Arabic Type-Making in the Machine Age: The Influence of Technology on the Form of Arabic Type, 1908-1993 [1 ed.] 9789004349308, 9789004303775

Citation preview

Arabic Type-Making in the Machine Age: The Influence of Technology on the Form of Arabic Type, 1908–1993

Islamic Manuscripts and Books Editorial Board Christoph Rauch Staatsbibliothek zu Berlin Arnoud Vrolijk Leiden University

VOLUME 14

The titles published in this series are listed at brill.com/imb

Arabic Type-Making in the Machine Age The Influence of Technology on the Form of Arabic Type, 1908–1993 By

Titus Nemeth

leiden‎ • ‎b oston

Cover illustration: Schematic view of the casting edge of a composed line of Arabic linecaster matrices. From Al-munaḍḍadä al-ʿarabiyyä, (Brooklyn, New York: Mergenthaler Linotype Company, 1929), 7. The Library of Congress Cataloging-in-Publication Data is available online at http://catalog.loc.gov

Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill-typeface. The Arabic sections have been typeset in DecoType “Emiri” in WinSoft Tasmeem. Typesetting by Titus Nemeth. issn ‎1 877-9964 isbn 9‎ 78-90-04-30377-5 (hardback)‎ ISBN ‎9 78-90-04-34930-8 (e-book) Copyright 2017 by Koninklijke Brill NV, Leiden, The Netherlands.‎ Koninklijke Brill NV incorporates the imprints Brill, Brill Hes & De Graaf, Brill Nijhoff, Brill Rodopi and Hotei ‎Publishing.‎ All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical,‎ photocopying, recording or otherwise, without prior written permission from the publisher.‎ Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NV provided that the appropriate fees are paid directly to The Copyright Clearance Center,‎ ‎222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA.‎ Fees are subject to change.‎ Brill has made all reasonable effforts to trace all rights holders to any copyrighted material used in this work. In cases where these effforts have not been successful the publisher welcomes communications from copyright holders, so that the appropriate acknowledgements can be made in future editions, and to settle other permission matters. This book is printed on acid-free paper and produced in a sustainable manner.

Für Wanni und León

∵

Vollkommene Typographie ist eher eine Wissenschaft denn eine Kunst. Beherrschung des Handwerks ist unerläßlich, aber ist nicht alles. Denn der sichere Geschmack, der das Vollendete auszeichnet, beruht auf einem klaren Wissen um die Gesetze harmonischer Gestaltung. Jan Tschichold, „Ton in des Töpfers Hand“, Schweizer Graphische Mitteilungen, 68, 1949, 78–79.

…

Perfect typography is more a science than an art. Mastery of the trade is indispensable, but it isn’t everything. Unerring taste, the hallmark of perfection, rests also upon a clear understanding of the laws of harmonious design. Jan Tschichold, ‘Clay in a potter’s hand’, The Form of the book, (Point Roberts & Vancouver: Hartley & Marks Publishers, 1991) 3.

∵

Contents Note on Typographical Conventions x Note on Terminology xi List of Archives xvi List of Figures xvii List of Tables xxvi Acknowledgements xxvii 1 Introduction 1 Arabic Type in the New Millennium 1 Situating this Book 8 Aspects of the Arabic Script 14 The Typographic Representation of the Arabic Script 22 Arabic Publishing at the Turn of the Twentieth Century 26 2 Mechanical Composition of Arabic 35 The Industrialisation of Printing 35 The Beginning of Arabic Mechanical Composition 45 Mergenthaler Linotype’s First Arabic Linecaster Founts 56 The Beginning of Arabic Type-Making at Linotype & Machinery  69 The Arabic Typewriter 73 Arabic Script Reform 82 Nasri Khattar’s Endeavours in Arabic Type Design 86 Ahmed Lakhdar Ghazal and the Asv Codar Type 90 The Printing Trade and the Middle East after the Second World War 104 Simplified Arabic, Kamel Mrowa, and Al-Hayat  109 Simplified Arabic: Proof of Concept 111 Intertype Abridged Arabic 120 A New Design: Series 8 with 9 124 Monotype Hot-Metal Composition 131 Monotype’s First Arabic Typeface: Series 507 139 A Design for Urdu 148

viii

A New Typeface for the Middle East 155 Monotype Series 549, the Unwitting Mechanisation of Qurʾān Typography 160 Monotype Series 589, Full Vocalisation 170 The Beginning of ­Computer-Aided Arabic Composition 183 The Development of Character Selection and Kashida Insertion Routines 187 3 Photocomposition: Towards Immaterial Type 205 Beginnings of a New Technological Era 205 Arabic on the Monophoto 213 Expansion of Monotype’s Arabic Type Catalogue 219 Simplified Arabic for Monophoto 225 Arabic on Linotype Photocomposition Machines 229 Towards an Arabic Photocomposition Programme 237 Arabic Type Developments by Linotype-Paul Ltd. 253 A Widening Field, the Emergence of New Developers  261 Arabic Type by the Photon Company 261 Endeavours in Arabic Type by Dr Edward Plooij 274 4 Persian Type and Typography 285 Early Persian Printing and Typography 285 A Simplified Typeface for Persian 290 Mitra and Nazanin, the Consolidation of a Persian Typographical Identity 306 5 Beginnings of Digital Arabic Type 319 Type Loses its Material Properties  319 Pierre MacKay and Computer-generated Arabic Type 323 The KITAB Software for the Hell Digiset 327 Compugraphic’s Venture into Arabic Type 335 Digital Tools in the Service of the ‘Hanging’ Styles Nastaʿlīq and Ruqʿä 340 Nastaʿlīq on the Monotype Lasercomp 346 Linotype Ruq’ah, the Exploration of a Writing Style 352 Linotype Sheeraz, a New Approach to Arabic Type 368 6 Looking Forward: PostScript and Beyond 383 The End of Machine-dependence 383 Linotype’s Arabic PostScript Programme 392



ix

Upheaval and Convergence, on the Road to Unicode 400 Desktop Publishing Tools for Arabic 406 DecoType 410 7 Complexities and Simplifications: Factors in the Evolution of Arabic Type-Making 435 Progress through Technology 435 Collaboration between East and West 439 Cultural Identity Expressed through Type 445 Radical Innovation and its Limits 451 Did Technological Advances Improve Arabic Type ? 458 Epilogue: Arabic Type-Making Today 467 Annex 1 475 Annex 2 478 Bibliography 483 Published Sources 483 Patents 494 Online Sources 496 Unpublished Sources 499 Index 500 People 500 Subjects 502

Note on Typographical Conventions Italic is used for multiple purposes. Besides the conventional use for publication titles, it is also used for transliterations of Arabic and Persian terms and other foreign language terms. Note that newspaper names are only italicised if the actual publication is meant; the newspaper as an institution or company is set in roman type. If Arabic names have a conventional Latin transcription it is generally used, with a transliteration according to the system of the Deutsche Morgenländische Gesellschaft provided upon the first occurrence. Any formatting in quotations is preserved from the original and no specific formatting is applied. Thus, any words emphasised with italics, capitals or underlines are maintained as in the source.

