t) though S2 (t). Figure 1 illustrates a part of transition diagram of the CTMC. In this case, we divide the system state into four global states, where each global state corresponds to a condition of BRs, i.e., it means a combination of s1 and s2 . On the other hand, the local state in each global state represents a condition of services on BRs. For example, the local state where both BRs fail (Global (0,0)) has no transition because BRs do accept no service. Another transition rule among global states depends on recovery policies. Under Policy I-A without APS, the transitions for recovery become the arrows that point to a fixed local state. If we take place Policy I-B without APS, the transitions do not change a configuration of local state before the completion of recovery. In the case of the policies with APS, all of recovery transitions become more complex. As shown in Figure 1, the CTMC has a large number states. If we set Cu = 10, Cd = 10, mA = 1, nA = 2, mB = 2 and nB = 1, then the total number of

1204

L. Zhang, H. Okamura, and T. Dohi

Global (1,1)

Global (0,1)

(1,1,2,1)

(2,1,1,1)

(1,1,2,1)

(2,1,1,1)

(1,1,1,1)

(1,1,1,1)

(1,1,1,2)

(1,2,1,1)

(1,1,1,2)

(1,2,1,1)

Global (1,0)

Global (0,0)

(1,1,2,1)

(2,1,1,1)

(1,1,2,1)

(2,1,1,1)

(1,1,1,1)

(1,1,1,1)

(1,1,1,2)

(1,2,1,1)

(1,1,1,2)

(1,2,1,1)

Fig. 1. State transition diagram of CTMC

C1

C2 Tr1

Tr2 t3 R1

t1

D1

U1 m2

m1

m1

m2 m2-1 n1

n1 m1 #(Su1+Sd1+U1+D1) #(Su2+Sd2+U2+D2) To1

m2

Ts1

#U1

Ts2 n2

n2-1

n2

Ts1

To2

Ts2 #D2 m2

n2

m1 n1

m2 Su1

n2

n1-1

m1-1 m2

#D1

n1

m1 n1

n1 m1

To2

m2 m1

n2

n1-1 n2 n2-1

m1-1

D2

U2

n1

m2-1

To1

t4

R2

t2

#Sd1

n2

m1

#U2

Sd1

n1

n2 m2

Sd2

Su2 #Su2

#Su1 Tb1

Tb2

#Sd2

Place Description Place Description Ci channel allocation Ui uplink channels Di downlink channels Sui serving uplink channels Sdi serving downlink channels Ri recovery Fig. 2. SPN model for two BRs without APS

states in the CTMC is 2304, and its associated total number of global and local transitions is 34802. For such a large scale CTMC, Stochastic Petri net (SPN) provides a smaller representation than the underlying CTMC by reducing the

Performance Evaluation for APS in a CDMA-TDD C1

1205

C2 Tr1

Tr2 t3 R1

t1

D1

U1 m2 m2-1

To1

n1

n2

m2 #U1

n2-1

n2

n2-1

n2 m2

m2

m1 n1

#Sd1

n2

m1

#U2

Sd1

To2

Ts2 #D2 m2

n2 m2

Su1

n2

n1-1

m1-1 n1 m1 Ts1

#D1

n1

m1 n1

m1

#(Su2+Sd2+U2+D2) To1

To2

Ts2 m2

m1

m2-1

#(Su1+Sd1+U1+D1)

n1-1 n2

m1-1

Ts1

D2

U2

n1

m1 m1 n1

t4

R2

t2

n1

n2 m2

Sd2

Su2 #Su2

#Su1

#Su1

Tb1

Tb2

#Sd2

Bu

t5

t7 #Sd1 t6

Bd t8

Place Description Place Description Ci channel allocation Ui uplink channels Di downlink channels Sui serving uplink channels Sdi serving downlink channels Ri recovery Bu uplink channels buffered by APS Bd downlink channels buffered by APS Fig. 3. SPN model for two BRs with APS

number of states. Figures 2 and 3 summarize the SPN models for CDMA-TDD system with and without APS. Since the SPN models are essentially same as the CTMC models described above, we need to generate the transition behavior of CTMC models from the SPN models in evaluating the performance measures. 3.2

QoS Performance Measures

Of our concern is the derivation of the performance measures in steady state. Define the steady-state probability for each system state in the underlying CTMC: P (a1 , b1 , s1 , a2 , b2 , s2 ) = lim Pr{A1 (t) = a1 , B1 (t) = b1 , S1 (t) = s1 , A2 (t) = a2 , B2 (t) = b2 , S2 (t) = s2 , }. t→∞

(1) Then QoS measures are defined by: – The expected number of services at a failure (ENs,i ): ENs,i is the expected number of services s in BRi at a time point of server failure. Under Policy I-A, this turns out the expected number of lost services. In the case of Policy I-B, this means the expected number of delayed services by a failure. – Blocking probability (BPs,i ): BPs,i is the probability that an arriving service s is rejected from BRi due to the shortage of channels and a failure.