Note on Terminology The words used to describe the processes and concepts of type-making and typography have evolved over the five hundred years of their practice in the Western hemisphere. As roughly four-fifths of this time was marked by the manual techniques of the various crafts involved in printing, terminology often reflects this technological legacy, as well as its origins from the Latin script. Many of the words that are still in use today have, however, all but lost the connection to their origins, leaving students and practitioners at a loss for their appropriate usage. Rather than an end in itself, the appropriateness of use arises from the capacity of chosen terms to describe and distinguish various aspects of at times complex matters; and whilst complexity may seem daunting, it is bound to occur in a specialised field at the crossroads of history, technology, linguistics and the interplay of different cultures. Indeed, complex issues require precise language as a guide towards a better understanding of the subject. In the present book some difficulty stems from the discrepancy of an established Western trade with a comprehensive and therefore dominant vocabulary, and a much younger Arabic script practice, in which a terminology is yet to be agreed on. Moreover, if typographical terms have some currency in languages using the Arabic script, they are not comprehensively documented, and thus remained opaque to me. Therefore no attempt will be made to coin novel words, or propose the use of specific terms for aspects of Arabic type. Instead, I have employed existing terms from related fields such as linguistics or calligraphy, or where reasonable from Western typographical practice. Furthermore, some familiarity with current typographical terminology such as ‘ascender’ or ‘contrast’ is assumed, and common sense is called for in the interpretation of concepts such as ‘colour’ or ‘interlinear space’. In general, definitions are only provided for the meaning of terms and concepts which are specific to the present book and subject, or wherever my usage may diverge from expectation. Baseline. The line on which Latin script letterforms appear to stand. Although Arabic also uses a baseline, not all letterforms are aligned with it. In some writing styles (e.g. Nastaʿlīq and Ruqʿä) only the last element of a letter group aligns with the baseline. All the preceding elements of the group relate to the one aligned to the baseline in accord with the morphological rules of the spe-

xii

cific letter sequence. This stacking of letters relative to the last letter of the group results in the appearance of a diagonal, cascading configuration within a line of text. Body (of type). The rectangle which holds the character configuration. The size of the character shape is not necessarily or usually congruent with the body, as commonly some distance is kept between the tallest and lowest elements of a character shape and the boundaries of the body. This discrepancy between body and actual character size explains why different types set in the nominally same size often have different dimensions. Casters were used for the casting of foundry type, and as an integral part of the Monotype system which consisted of a keyboard and a composition caster; its distinguishing feature is apparent from the name as the Monotype caster also composed lines of freshly cast sorts for immediate page make-up, whereas conventional casters only produced unordered sorts for manual foundry type composition. Character is the smallest element of a fount of type. All characters in a fount define the character set. Note that this definition applies prior to the general use of the Unicode standard in which the term ‘glyph’ assumes the meaning that character previously held. As the present book largely precedes the Unicode era, the term ‘glyph’ would be anachronistic and is therefore avoided. Character set describes which characters are contained in a fount. Character repertoire and fount synopsis or extent are used synonymously in this book. Diacritics or diacritical signs are marks – usually smaller than a letter and placed above or below it – which modify the sound that the letter represents. Foundry type is movable type cast from a lead-alloy for the manual composition of individual sorts into lines and paragraphs. It is the typographical technology which defined typography from Gutenberg until the introduction of the Linotype. Font. The digital representation of one specific type style. It is not to be confused with the term typeface. A typeface containing regular, italic and bold styles therefore relies on three distinct fonts. Usage in this book reflects the



xiii

historical evolution of the term, as the British spelling ‘fount’ was superseded by the US spelling ‘font’ towards the end of the twentieth century: with the beginning of digital type-making and the consequential influence of US technology companies also the language of the trade changed. Fount is the collection of individual characters of an instance of a typeface in analogue typesetting technologies. A fount of foundry type or hot-metal is dependent on size and quantity of sorts, and as such an instance of a typeface that might entail multiple sizes and styles. In analogue photocomposition the fount is the film matrix that holds the characters, and is largely size-independent. Friskets were employed in the production of founts for photocomposition. They were made from a transparent laminate that was covered on one side with rubilith film. On the film the enlarged outline of a character was traced with a scalpel, then the cut-out was removed to reveal the typeform in negative against a red background. The frisket could then be used as a negative for the production of phototypesetting founts, assuming the role that the pattern had played in hot-metal matrix manufacture. Hot-metal refers to typographical composition with a Monotype system or a linecasting machine. In both cases the type is freshly cast and composed mechanically; as individual sorts and as lines respectively. Justification is the adjustment of typeset matter with the goal to achieve identical line widths. Whereas Latin script typography primarily relies on the modification of the spaces between words, Arabic typography may also employ the extension of letterforms and the use of letter variants to achieve the same goal. Kern. The part of a typeform that projects beyond the body of a character into the space above or below an adjacent character. In foundry type and Monotype hot-metal composition kerns rested on the shoulders of neighbouring sorts and were prone to breaking under the pressure of the press. The verb kerning describes the process of defining kerns in type-making or typography. Letterform is the visual representation of a sound and/or concept in a writing system. In the Arabic script individual letters have multiple letterforms. In type a single letterform may be represented with one or more characters.

xiv

Letter group. All but seven letters of the Arabic script are written in a continuous, joined way. Because seven letters cannot join to the left, words may be interrupted by spaces. Thus, a single word may contain multiple letter groups, each formed of multiple letterforms. Ligatures are a typographical method for the rendering of multiple letterforms with a single character of movable type. Although frequently employed in Arabic type-making, ligatures are neither specific to the script, nor particularly suitable for it, as will be discussed later in the book. The notion that Arabic is defined by a high number of ligatures is therefore a categorical mistake as it projects a typographical concept on the script itself. Linecasting is the composition process of linecasters, principally the Linotype and the Intertype machine. It differs from most other composition processes because its output consists of solid lines of type, also called slugs, as opposed to sequences of individual characters. Whilst advantageous for speed and facility of page make-up, linecasting had considerable drawbacks for typographical quality. Matrix. A die, usually made from brass or copper, into which a character is stamped or engraved. In type-founding and hot-metal composition matrices are used in combination with moulds to cast individual sorts or lines of type from molten metal. Morphology is borrowed from the study of linguistics to describe the rule-based characteristics of word formation, letter variation, and vocalisation found in the Arabic script. Patterns were used in the mechanical production of hot-metal founts. Typically made from bras, patterns were employed by all typesetting machinery manu­ facturers for the pantographic cutting of punches. The pattern thus stored the character configuration for the virtually unlimited production of identical matrices. In photocomposition friskets fulfilled the same role. Rasm is the Arabic term for the skeleton of a letterform without diacritical signs. It is employed here for lack of an equivalent in English typographical terminology.



xv

Sort. An individual character of foundry type. Typeface refers to the visual configuration of a set of related characters and founts which share a unifying gestalt. Typeface is used here similarly to ‘design’ in the narrow sense of appearance. Type describes the actual implementation of a typeface in a composition technology. Thus, a typeface may be more successful as a hot-metal type than as a photocomposition type. Typography is the use of type for the design of text. It is not type design, and most typographers do not design type, nor do type designers usually practise typography. Vocalisation is used as shorthand for the notation of vowels, orthographic and grammatical marks in a consonantal writing system. Because vocalisation is optional in Arabic it is considered as a distinct aspect of the script.

List of Archives Non-Latin type collection at the Department of Typography & Graphic Communication, University of Reading, United Kingdom. Abbreviated as DTGC. Musée de l’imprimerie et de la communication graphique, Lyon, France. Abbreviated as MI. Archives of Monotype, Ltd, Salfords, Surrey, United Kingdom. Abbreviated as MT. Mergenthaler Linotype Company Records, 1905–1993, Archives Center, National Museum of American History, Washington, DC. Abbreviated as NMAH.

List of Figures 1.1 1.2 1.3 1.4 1.5 1.6 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11 2.12 2.13 2.14 2.15 2.16 2.17 2.18 2.19 2.20 2.21 2.22

Approximate distribution of Arabic script use along current boundaries 2 ‫ٱ‬ ‫ ت ت‬national ‫ك‬ The word �‫�ا ب‬ �� ��‫ � ����س‬stiktāb in its word-formation, and divided into the individual letters of the word 18 Layers of the Arabic script 20 Rasm of the letters �‫ ب‬bāʾ and �‫ �ي‬yāʾ 23 Foundry type sort with a ligature of letters ‫ ك‬kāf and �‫ �ي‬yāʾ 24 Notional arrangement of ­vowel sorts in relation to other ­characters 24 Punch and corresponding matrix as used in manual type-making 35 Punch Cutting Machine 40 The first commercially available Linotype 42 Schematic illustration of a composed line of single letter matrices 42 Diagram of the progressive positions of the matrices travelling through the Linotype machine 44 Illustration of a duplexed matrix 44 Arabic Linotype linecasting machine as built in the 1910s 46 Salloum Mokarzel, circa 1940 50 Mergenthaler 22 pt Arabic specimen 53 Detail of the opening page of the chapter about linecasters in Legros and Grant, Typographical Printing-Surfaces 54 Stages of pattern making 58 Schematic drawing of the pattern’s function in guiding the punchcutting machine 58 Comparison of Arabic fount extent 60 Schematic view of the casting edge of a composed line of Arabic linecaster matrices 62 Illustration of kerning problems 62 Illustration of inadequate mark positioning 63  Detail of al-Mustaqbal newspaper 64  Specimen of Mergenthaler’s Arabic founts 66 “14 point Arabic” specimen 67 Details of Figure 2.19 68 Type specimen from Mākīnāt al-Laynūtayb mākīnāt aṭṭibāʿä mākīnāt Ūtūblayt liṣabbi aṣ-ṣafaḥāt 70 Patent drawings from Selim S. Haddad, Types for Type-Writers or Printing-Presses 75

xviii

2.23 2.24 2.25 2.26 2.27 2.28 2.29 2.30 2.31 2.32 2.33 2.34 2.35 2.36 2.37 2.38 2.39 2.40 2.41 2.42 2.43 2.44 2.45 2.46 2.47 2.48 2.49 2.50 2.51 2.52

Keyboard layout of Hammond Arabic typewriter 76 Arabic Olympia SM9 typewriter 77 Examples of character design principles as employed in typewriter founts 78 Typewriter composition sample demonstrating the character design principles in actual words 78 Diagrammatic keyboard layout of the Arabic Model 5 Remington Portable typewriter 79 Diagrammatic keyboard layout of an Arabic Continental typewriter 80 Diagrammatic Olympia keyboard layouts for Pashto, Persian and Urdu 81 Illustration of the design principles of Unified Arabic 87 Type specimen from Al-ḥurūf Al-Mūḥḥadä 89 Hand-lettered headline alphabet inspired by Unified Arabic 90 “Systeme ASV-Codar Naskhi-2” specimen 92 “Trial No. 1” of Monotype Series No. 640, No. 641, No. 642 94 Illustrations of the design process for a Ruqʿä typeface 102 Front cover and spread of It can now be revealed 105 Photograph of the Linotypes in the machine room of the Government Press of Pakistan 107 “12 point Urdu No. 2” specimen 108 Kamel Mrowa, probably in 1966 109 Type sample produced on an Arabic German Continental typewriter 110  Forms of �‫ ب‬bāʾ in different contexts 113 The characters for �‫ ب‬bāʾ from “­Linotype Arabic Light with Bold” 113  Simplification principle of �‫ ب‬bāʾ 114 “Simplified Arabic” proof of the 12D size under development 115 Detail of front page of al-Hayat 117 Article announcing the release of MrowaLinotype Simplified Arabic 118  Cover page of Linotype Matrix, 33 121 Specimen of three sizes of Mrowa-Linotype Simplified Arabic 121 Intertype Abridged Arabic, early proof of the fount under development 123 “Namūḏaǧ min al-ʿarabiyy al-muḫtaṣar ǧism 16 bunṭ” advance specimen 123  Detail of al-Ahram newspaper 125 “A new design in Mrowa-Linotype Simplified Arabic” specimen 127



2.53 2.54 2.55 2.56 2.57 2.58 2.59 2.60 2.61 2.62 2.63 2.64 2.65 2.66 2.67 2.68 2.69 2.70 2.71 2.72 2.73 2.74 2.75 2.76 2.77 2.78 2.79 2.80 2.81 2.82 2.83 2.84 2.85 2.86 2.87 2.88

xix

Keyboard layouts for Mrowa-Linotype Simplified Arabic 127 “Yakout” typeface specimen 130 Monotype DD style keyboard and caster 132  Monotype D style keyboard, counting and perforating mechanism 133 The matrix case in its pre-1963 configuration with 15×15 rows 135 Partly filled matrix case 136 Drawing of a 0.2″ × 0.2″ Monotype matrix 136 Comparison of different MCAs 136 Draft Arabic MCA 141 Draft of Arabic keyboard layout 141 “Trial No. 1 Series No. 507, 14-pt” 142 “Trial No. 2 Series 507, 12D” 143 “Trial No. 4a Series No. 507, 14D” 146 “Trial No. 8, Series No. 507, 12D (redesigned)” 147 Page of a lithographic copy of a handwritten Bible 153 “Monotype Urdu Series No. 507” specimen 153 Composition of characters from figure 2.68 154 Specimen of the Egyptian Government Press fount in 15 pt 157 “Trial No. 1, Series 409, 16-pt”  159 Page from the King Fūād Qurʾān 161 Diagram of the nominal configuration of foundry type sorts for vocalised text setting 162 “Trial No. 1, Series No. 549, 16D and 18D” 164 “Trial No. 2, ­Series No. 549, 16D” 165 “A Specimen of ‘Monotype’ Arabic” 167 Arabic text shown in “A Specimen of ‘Monotype’ Arabic” 169 An iǧāzä, the Islamic certificate of competence in calligraphy 170 Detail of a page from The M.E.C.A.S. Grammar of Modern Literary Arabic 171 Justified row of Arabic Linotype matrices 171 Specimen of vowel matrices set of the Linotype 24 pt Arabic 26◊ 544 fount 172 Detail of specimen of vocalised Linotype composition 172 Page of The Accented Arabic elementary school books 174 “Trial No. 1, Arabic Naskh Accented, Series No. 589, 24D” 178 “Trial No. 2, Arabic Naskh Accented, Series No. 589, 24D” 179 Illustration of the interlocking sorts of Monotype’s system for the composition of floating marks and vocalisation 180 Detail of figure 2.88 180 “Trial No. 3, Arabic Naskh Accented, Series No. 589, 24D” 181

xx

2.89 2.90 2.91 2.92 2.93 2.94 2.95 2.96 2.97 2.98 2.99 2.100 2.101 2.102 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15 3.16

Friden Flexowriter 2300 series 8-level punched tape typewriter 184 Fairchild Teletypesetter operating unit 184  “T.T.S perforator keyboard Mrowa Linotype Simplified Arabic” 187 Detail of al-Ahram newspaper 188 Linotype Elektron linecaster 189 The Justape special purpose justifying computer 191 Hrant Gabeyan at his desk in the offices of Antoine Homsy’s firm 192 Hrant Gabeyan in front of Linotype House, 21 John Street, London 194 “‘Kashida’ Logic for Linotype Simplified Arabic” 196 “Arabic Perforator Keyboard Logic for Linotype Traditional Arabic” 198 Detail of figure 2.98 199 “‘One-Character’ Keyboard for Simplified Arabic” 200 Meeting at the MELO, Beirut, at the occasion of the launch of the Elektron 202 Tom Marsden explaining the features of the Elektron at the MELO, Beirut 203 Comparison between the drawings by Jan van Krimpen and the types made from them 207 “Intertype presents New Horizons for the modern printer,” promotional publication for the Intertype Fotosetter 208 Fotosetter single- and double-letter matrices 208 Optical system of the Fotosetter 210 Close-up of individual Monophoto film matrices against a lit background 210 “Arabic in All Its Glory”, announcement of Monophoto Arabic Series 589 215 Comparison of hot-metal and photocomposition versions of Series 589 215 “Page from Telephone Directory, ‘Monophoto’ Filmsetter work, printed by Bank Melli Press Iran, Tehran” 217 “12 point Urdu No. 2” 218 “‘Monophoto’ Urdu Series 549” specimen 219 “Trial No. 1 ‘Monotype’ Arabic Naskh Bold, Series No. 649, 16D” 220 “Trial No. 2 ‘Monotype’ Arabic Naskh Bold, Series No. 649, 16D” 221 Comparison of set widths of characters in Monotype Series 549 and Series 649 222 Photographic reproduction of a Nastaʿlīq design provided by a certain Mr Minai 224 “Urdu Nastaliq, Series 679, 10-pt” proof  224 “Arabic 749” proof 228