1206

L. Zhang, H. Okamura, and T. Dohi

The performance measures can be computed as   a1 P (a1 , b1 , 0, a2 , b2 , s2 )/ ENA,1 = a1 ,b1 ,a2 ,b2 ,s2

a1 ,b1 ,a2 ,b2 ,s2





ENB,1 =

b1 P (a1 , b1 , 0, a2 , b2 , s2 )/

a1 ,b1 ,a2 ,b2 ,s2

a1 ,b1 ,a2 ,b2 ,s2





ENA,2 =

a2 P (a1 , b1 , s1 , a2 , b2 , 0)/

a1 ,b1 ,s1 ,a2 ,b2



ENB,2 = 

b2 P (a1 , b1 , s1 , a2 , b2 , 0)/

BPs,2 =

(3) P (a1 , b1 , s1 , a2 , b2 , 0),

 



(4) P (a1 , b1 , s1 , a2 , b2 , 0),

a1 ,b1 ,s1 ,a2 ,b2

P (a1 , b1 , 1, a2 , b2 , s2 ) +

a1 ,b1 ∈Ls a2 ,b2 ,s2



(2) P (a1 , b1 , 0, a2 , b2 , s2 ),

a1 ,b1 ,s1 ,a2 ,b2

a1 ,b1 ,s1 ,a2 ,b2

BPs,1 =

P (a1 , b1 , 0, a2 , b2 , s2 ),



(5) P (a1 , b1 , 0, a2 , b2 , s2 ),

a1 ,b1 ,a2 ,b2 ,s2

P (a1 , b1 , s1 , a2 , b2 , 1) +

a1 ,b1 ,s1 a2 ,b2 ∈Ls



(6) P (a1 , b1 , s1 , a2 , b2 , 0),

a1 ,b1 ,s1 ,a2 ,b2

(7) where Ls is the set of all combinations of services A and B and corresponds to the limit number of channels.

4

Numerical Experiments

This section presents numerical experiments on the QoS measures. We assume that BRs are placed in the same operation environment, i.e. λs,1 = λs,2 , μs,1 = μs,2 , γ1 = γ2 and η1 = η2 . Note that QoS measures for BRs under consideration are exactly same in this setting. Service A occupies 1 uplink and 2 downlink channels, and Service B does 2 uplink and 1 downlink channels. Also, throughout numerical examples, the processing rate, the failure rate and the repair rate are given by μA,· = 1.0, μB,· = 0.1, γ· = 0.00001, η· = 0.0002. It is assumed that the expected number of services at a failure is relatively small under these parameters. In fact, there is no remarkable difference between Policy I-A and Policy I-B. Thus we focus on only Policy I-B in the latter discussion. Figures 4 and 5 show the blocking probabilities (BPs) and the expected numbers of services at a failure (ENs), respectively, for Service A and Service B with different arrival rate of Service A. Here the varying arrival rates of Service A are given in the range of λA,· = 0.1, . . . , 0.9, and the arrival rate of Service B is fixed as λB,· = 0.05. The traffic intensities for Service A and Service B are ρA,· = 0.1, . . . , 0.9 and ρB,· = 0.5, respectively. The initial allocations of uplink and downlink channels are Cu = 10 and Cd = 10. In Fig. 4 it can be seen that the blocking probabilities with APS take the same values as the ones without APS. Since our example here implicitly supposes that the services are not cancelled by a failure (Policy I-B) and that the abilities of serving of both BRs are same,

Performance Evaluation for APS in a CDMA-TDD 0.054

0.054 No failover (Service A) No failover (Service B)

0.053

0.053

0.052

0.052 Blocking probability

Blocking probability

1207

0.051

0.05

0.051

0.05

0.049

0.049

0.048

0.048

0.047

Policy II-A (Service A) Policy II-A (Service B) Policy II-B (Service A) Policy II-B (Service B)