3.17 3.18 3.19 3.20 3.21 3.22 3.23 3.24 3.25 3.26 3.27 3.28 3.29 3.30 3.31 3.32 3.33 3.34 3.35 3.36 3.37 3.38 3.39 3.40 3.41 3.42 3.43 3.44 3.45 4.1 4.2

xxi

Linofilm system 230 Linofilm fount grid 230 Linofilm Quick filmsetter 232 Comparison of boundaries of hot-metal characters to photocomposition characters 232 Linotron 505 236 Linofilm V-I-P 237 Linotron 505C character-generating system 238 Linofilm V-I-P A-range fount strip of Helvetica Condensed 239 Schematic diagram of the V-I-P’s optical system 240 “Arabic ‘Standard’ Alignment” 242 Sample of Osman Hussein’s lettering 244 Artwork provided by Osman Hussein 246 Photographic reduction of Osman trial characters and comparison to Monotype Series 549 247 Comparison of the vertical space available on the Linotron with Monotype’s system 248 Diagram comparing nominal point size to actual face size for Arabic typefaces on the Linotron 505C 248 Text sample of vocalised Arabic text composed in Osman on a Linotron 505C 250 Sample of individual characters in regular and bold weights of the finished Osman typeface 252 Drawing for ‫ ط‬ṭāʾ, Jalal 257 Drawings of characters ḥāʾ, ‫ �ه‬hāʾ, and the ‫ح‬ ‫ف‬ rasm of a �� fāʾ and �‫ �ي‬yāʾ ligature 258 The Department of Typographic Development, letterdrawing studio and research staff, 1983 260 Photon-Lumitype 200-500 fount matrix disc 263 Mixing of styles and sizes on the Photon-Lumitype 263 Detail from the Tunisian newspaper as-Sabah 266 Nasriphot 201.55 proof 267 Large-size lettering for Nasriphot 268 “Arabica” specimen 271 Detail of a page of Utusan Melayu 273 “Inrichting voor het machineschrijven of zetten van Arabisch en daarmee verwant schrift,” Dutch Patent NL113873 275 Specimen fragment by Edward Bernard Plooij 280 “The fainting of Laylah and Majnun” 284 Alahram specimen 289

xxii

4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 4.13 4.14 4.15 4.16 4.17 4.18 4.19 4.20 4.21 4.22 4.23 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10 5.11 5.12 5.13 5.14 5.15 5.16

“Persian by Hermann Zapf & Arabic” specimen 290 Proof of Linotype Simplified Persian fount 291 “Specimen of 10-pt loose type selected by Ettela’at” 291 Raed specimen 293 Preliminary specimen Raed Persian 293 Sample of Persian fount provided by Massoudi 295 Page from Douglas Turner to Walter Tracy, “Simplified Persian” 296 Artwork for Simplified Persian 299 Notional indication of the matrix widths of selected characters in artwork for Simplified Persian 301 Illustration of a composed line of double letter matrices 301 Ettelaʿat Simplified revised artwork 302 “10D-point Trial Characters for Ettela’at Simplified” 303 “Linotype Simplified Persian Keyboard 387 with French Keyboard 103” 304 Components of the Linotron 303 system 307 The character-generating system of the Linotron 303 307 Fount grid of the Linotron 303 307 Illustration of the maximum character area available on the 505C 309 Character drawing for mīm, Mitra bold 311 ‫م‬ Mitra proof 312 Character drawing of the ���‫ ش‬šīn �‫ �ي‬yāʾ ligature of Mitra 313 Nazanin specimen 317 Diagrammatic illustration of character-generation by strokes 320 Comparison of output from Linotron 404 and Linofilm V-I-P 320 Bitmap edit sheet 320 Hell-Digiset filmsetter 322 Bitmaps for the typeface Edison 322 Page 20 from On Burning Mirrors 326 Adly Boulos and Hrant Gabeyan 329 The calligram for KITAB 330 Bitmap of the isolated character of ‫ ك‬kāf of the Klett-Interpart KITAB system 330 Page composed with Klett-Interpart KITAB system 332 Detail of figure 5.10 332 The Compugraphic EditWriter 7500 335 Video screen of the EditWriter 7500 336 Cover page of the “Arabic Documentation Package” 338 Schematic diagram of the Linotron 202 341 Diagrammatic illustration of outline description of character image as used for Linotron 202 fonts 342



5.17 5.18 5.19 5.20 5.21 5.22 5.23 5.24 5.25 5.26 5.27 5.28 5.29 5.30 5.31 5.32 5.33 5.34 5.35 5.36 5.37 5.38 5.39 5.40 5.41 5.42 5.43 5.44 5.45 5.46 5.47 5.48

xxiii

CRTronic output of Jalal 343 Arabic CRTronic S 343 Monotype System 3000 Lasercomp 345 Matlubul Hasan Saiyid and Ahmad Mirza Jamil 346 Sample of digital Urdu Nastaʿlīq corpus compiled by Matlubul Hasan Saiyid and Ahmad Mirza Jamil 348 Announcement of Monotype Nastaliq 350 Handwritten and typeset columns from The Daily Jang 351 Ruqʿä sample by the Ottoman calligrapher Mustafa İzzet Efendi 353 Hand-lettered street sign 354 Roqua‘i, corps 36, specimen 355 Detail of cover page of al-Ahram newspaper 355 Character drawings for hot-metal Ruqʿä fount 356 Drawings of additional Ruqʿä characters ordered by al-Ahram newspaper 357 Diagrams showing the exposure fields of the V-I-P and the Linotron 606 359 “Drawing guidelines for Arabic typefaces” 359 Drawing of Ruqʿä extenders in relation to the exposure window of the Linotron 202 360 Character drawings for Linotype Ruq’ah 360 Detail of a dot-matrix printer proof of Linotype Ruq’ah 362 Detail of Linotron 202 proof of Linotype Ruq’ah 363 Character drawings of of Linotype Ruq’ah and hand lettering of same letter sequence 364 Comparison of typical Ruqʿä morphology with Linotype Ruq’ah typeface 364 Sketch explaining the use of the Kashida character in the Linotype Ruq’ah font 366 Sketch explaining design characteristics of the Linotype Ruq’ah typeface 366 Page from “Layinūtāīb tuqaddim ǧihāz al-Layinūtrūnīk 500” 367 Linotype’s proposed scheme for typesetting Nasta’liq 369 Photographic trial setting of Holloway’s Nastaʿlīq artwork 370 Selection of manuscript samples from Pakistan 370 Nastaʿlīq letter drawings by Holloway 372 Page from Sheeraz Nastaliq software specification draft 373 Illustrations of ‘knock-on value’ and ‘zone of influence’ concepts 375 Explanatory note specifying font behaviour 375 Proof of the Sheeraz Nastaliq font marked up for errors 375

xxiv

5.49 5.50 5.51 5.52 5.53 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 6.12 6.13 6.14 6.15 6.16 6.17 6.18 6.19 6.20 6.21 6.22 6.23 6.24 6.25 6.26

“Nastaliq milestone chart” 377 Proof of Sheeraz Nastaliq vowel positioning in various contexts 379 Illustration accompanying suggestions for font behaviour changes by Holloway 379 Clipping of Nawa-I-Waqt, column composed in Linotype Sheeraz 380 Detail of “Sheeraz” specimen 381 Comparison of high-resolution phototypesetter output and 300 dpi resolution laser printer output 384 Schematic illustration of a setup combining a low-resolution laser proofing printer and high-resolution output device 386 Article announcing ITC’s publication of Arabic typefaces 388 “Stempel Arabic typefaces” specimen 393 Diagrammatic illustration of the outline description employed in the Laserfont format 393 Design review of floating marks 395 Review process of the character for final ʿayn 395 ‫ع‬ Accent test Jalal light 396 Apple Macintosh Arabic font layout 397 Comparison of font production processes 399 “Oryx Page Make-up” cover 399 Illustration and explanatory text about the ATM 3.9 401 Wood lettering type case by DecoType 411 Side-view of wood lettering modules by DecoType 411 Detail of a photocopy of an Ottoman manuscript 413 Close-up photograph of brass matrix and proof casting in resin 414 Close-up photograph of custom type petals made for a Brother daisy wheel printer 415 Detail of a proof from a Brother daisy wheel printer 415 Two publications by the municipality of Amsterdam from the late 1980s 416 Spread from English-Persian Dictionary of Legal and Commercial Terms 417 Thomas Milo with his daughter Sarah 418 Proof of DecoType’s Ruqʿä style bitmap font 418 A caption from the calligraphy exhibition at the Zamana Gallery London 419 “Typesetting Islamic Calligraphy” promotional leaflet 420 Selection of type drawings made in the development of DecoType’s first Ruqʿä font 422 Sample setting of DecoType’s earliest Ruqʿä font 423



6.27 6.28 6.29 6.30 6.31 6.32 6.33 6.34 6.35 6.36 6.37 6.38 6.39 6.40 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10 7.11

xxv

Contents of the Microsoft Office 6 Arabic retail package 423 Samples of typographical options provided through the DTP Naskh font set 424 Cover of “DTP Naskh, The DecoType Professional Font Series” 426 Sample setting of DTP Naskh 426 Interface of the DecoType Authentic Ruq’ah Server 428 Sample setting of the DecoType Authentic Ruq’ah Server 428 Window of WinSoft PageMaker 5.5 ME 429 Explanatory text and illustrations of the DecoType Setter 430 Sample setting of the DecoType Authentic Naskh Server 431 Interface of the DecoType Authentic Naskh Server 431 Word Shaping interface of Tasmeem 6 in Adobe InDesign CS 6 433 Control of letter variations with the Text Shaping interface 433 Illustration of the break-down of lettershapes into smaller, recurring elements 434 Interface of the Position Tuner tool 434 “We write a different Arabic,” advertisement in an-Nahar newspaper 438 Cutting from an-Nahar newspaper 438 Page composed in the hot-metal version of Monotype Series 549 440 Page composed with the hot-metal version of Monotype Series 507 447 Page from the 1938 Hyderabad Government Press Qurʾān 449 Ettelaʿat Simplified, composition of Haghighi’s full name from character drawings 450 Nazanin, proof composed on a Linotronic 300 in 18-pt 450 Mitra, proof composed on a Linotronic 300 in 18-pt 450 Sample of Compugraphic’s first Arabic typefaces 461 Sample setting of Microsoft fonts Arial, Times New Roman, Simplified Arabic and Traditional Arabic 463 Monotype Naskh specimen 465

List of Tables 1.1 1.2 1.3 1.4 6.1

The Arabic abjad 16 Changes in meaning with varying vocalisation 17  ‫ ت‬t � b 18 Transformations of the root ‫ ك‬k � ‫ب‬ Disambiguating function of diacritical signs for Arabic 19 The Unicode character-glyph model 403

Acknowledgements I owe much to the Department of Typography and Graphic Communication at the University of Reading, UK, where this book originated. Based on my doctoral research, and rooted in the Department’s synthesis of theory and practice, the present book shows the influence of this institution and its staff. I am indebted in particular to Fiona Ross, Gerry Leonidas and Eric Kindel, whose sensitivity to letterforms and scripts, conceptual vision, and attention to the minutiae of craft – from type-making to the choice of words – were central in shaping my understanding of our field. I am most grateful to Christopher Burke for reading drafts of this book as a friend and exemplar of the informed reader to whom this book is adressed. My research would not have been possible without the archives that contain the artefacts and traces that were left in making Arabic type, nor without the people who enabled me to research their contents. I would like to thank the National Museum of American History, Washington DC; the Monotype Archives in Salfords, Surrey; the Musée de l’imprimérie et de la communication graphique in Lyon, France; and the former Linotype GmbH in Bad Homburg, Germany. Help in accessing these collections was kindly provided by Cathy Keen, Carrol and Karl Kindel (Washington); Patrick Giasson, Ian Bezer, Robin Nicholas (Salfords); Alan Marshal, Alice Savoie (Lyon); Otmar Hoefer, and Dan Reynolds (Bad Homburg). I also want to acknowledge Nada Itani of the Dar alHayat Information Centre, Beirut, and Hannelore Herpertz of the IZM Aachen. Special thanks to Frank Romano for his generous gift of Arabic specimens. In my search to piece together the history of Arabic type-making, I was fortunate to meet and exchange with some of the contributors of the recent, and not so recent, past. Here, I would like to thank especially Hrant Gabeyan, Thomas Milo, Mirjam Somers, Michael Fellows, Kamal Mansour, and Hans Jürg-Hunziker. Indeed, I feel privileged not only to have met, but to have developed friendships with some of the forebears of my profession. Amongst my contemporary fellow practitioners and friends, I would like to extend special thanks to John Hudson, Tim Holloway, and Martin Tiefenthaler. Finally, I would like to thank my family for their support in all strands of life. Lisa and Volker, you excel the fabulous parents that you are as grandparents. Johannes, always my big brother. Wanni, for making it all happen. León, for lending the future to the past. I would not be able to do what I am doing without your help, love and care. TN, Vienna, June 2017

Chapter 1

Introduction Arabic Type in the New Millennium Arabic type is not new. The first Arabic book composed with movable type was printed in Europe in the early sixteenth century, and from the eighteenth century printing with Arabic type was practised in print shops in the Ottoman Empire.1 However, it was during the twentieth century that Arabic typemaking became a domain in its own right, and evolved into the specialised field we know today. Since the turn of the millennium, Arabic type design has become a prominently discussed topic, a development recognised in the wider field.2 Indeed, Arabic type is probably more widely discussed than it has ever been.3 The subject appears in specialised publications, websites, conferences, and the occasional mainstream media coverage, the latter being a rare feat in the world of type at large.4 That Arabic is among the more discussed subjects in the typographical field is no coincidence, but a reflection of wider trends, as described by Gerry Leonidas: Typefaces becoming part of the mainstream is neither accidental, nor a fashionable blip. It was foreseeable many years ago, and has been accelerating under the dual impetus of the accelerating move to a standards-

1 See Miroslav Krek, “The Enigma of the First Arabic Book Printed from Movable Type,” Journal of Near Eastern Studies 38, no. 3 (1 July 1979): 203–12. 2 See for example Robin Kinross, Modern Typography: An Essay in Critical History, 2nd ed. (London: Hyphen Press, 2004), 181. 3 My emphasis on Arabic type design, as opposed to typography, is deliberate: curiously, there is no comparably visible debate about the use of Arabic type, as there is about its design. Moreover, discussions of Arabic type are commonly framed in a narrow sense of visual design, and omit the underlying technological conditions that prescribe how type is made and shaped; a shortcoming that the present book seeks to address. 4 For publications see the following discussion; for websites see for example the two Dutch websites www.khtt.net and www.arabictypography.com, and the Lebanese www.29arabicletters.com; for conferences Kitabat, held in 2006 in Dubai, was an early expression of the increasing activity in the field and was followed by numerous subsequent events.

2

Figure 1.1

Chapter 1

Approximate distribution of Arabic script use along current national boundaries. Solid colours: contemporary primary script; vertical stripes: contemporary use as a secondary national script; horizontal stripes historical use. Colours loosely indicate cultural zones of influence that are reflected in the dominant writing styles. Yellow: North African, Maġrabī; purple: Arabic, Ottoman, Nasḫ, Ruqʿä; red: SubSaharan styles; green: Persian, Nastaʿlīq; grey: East-Asian writing styles which somewhat fall outside the conventional script nomenclature. Evidently, cultural influences often extend beyond national boundaries, the present map is thus more an aid to orientation than a definitive description. Map adapted from Wikipedia by the author.

Introduction

3

compliant, text-orientated internet, and the growth of mobile devices with usable browsers.5 As Leonidas points out, within this general phenomenon, type design for multiple scripts has attained a place of particular importance. Globalised communications increasingly call for international document exchange, relying on fonts that can display more than one language. Commercial interests are decisive, as companies seek new markets whilst aiming for unified visual appearances across language and cultural borders. Consequently, bespoke typefaces that cover two, three or more scripts are becoming more common as global enterprises are recognising the importance of type that corresponds to cultural expectations. Such commercial demands provide the impetus for increasing design activity, contributing to the growth in releases of so-called multi-script typefaces.6 It comes as no surprise that Arabic, the second most-widely used phonemic script in the world, plays a leading role in this trend (Figure 1.1). Not least because of the economic importance of the oil-producing Gulf States and the associated business interests in the region, recent years saw a sharp increase in the number of Arabic typefaces released and commissioned. This commercially-induced expansion has not, however, been accompanied by comparably active research. Taking the year 2000 as a nominal beginning for the increasing interest in Arabic type, then seventeen years seem like a sufficiently long period to see the emergence of substantial discourse about the subject, including considerations of recent history and current practice. Yet so far, this has not been the case. The majority of literature concerned with Arabic type and typography engages with pre-industrial developments. A growing body of historical studies concentrates on the first European printers cutting Arabic types, and the pioneers of the trade in the Middle East. Notable recent publications representative of this branch of research are the edited volumes Middle Eastern Languages and the Print Revolution, Historical aspects of printing and publishing in languages of the Middle East, and The History of the Book in the Middle East, as well as Hendrik D. L. Vervliet’s study of Robert Granjon’s Arabic types.7 While these studies provide important historical research, they 5

Gerry Leonidas, “The Next Ten Years,” 12 March, 2013, http://leonidas.org/2013/03/12/ thenexttenyears/ (accessed 16 May 2017).