0.047 0.1

0.2

0.3

0.4 0.5 0.6 Traffic intensity (Service A)

0.7

0.8

0.9

0.1

0.2

0.3

w/o APS

0.4 0.5 0.6 Traffic intensity (Service A)

0.7

0.8

0.9

0.7

0.8

0.9

APS

Fig. 4. Dependence of BPs on traffic intensity

0.9

0.09 No failover (Service A) No failover (Service B)

Policy II-A (Service A) Policy II-A (Service B) Policy II-B (Service A) Policy II-B (Service B)

0.08 The expected number of services at a failure

The expected number of services at a failure

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.07

0.06

0.05

0.04

0.03

0.02

0.01

0

0 0.1

0.2

0.3

0.4 0.5 0.6 Traffic intensity (Service A)

w/o APS

0.7

0.8

0.9

0.1

0.2

0.3

0.4 0.5 0.6 Traffic intensity (Service A)

APS

Fig. 5. Dependence of ENs on traffic intensity

the conservation law in the queueing networks may lead to this result that the blocking probabilities are exactly same. Also, as the traffic intensity of Service A increases, the blocking probability of Service B also increases. This is because Service A and Service B share uplink and downlink channels. When the traffic intensity of Service A is large, most of channels tend to be occupied by Service A. Thus the blocking probability of Service B becomes large. On the other hand, we can see the remarkable effects of APS on the expected number of services at a failure, as shown in Figure 5. In particular, the number of services at a failure for the high priority of service becomes smaller than the low priority of service. Next we investigate the QoS measures for varying channels. Figures 6 and 7 depict the blocking probability and the expected number of services at a failure in the cases where the number of uplink channels is given by 0, . . . , 20. The total number of channels are 20 in all the cases. Moreover, the traffic intensities of Service A and Service B are given by ρA,· = 0.9 and ρB,· = 0.5, respectively. For the other traffic intensities, we found the similar tendency on blocking probabilities and the expected number of services at a failure. The result in the case of ρA,· = 0.9 implies the largest difference between the results of Service A and Service B. Similar to the previous result, the blocking probabilities with APS equal those without APS. The expected number of services at a failure without

1208

L. Zhang, H. Okamura, and T. Dohi 1

1 0.9

0.8

0.8

0.7

0.7 Blocking probability

Blocking probability

No failover (Service A) No failover (Service B) 0.9

0.6 0.5 0.4

0.6 0.5 0.4

0.3

0.3

0.2

0.2

0.1

Policy II-A (Service A) Policy II-A (Service B) Policy II-B (Service A) Policy II-B (Service B)

0.1

0

0 0

5

10 The number of uplink channels

15

20

0

5

w/o APS

10 The number of uplink channels

15

20

15

20

APS

Fig. 6. Dependence of BPs on channel allocation

0.9

0.35 No failover (Service A) No failover (Service B)

Policy II-A (Service A) Policy II-A (Service B) Policy II-B (Service A) Policy II-B (Service B)

0.8 The expected number of services at a failure

The expected number of services at a failure

0.3 0.7

0.6

0.5

0.4

0.3

0.2

0.25

0.2

0.15

0.1

0.05 0.1

0

0 0

5

10 The number of uplink channels

w/o APS

15

20

0

5

10 The number of uplink channels

APS

Fig. 7. Dependence of ENs on channel allocation

APS takes the maximum value at the allocation where uplink and downlink channels are symmetric. In the symmetric allocation, each BR can accept more services than the extremely asymmetric cases. That is, the increase of the expected number of services at a failure without APS depends on the number of accepted services. On the other hand, the results with APS leads to the minimum value at the symmetric allocation. This means that it is important to make a large capacity for services with both BRs in the case with APS.

5

Conclusion

In this paper, we have described the CDMA-TDD wireless communication system with/without APS and analyzed it by means of CTMC and SPN models. The performance measures, like blocking probability and the expected number of services at a failure, are numerically derived to reflect the performance improvement caused by APS. Throughout numerical experiments, it has been shown that APS could increase the expected number of services at a failure, without any change of blocking probability.

Performance Evaluation for APS in a CDMA-TDD

1209

As a future work, more complex CDMA-TDD wireless communication systems with more kinds of services should be considered to examine the effect of APS. If the number of base repeaters is larger than two, the underlying CTMC will be more complex. Then, any effective numerical technique to analyze a large scale CTMC should be utilized for comparison of CDMA-TDD systems with/without APS.