6

The establishment of the Rosetta Type Foundry in 2011, the first type-maker explicitly dedicated to the development of non-Latin typefaces, is indicative of these wider trends. Rosetta Type Foundry, http://www.rosettatype.com (accessed 16 May 2017).

7

Eva Hanebütt-Benz, Dagmar Glass, and Geoffrey Roper eds. Middle Eastern Languages and the Print Revolution: a Cross-Cultural Encounter: a Catalogue and Companion to the

4

Chapter 1

generally afford less consideration to detailed assessments of the types under discussion. In these writings, the general narrative – the chronology of events and their situation within a larger socio-historical perspective – generally outweigh considerations of the design processes, the artistic achievements, and their relevance for subsequent typographical developments. A practitioner’s critical perspective on the actual typefaces, the processes and judgements involved in their making and the type’s qualitative appraisal is not present.8 Similarly, when literature in the related domain of calligraphy discusses the subject, it rarely finds a successful approach to the specific questions of type; a failure for which it cannot be blamed, given the novelty and limited amount of academic discourse about Arabic typography and type design.9 While no history of Arabic typography in the twentieth century has yet been written, individual works provide important starting points for further research. Notable are the publications by Walter Tracy, long-time typographical adviser of the British Linotype & Machinery Ltd company. Although few in number and not exhaustive in coverage, they provide first-hand insight into a number of important Arabic type developments.10 As these accounts were Exhibition (Westhofen: WVA-Verlag Skulima, 2002); Geoffrey Roper ed. Historical aspects of printing and publishing in languages of the Middle East : papers from the Third Symposium on the History of Printing and Publishing in the Languages and Countries of the Middle East, University of Leipzig, September 2008 (Leiden: Brill, 2014); Geoffrey Roper ed. The History of the Book in the Middle East (Farnham: Ashgate, 2013); Hendrik D. L. Vervliet, “Cyrillic & Oriental Typography in Rome at the End of the Sixteenth Century: An Inquiry into the Later Work of Robert Granjon; 1578–1590,” in The Palaeotypography of the French Renaissance: Selected Papers on Sixteenth-century Typefaces, vol. 2, 2 vols. (Leiden: Brill, 2008), 427–74. 8

Robin Kinross warned that typographical history, as produced by practising typographers, “has been of mixed benefit”, for although “this kind of history is the only one to recognise the aesthetic factor in printing, […] it has had the tendency to do little else but view.” Kinross, Modern Typography, 17. This study tries to avoid this shortcoming by having done “its time in the archives” as Kinross suggests, and by the critical analysis of circumstances leading to design decisions, rather than focussing on finished products.

9

Sheila Blair’s Islamic Calligraphy is a recent example of a study in which the chapter on Arabic typography is not of the same depth and quality as the rest. Blair, Islamic Calligraphy (Edinburgh: Edinburgh University Press, 2008).

10

Walter Tracy, “The Flourishing Reed,” in Alphabet: International Annual of Letterforms (London: James Moran Limited, 1964); Walter Tracy, “Arabic Without Tears,” in The Penrose Graphic Arts International Annual, vol. 68 (London: Northwood, 1975), 121–26; Walter Tracy, “Advances in Arabic Printing,” Bulletin (British Society for Middle Eastern Studies) 2, no. 2 (1975): 87–93.

Introduction

5

written shortly after the events they describe, they are significant historical sources, providing largely more reliable and detailed information than later secondary sources.11 Similarly, recent publications by Fiona Ross, effectively Tracy’s successor at the company, provide a relevant perspective for this study.12 Overall, however, the history of Arabic type and typography remains a subject that has only been researched fragmentarily. Reliable published sources are sparse, and primary research about Arabic typographical history is wanting. Students and practitioners of the subject rely on the few publications which are easily accessible, yet unfortunately not all of them exhibit the hallmarks of sound research. This is particularly marked in literature aimed at a general design audience – a substantial element in a field dominated by practitioners – where scrutiny and depth of content are sometimes side-lined in favour of exceedingly simplified narratives. An example of such literature is Arabic Typography: a Comprehensive Sourcebook.13 Published in 2001 it became the first visible trace of the broadening interest in the subject. As an English language mainstream publication with a subtitle promising thorough treatment, it was welcomed by the wider design field.14 Based on this publication Smitshuijzen AbiFarès was rapidly recognised as an authority in the field. Attracting funding from cultural bodies in The Netherlands, numerous typography related projects were initiated, making laudable contributions to a widening awareness of Arabic type design. Concomitant publications, workshops and efficient media use provided the growing body of interested students and practitioners with readily available information. However, the expertise of Smitshuijzen AbiFarès was soon challenged. As convincingly demonstrated in the joint review of Arabic Typography by Thomas Milo and Mohamed Zakariya, authorities in Arabic type-making and 11

This applies least to Tracy’s latest, anecdotal and partly inaccurate recollections of his working life. See Walter Tracy, “Composing Room Days (and After),” Printing Historical Society Bulletin 40, no. Winter 1995/96 (1995): 3–15.

12

Fiona Ross, “The Type Design Process for Non-Latin Scripts,” in Non-Latin scripts, from metal to digital type (London: St Bride Library, 2012), 125–54; Rob Banham and Fiona Ross, Non-Latin Typefaces: At St Bride Library, London and Department of Typography & Graphic Communication, University of Reading (London: St Bride Library, 2008).

13

Huda Smitshuijzen AbiFarès, Arabic Typography: a Comprehensive Sourcebook. (London: Saqi Books, 2001).

14

See for example Jason Smith, “Book Review,” Eye Magazine 12, no. 45 (2002): 82. The title of the book is somewhat misleading, as about half of the book is explicitly dedicated to type design.

6

Chapter 1

­calligraphy respectively, the publication is marred by numerous shortcomings.15 According to Milo and Zakariya, its generally shallow treatment of the subject is compounded by factual errors and an overall lack of first-hand information. Without subject-specific training or actual experience in type design, exacerbated by the lack of primary sources to substantiate claims, Smitshuijzen AbiFarès’ grasp of the topic remains superficial and tainted by apparent preconceived ideas. As Milo pointed out, this is exemplified by the apparent “uncritical faith in Digital Omnipotence [which] is characteristic of the uninitiated”.16 A recurring feature that determines many of the shortcomings of popular literature, is the apparent ignorance of Arabic typographical history. Smitshuijzen AbiFarès, for example, dismisses two decades of well documented design and invention in passing, stating that between 1965 and 1985 “all type manufacturers were busy converting their existing typefaces from metal to photocomposition to digital typesetting” and that “no new designs were created for Arabic text fonts”.17 Neither claim is substantiated; both are factually wrong. Elisions and simplifications are, however, not exclusive to Smitshuijzen AbiFarès. Other recent publications also fail to present an informed and balanced image of historical developments. Stefan Moginet’s 2009 Du calame à l’ordinateur : l’évolution graphique de l’écriture arabe promises an account of the 15

Thomas Milo and Mohamed Zakaria, “Beautiful Book, Made to the Highest Western Standards,” Al-Computer, Communications & Electronics, 2006, 64–73.

16

Ibid., 71–72. Further publications by Smitshuijzen AbiFarès unfortunately do not rectify the demonstrated problems. Whilst the 2006 Typographic Matchmaking: Building Cultural Bridges with Typeface Design aspired to become “the benchmark for future developments in this field”, the practical reality fell short of its claims. Barely any applicable guidance for the design of Arabic type is provided in the publication and new Arabic fonts supplied with the book were largely dysfunctional, demonstrating a stark contrast between rhetoric and factual competence. Huda Smitshuijzen AbiFarès, Typographic Matchmaking: Building Cultural Bridges with Typeface Design (Amsterdam: BIS Publishers, 2006), 8. A recent book by Edo Smitshuijzen further descends into polemic as illustrated by his remarks about the Arabic script world: “First, Arabic [sic] societies show generally a very low interest in graphic design and typeface design. Just go to an Arabic book fair, buy a newspaper or a magazine. It will make any ambitious graphic designer cry. The quality is just painful to look at. And everything looks the same.” etc. Edo Smitshuijzen, Arabic Font Specimen Book (Amsterdam: De Buitenkant, 2009), 65. The similarity of the authors’ names is not coincidental: Edo Smitshuijzen is the partner of the author of Arabic Typography. For a comprehensive critical review see Titus Nemeth, “Complex Dutch Arab Complex,” TYPO (2009): 86–93.

17

Huda Smitshuijzen AbiFarès, Arabic Typography: a Comprehensive Sourcebook., 81.

Introduction

7

visual evolution of the Arabic script, yet only dedicates a fraction of the book to a superficial and inconclusive glance at typographical history.18 Mourad Boutros’ Arabic for Designers is not intended as a historical work but a general introduction to a multitude of aspects surrounding Arabic design practice. Yet, the importance Boutros gives to Letraset dry-transfer lettering, without mentioning Linotype’s work in Arabic type-making, give a pointed insight into Boutros’ selection criteria.19 Similarly, Boutros et al.’s Talking about Arabic maintains the questionable practice to discuss almost exclusively examples of Arabic type-making which are directly linked to the contributors, and fails to provide any traceable evidence or sources for its claims.20 The bulk of literature thus brushes aside substantive historical enquiry and presents a view which paves the way for practice that is severed from historical precedents.21 It suggests exaggerated novelty and exceptional achievements in current Arabic type design, and thereby implicitly diminishes previous contributions. Indeed, in view of current literature the student of Arabic type history is bound to conclude that the twentieth century was an era of stagnation for this field, and that Arabic type-making only attained a level worthy of consideration when digital production methods became mainstream. Yet as will be shown throughout this book, quite the contrary was the case. Not only is Arabic type nothing new, it has received the attention of generations of printers, draughtspeople, engineers, company representatives, typographical advisers, programmers, publishers, designers, calligraphers and even politicians and clergymen. Their ideas, concepts, skills, ideologies and aesthetics all have shaped the path on which Arabic type evolved, and provided many 18

Stefan F. Moginet, Du calame à l’ordinateur : l’évolution graphique de l’écriture arabe (Gap: Atelier Perrousseaux, 2009). Also published as Writing Arabic: From Script to Type (Cairo: The American University in Cairo Press, 2009). Evidence from the collection at the DTGC indicates that Moginet had unsuccessfully offered a typeface for production to Linotype & Machinery Ltd, suggesting familiarity with the typographical world of the 1970s. Stefan F. Moginet to Linotype & Machinery Ltd, Letter, 12 September 1977, WT correspondence, folder 18b, DTGC.

19

Mourad Boutros, Arabic for Designers (New York: Mark Batty Publisher, 2007). Boutros’ personal involvement in the development of Arabic Letraset designs explains, but does not justify such disproportionate discussion. The book’s historical negligence is demonstrated by Boutros’ misattribution of a typeface to Tracy, and the misspelling of his name.

20

Mourad Boutros et. al., Talking About Arabic, (New York: Mark Batty Publisher, 2009).

21

J. R. Osborn’s recent Letters of Light is a welcome corrective that tentatively seeks to address this problem, but continues to rely on published sources and therefore foregoes most of the twentieth century from its account (Cambridge, Massachusetts & London: Harvard University Press, 2017).

8

Chapter 1

of the foundations upon which the trade rests today. This legacy is found in typeforms, font synopses, structural principles and stylistic interpretations, and quintessentially in conventions that are now often taken for granted. Yet cognisance of these historical facts and their influence on the evolution of this field is essential. Only a critical appreciation of the decisive forces that carved the contours of Arabic type-making to their present image can prevent the naïveté and ignorance that lie at the heart of poor practice. Arguably, all of the facets that defined Arabic type, from the macroscopic motivations and circumstances – economic, political, religious – to the minutiae of technical aspects and the complex interplay between tools, form, and identity, ought to be known and understood in order to fully grasp Arabic type-making today. Based on this understanding the present book aspires to shed some light on the history of a field which is too often framed ahistorically. Situating this Book In the West the training of most typography and type design students includes lessons about the history of the trade, its evolution through different periods, stylistic and technological changes, and ideally the relevance of typography in a larger societal context. A distinction is even made between different aspects of history, as specialists of printing history may not be considered as specialists of type design history, and vice versa. Few would contest that familiarity with the historical evolution of typography helps the contemporary practitioner to make informed decisions, and that an appreciation of the history of conventions and customs can provide the critical judgement necessary for their evaluation. What the historian John Tosh articulated regarding the use of history for political purposes, applies equally to the history of type-making: To know about the past is to know that things have not always been as they are now, and by implication that they need not remain the same in the future. History can be the basis for scepticism about received ideas, which is why in totalitarian societies history-writing is not regarded as a harmless escape from the present.22 By contrast, Arabic typographical history has not yet been recognised as this rich source of information for current students and practitioners. As noted 22

John Tosh, The Pursuit of History. Aims, Methods, & New Directions in the Study of Modern History, 2nd rev. ed. (London and New York: Longman, 1991), 6.

Introduction

9

above, academic studies of Arabic typography tend to be histories of printing, and make few attempts to establish connections to current practice. By and large, the history of Arabic typography is a self-contained discipline, which rarely extends as far as the twentieth century. History, it appears, is only researched if it is at a historically safe distance. This is surprising, for Arabic typography – although practised – could hardly be called a field in its own right before the twentieth century. Indeed, if one follows the ‘simplified’ distinction between printing and typography that Robin Kinross defines, it is questionable whether one can even talk about Arabic typography prior to the nineteenth century: […] the difference is between inarticulate practice with the materials of production (‘printing’), and conscious shaping of the product, by instruction (‘typography’). One might well argue, with this distinction in mind, that ‘modern’ typography is indeed a duplication of sense, because when printing becomes typography is also when printing becomes modern. Printing becomes modern with the spreading of knowledge about itself: with the published description of its practices; with the classification of its materials and processes; with co-ordination of dimensions of materials, enabling their exchange and better conjunction; with the establishment of a record of its history.23 Printing in Arabic had few, if any, of these characteristics before the nineteenth century, when tentative traces of the beginning of a print culture began to emerge.24 Indeed, the inappropriateness of the term Arabic typography before 1900 is apparent when considering that “in the Arab world, the art of printing, publishing, and distribution of books on a large scale among all classes and among young and old approached the European level and acquired some semblance of maturity almost four centuries after Gutenberg’s invention”.25 Given this delay, increased attention should be given to the history of Arabic typography as it evolved in the last 100 years. Indubitably, current contributions to 23

Kinross, Modern Typography, 15–16. Kinross proposes this distinction to define the scope of his essay about the notion of ‘modern’ typography.

24

See for example Geoffrey Roper, “Fāris Al-Shidyāq and the Transition from Scribal to Print Culture in the Middle East,” in The Book in the Islamic world: The Written Word and Communication in the Middle East, edited by George N. Atiyeh, 209–32 (Albany: State University of New York Press, 1995).

25

Atiyeh, “The Book in the Modern Arab World: The Cases of Lebanon and Egypt,” in The Book in the Islamic world, 233.

10

Chapter 1

Arabic typography and type design would have much to gain from an appreciation of developments that happened during the period in which Arabic printing and publishing “acquired some semblance of maturity”, for this recent history is still most tangible in today’s practice. This study rests on the conviction that for Arabic type-making and typography, an historical perspective which maintains a connection to current issues provides clues to the origins of established norms, and thus facilitates their assessment, review and potentially reconsideration. Moreover, the increased breadth of comparison also assists in evaluating the merits of contemporary work. As Ross put it, “best practice is built upon sound research”.26 Ross’ approach to historical enquiry, which is not only applied in the standard work on Bengali type history, but also embodied in her practice, informs the present study.27 Whilst multi-faceted and in Ross’ work specifically adapted for Bengali type-making, a distinguishing feature of this approach is the emphasis put on technology. It considers technological constraints and their change through time as a principal determinant of the evolution of non-Latin type, emphasising the disproportionate influence mechanical limitations had for scripts with a complex morphology.28 As such, it provides a constant against which to measure typographical developments, and opens various directions of enquiry through which this study tries to shed light on the evolution of Arabic typemaking. This technological lens reflects economic considerations, another decisive element for an understanding of a field so strongly marked by the pragmatic needs of the trade; it allows an assessment of the historical reasons for design decisions; it illuminates conceptual and artistic skills in the adaptation to given constraints; it provides a gauge to understand the priorities of manufacturers and clients; it makes for an important selection criterion to distinguish novel 26

Ross, “The Type Design Process for Non-Latin Scripts,” 151.