References 1. Falconer, D.D., Adachi, F., Gudmundson, B.: Time division multiple access methods for wireless personal communications. IEEE Communications Magazine 1, 50–57 (1995) 2. Giannakakis, G., Todd, C.: Techno-economic evaluation of 3rd generation mobile systems, Working Paper at Department of Electronic and Electrical Engineering, Telecommunications for Industry (2003) 3. Hong, D., Rappaport, S.S.: Traffic model and performance analysis for cellular mobile radio telephone systems with prioritized and non-prioritized hand off procedures,”. IEEE Trans. on Vehicular Technology 35(3), 77–92 (1986) 4. Ma, Y., Han, J., Trivedi, K.S.: Composite performance and availability analysis of wireless communication networks. IEEE Trans. on Vehicular Technology 50(5), 1216–1223 (2001) 5. Padgett, J.E., Gunther, C.G., Hattori, T.: Overview of wireless personal communications. IEEE Communications Magazine 1, 28–41 (1995) 6. Sahner, R., Trivedi, K.S., Puliafito, A.: Performance and Reliability Analysis of Computer System: An Example Based Approach Using the SHARPE Software Package. Kluwer Academic Publishers, Dordrecht (1995) 7. Sun, H., Cao, Y., Han, J., Trivedi, K.S.: Availability and performance evaluation for automatic protection switching in TDMA wireless system. Proceedings of 1999 Pacific Rim International Symposium on Dependable Computing, pp. 15–22. IEEE Computer Society Press, Los Alamitos (1999)

A Routing Algorithm for Random Error Tolerance in Network-on-Chip* Lei Zhang1,2, Huawei Li1, and Xiaowei Li1 1

Key Laboratory of Computer System and Architecture, Institute of Computing Technology, Chinese Academy of Sciences, 100080 Beijing, China 2 Graduate University of Chinese Academy of Sciences, Beijing, 100080, China {zlei,lihuawei,lxw}@ict.ac.cn

Abstract. In DSM and nanometer technology, there will present more and more new fault types, which are difficult to predict and avoid. Applying fault tolerant algorithms to achieve reliable on-chip communication is one of the most important issues of Network-on-Chip (NoC). This paper reviews the main on-chip fault tolerant communication algorithms and then proposes a new routing algorithm with end-to-end feedback. The average transmission latency, power consumption and reliability are compared with other techniques. As experiments show, the proposed algorithm has lower latency, lower power consumption compared with those of others, and it can provide high reliability. Keywords: fault tolerance, reliability, Network-on-Chip, routing, error correcting, error detecting.

1 Introduction According to ITRS’s report [11], the feature size will shrink to 50nm with 4 billion transistors on a single chip, and the frequency can be up to 10GHz while the working voltage is around 1v. The more sophisticated semiconductor technology makes the more powerful and complex SoCs come true. There will be hundreds of, maybe even more, processing elements, storage elements, DSP cores, I/O modules etc. on a die. To reduce the time-to-market, SoC chips must reuse available components in a plug & play fashion. Therefore, the interconnection mechanism becomes a critical part in such a design methodology [6]. Traditional on-chip interconnection schemes include shared bus (CoreConnectTM, AMBA) and proprietary interconnection mechanisms. As those methods have limited scalability, they can not fulfill the requirements of future on-chip communication [7]. To deal with this problem, computation and communication should be separated. The computing resources are connected with interconnection fabric. The general interconnection structure is termed as Network-on-Chip (NoC) [1] [8] [9]. The client of the network can *

This paper is supported in part by National Natural Science Foundation of China (NSFC) under grant No.(60633060, 60606008, 60576031), and in part by National Basic Research Program of China (973) under grant No.(2005CB321604, 2005CB321605).

J. Jacko (Ed.): Human-Computer Interaction, Part IV, HCII 2007, LNCS 4553, pp. 1210–1219, 2007. © Springer-Verlag Berlin Heidelberg 2007

A Routing Algorithm for Random Error Tolerance in Network-on-Chip

1211

be a processor, DSP, ASIC or memory. Each client has a network interface, through which the resource is connected to the network. The on-chip network is the backbone used for communication between various computing resources. The task of on-chip network is to provide pre-specified quality of service (QoS) for communication, including bandwidth, latency and reliability. NoC has adopted a variety of techniques from computer network and parallel computing [12] and also has many modifications and variations. As for the QoS provided by NoC, the reliability of transmission is one of the most important aspects. The fault tolerant mechanism is indispensable in DSM and nanometer technology. The factors which may introduce errors into on-chip network can be categorized into two classes [2]: one is for the crosstalk between long wires and the other for the cosmic rays, electronic-magnetic interference, variation of process etc. Those can increase the probability of transient errors and soft errors significantly which may be hardly avoided or eliminated. New algorithms that are conductive to NoC should be adapted in that traditional algorithms for fault tolerance are not suitable due to their unacceptable overhead in power or area. Several algorithms already exist are reviewed and a novel routing algorithm with end-to-end feedback to tolerate random errors is proposed. Experiment results and analysis of the algorithm is given. This paper is organized as follows. Section 2 gives some related works. In section 3, the new random error tolerance routing algorithm is proposed. Experimental results are shown in section 4. Finally conclusions are given.

2 Related Works There will be more and more fault types affecting on-chip interconnected wires as technology is getting improved [5]. However, it’s impossible or very expensive to avoid these faults completely at design stage. Thereby error recovery and fault tolerant mechanisms are needed. These mechanisms can provide high reliability with low test and verification cost. However, traditional fault tolerant algorithms can not be used on chip directly due to the constraint of resource, power consumption, latency and chip area [2]. Existing fault tolerant mechanisms fall into two categories: one is the request retransmission mechanism based on error checking and correcting code, the other is stochastic communication. We will give a brief introduction of these algorithms in this section. A generic NoC model is described firstly, which is the baseline used in this paper. 2.1 A Generic Network-On-Chip Model As depicted in Fig. 1, a die is divided into 16 symmetrical tiles. Computing resources in a system are mapped to each tile. Those tiles communicate with each other using on-chip network. Each tile has a network interface connected to a switch. Each switch is connected to adjacent neighbor switches [10]. Wormhole switching [12] is used in this model.

1212

L. Zhang, H. Li, and X. Li

00 tile 10 20 30

s s s s

01 11 21 31

s Switch

s s s s

02 12 22 32

s s s s

03 13 23 33

s

s

00

01

02

03

Source 10 (Producer)

11

12

13

20

21

22

23

30

31

32

33

s Destination (Consumer)

s

Network interface

Fig. 1. A 2D Mesh NoC [14]

Data path

Ack/nack path

Fig. 2. Examples of routing vectors

The format of the data packet can be customized for different applications. Here takes a generic format for ease of analysis [1]: Header Type

Size

Direction

Payload Routing

CRC

Data

Like the computer network, the communication protocol for on-chip network is organized as a layered stack [6]: physical, link, network, transport, and application layers from bottom to top. 2.2 Request Retransmission There are three types of retransmission: end-to-end, switch-to-switch and hybrid [4]. End-to-end retransmission can use parity or CRC code to check packets. Packet checking is only performed at destination end. Intermediate switches only forward packets and have little knowledge about whether the packet is destroyed. The source end should wait for ack or nack packets fed back by the destination and then decide whether to retransmit the packet or not. In switch-to-switch mechanism, on the other hand, each intermediate switch checks all the passing packets. Therefore decoders are needed in all switches. Hybrid mechanism makes use of end-to-end error correcting and in some degree has lower retransmission overhead. But error correcting logic is complex and consumes large chip area. 2.3 Stochastic Communication Retransmission needs a lot of resources, like logic and storage, and has long latency which is critical for on-chip communication. On-chip network has a large amount of interconnecting wires. To make use of these rich wiring resources, stochastic communication algorithms were proposed [2, 3].

A Routing Algorithm for Random Error Tolerance in Network-on-Chip

1213

Probabilistic flooding [2]. In this algorithm, the data source, named producer, broadcasts a packet with probability p. Then some of its neighbors have a copy of the packet. In the next flooding cycle, all copy-owners also broadcast the packet to their neighbors with probability p. This will continue until the data sink, named consumer, receives a correct copy. Probability flooding does not need retransmission because consumer will receive redundant copies from its neighbors and this algorithm has exponential broadcasting speed. Direct flooding and N-redundant random walk [3]. Direct flooding overcomes the destination ignorance problem of probability flooding. However, flooding algorithms, both probability and direct flooding, have very high overhead in terms of the amount of total packets to be transferred through the network. In N-redundant random walk algorithm, each switch only forwards the packet to one of its neighbors, which is randomly chosen instead of broadcasting. This algorithm sends deterministic amount of N packets, which is pre-computed, into the network to achieve the required reliability. Each of the N packets follows a slightly different way from producer to consumer. It is proved that random walk can significantly reduce the amount of total packets and then power consumption is saved.

3 The Proposed Random Error Tolerance Routing Algorithm For end-to-end retransmission mechanism, the transmission latency may be long because packets may be destroyed on the way, but intermediate switches do not have such knowledge. Switch-to-switch retransmission responds more quickly, as each switch checks the passing packets and requests retransmission when necessary. However, it brings more computing and processing on the way. In addition, end-to-end error correcting has very complex logic. Stochastic communication is used to overcome the long transmission latency of request retransmission. In flooding there is no retransmission, packets can be forwarded very fast, because there are many redundant packets in the network. However, redundant packets will result in very high power consumption, which is not acceptable. N-redundant random walk algorithm is proposed to deal with the above-mentioned problem. However, it also has many constraints. First, each switch must have a random number generator, which increases chip area and slow down forwarding speed. Second, choosing of N is complex and it can not be used in arbitrary on chip environments. Last, this algorithm can not achieve 100% reliability. Suppose the consumer receives Nth packet, which has been destroyed, while the algorithm has already stopped, thus the consumer can not get the correct packet. On the other hand, if the consumer receives the correct packet before Nth one, producer will continue to send redundant packets until Nth copy. As a result, there will be a waste of bandwidth and power. Based on the analysis above, we propose a Random Error Tolerance (RET) routing algorithm with end-to-end feedback. Fig. 2 gives a 4X4 on-chip mesh network. Suppose (00) is the producer and (33) is the consumer. To avoid unnecessary overhead we choose shortest path routing, packets can only be routed rightward or downward in this case. Let 0 for rightward and 1 for downward, the solid line represents routing vector (001101). All routing vectors in this case are composed of 6 bits, among which three are 1s and the other three are 0s. Then there are total 20 (6! / (3!*3!)) routing vectors.

1214

L. Zhang, H. Li, and X. Li

We store these routing vectors in producer’s memory, and generate a random number to choose one of the routing vectors, which will be encapsulated in the packet later. Each intermediate switch reads one of the six bits to determine whether to deliver the packet rightward or downward. When a packet arrives at the consumer, it will be checked. The consumer feeds back ack or nack to inform producer whether it has received a correct copy. To send ack/nack control packets, routing vector is extracted and reversed. For example, if the routing vector is (001101) in the original data packet, ack/nack will follow (101100) in which 0 for leftward and 1 for upward. The meaning of 0 and 1 is determined by DF described later. Ack/nack packets will be routed along the same way back as shown in Fig. 2. This eases routing vector choosing at consumer end. Producer will continually send redundant packets into network, one for each cycle, until it gets a correct ack. Each packet will follow a shortest but randomly chosen routing. Suppose producer with the coordinate (i, j) and consumer (m, n), let Δx = m − i ,

(Δx + Δy )! entries and each entry is one Δy = n − j .The routing vector table contains Δx!*Δy! of the full permutations of ( Δx + Δy ) bits, including Δx 1s and Δy 0s. A special Direction Field (DF) is needed to indicate the relative position of producer and consumer as shown in table 1. We also need a Bit Check Pointer (BCP) in the packet to tell the switch which bit in the routing vector needs checking. The related pseudo-code of the algorithm is described below. Pseudo-code of RET routing algorithm Producer () { every cycle do { randomly choose a vector in routing vector table; pack DF, routing vector, payload, CRC etc. into a packet; set BCP to point to the first bit of routing vector; deliver the packet to the affiliated switch; if (ack correctly received) break;} } Consumer () { while (a packet received) { extract DF and routing vector in packet; reverse routing vector and convert DF to form new pairs; BCP points to the first bit of the new routing vector; if (CRC correct) {generate ack packet and deliver; break;} else {generate nack packet and deliver;} } Switch() { while (a packet received) { check DF and the bit in routing vector pointed by BCP; Look up forwarding action table (table 2) to determine the action of next hop; move BCP to point to the next bit; deliver the packet;} }

A Routing Algorithm for Random Error Tolerance in Network-on-Chip

1215

Table 1. Direction Field Condition m>i; m>i; m