27

Fiona Ross, The Printed Bengali Character and Its Evolution (Richmond, Surrey: Curzon, 1999); More recently also Jo De Baerdemaeker’s PhD study on Tibetan built on methods articulated by Ross. See Jo De Baerdemaeker, “Tibetan Typeforms: An Historical and Visual Analysis of Tibetan Typefaces from Their Inception in 1738 up to 2009” (PhD thesis, University of Reading, 2009).

28

The term non-Latin is widely recognised as not being satisfactory. It lumps together scripts without any logic other than them not being Latin. A better term, or indeed a range of terms are lacking, and sorely needed. Here, however, it is used for the single merit it has: most scripts are more complex – either in morphology or quantity of letterforms – than the Latin script. Arguably, ‘non-Latin type’ has the merit of acknowledging this difference, and the importance it had whenever technology conceived for the Latin script was employed for the composition of other scripts.

Introduction

11

from generic approaches; and lastly, it spins a thread through disparate historical events and eras, connecting current to past practice, and thus aspires to offer insights valuable to historians, students and practitioners alike. This is the first study of Arabic type-making in the twentieth century that systematically employs primary sources – a challenging and rewarding task. Much of recent Arabic typographical history is documented in professional archives of the companies that worked (and sometimes still work) in the field. These collections amount to what have been called “the ‘records’ of history par excellence”.29 They contain vast numbers of business letters, memoranda and notes, original artwork and other documentation of some of the most influential developments of Arabic typographical history. Started as records of everyday practice, the materials offer a degree of organisation, but are not systematically catalogued. Generally held in boxes and folders with crude and at times misleading labels such as ‘Arabic’, ‘Orientals’, ‘Middle East’, the documents are diverse and provide a wide range of evidence, from the irrelevant to the historic. The sheer quantity of material warranting examination was daunting, as hundreds, possibly thousands of records had to be examined, assessed, referenced and interpreted. Particularly the latter proved difficult, as contrary to assumption, well documented records do not simply provide facts. Often chronological order was not readily apparent from the sources, and had to be established. But more importantly, the relevant aspects of this mass of information had to be identified, and the relations between events established. Descriptions of events in the sources provided important perspectives, which then had to be assessed from the vantage point of historical hindsight. In this interpretative work, consultation of different archives proved essential for a balanced perspective. Thus, I could verify claims of one manufacturer against the records kept by a competitor – or in the case of the archives of the British Linotype & Machinery Ltd and the US Mergenthaler-Linotype companies, corroborate evidence from these complementary sources: often correspondence between the two sister companies only survived at one end, and comparison of records in both locations provided clues to the historical jigsaw puzzle of archival research. As Tosh notes, the process “has all the fascination of eavesdropping”, and may even bring about the excitement of detective work.30 But the main reward was of a more historical satisfaction. Using untapped sources, many more novel findings could be unearthed than would have been the case in a better charted field. 29

John Tosh, The pursuit of history: aims, methods and new directions in the study of history, 6th rev. ed. (Abingdon-on-Thames & New York: Routledge, 2015), 75.

30

Tosh, The Pursuit of History, 75.

12

Chapter 1

To complement this archival work, interviews with selected participants of the investigated events were conducted. These additional accounts were sought for a wider variety of primary sources, to fill gaps within archival records and to verify inconclusive evidence. Participants were chosen for their documented relevance to the subject, and for their availability. As a significant part of this study deals with eras for which no living memory remains, only the second half of its time-span could be thus complemented. Moreover, the majority of employees of businesses active in type-making used to be anonymous as far as the final products and their documentation were concerned. Many of the people who worked for some part of their lives in type-making left the field when a business closed, or personal circumstances changed. Prior to the late twentieth century, only a select few employees of type businesses had public faces and left any historical traces. The use of oral history sources therefore remained limited, but where applicable it contributed to the depth of enquiry. Unavoidably, and as is widely recognised, the use of oral sources entails its own difficulties, and potential pitfalls. Not only do the questions and presence of the interviewer affect the testimony of the witness, but personal recollection is prone to distortions: His or her memories, however precise and vivid, are filtered through subsequent experience. They may be contaminated by what has been absorbed from other sources (especially the media); they may be overlaid by nostalgia (‘times were good then’), or distorted by a sense of grievance about deprivation in childhood which took root only in later life. To anyone listening, the feelings and attitudes – say of affection towards a parent or distrust of union officials – are often what lend conviction to the testimony, yet they may be the emotional residue of later experience rather than the period in question.31 The risk of distortion appears to be the more pronounced the more time has passed between the event and its recollection, and seems unavoidable where autobiographical details feature in the account. Here, scrutiny of other sources was necessary to verify claims, sometimes resulting in contradiction of the interviewee, as the author occasionally experienced during this research. Inescapably, this study has its limitations. Based primarily on the records of type manufacturers, and written from the perspective of a practitioner, this study has a bias towards the making of type, rather than its use. The views of customers, though partially documented in the archives of the companies investigated, were not actively sought and may provide routes of enquiry for 31

Ibid., 269.

Introduction

13

further research. In particular, archives of printers and publishing houses in the Middle East could, if extant and accessible, shed further light on the interaction between regional clients and (mostly) European type manufacturers. Moreover, Arabic, Persian, Urdu and other language sources were not considered in this study, and may prove relevant in further research. Another, deliberate, limitation of the current study is the selection of cases. Priority was given to developments and events which illustrate turning points in practice and which demonstrate a range of varying approaches to similar typographical problems. A selection of the historically most significant cases for the evolution of Arabic type was sought, rather than a comprehensive account of all Arabic type developments which would have sacrificed depth of treatment for breadth of coverage. The selection was informed by the question how processes of type-making and composition influenced the form of Arabic type from the introduction of mechanical typesetting to the advent of Unicode, and the lateral themes that emerged from this enquiry. Thus, developments that illustrate different aspects of the making of Arabic type were included instead of cases that followed repetitive patterns. The rationale of this selection was to demonstrate the multifaceted nature of Arabic type-making during the time frame under consideration. Although the assessment of typefaces must entail subjective considerations, there are criteria which allow for evaluations independent of personal preference. This study adopts a framework for the appraisal of typefaces which loosely follows Ross’ principles, defined in the context of Bengali type. Fundamental prerequisites for a satisfactory design are the accurate reproduction of the script’s morphology, the apparent understanding of its historical evolution – articulated through the consistent and culturally appropriate use of formal expression – and, where applicable, the correct representation of typographical customs. Crucially, proportions and principal configurations of letterforms have to correspond to established conventions to be readable, and at the same time all letterforms of a typeface should be individually distinctive and yet form a cohesive whole. Indeed, while the stylistic treatment of letterforms is partly guided by taste, one can identify criteria for appraisals of form: the consistent application of stylistic choices, the proportional relations of characters, the evenness of colour and density of text, and the clarity and distinctiveness of the image, partly defined by the quality of drawings, are decisive for the success of a design. Another category for the assessment of a typeface is found in its intended purpose. A design for newspaper composition generally has to follow different considerations than a typeface for fine book printing, and wherever applications are expressly articulated, the gauge for its success lies in its intended use.

14

Chapter 1

Legibility in small sizes, sufficient contrast between regular and bold weights to create hierarchies, and the graphic effect of shapes in large sizes are just some of the criteria that may determine the fitness for purpose of a design. Moreover, as will emerge from this study, factors unrelated to the visual characteristics often prove decisive for how a type fares in practice: technological circumstances, the evolutionary stage of the field, economic, political, and general cultural aspects all influence its actual success. The historical scope of this book is defined by technological evolution, as suggested in its title. The start-date of 1908 refers to the development of the first Arabic typesetting machine, the Linotype linecaster, marking the beginning of unprecedented influence of technology on the form of Arabic type. This machine age then bridges a range of different technologies, each providing the background for new and different influences, limitations and opportunities for Arabic type-making. The end-date is not as clearly demarcated, as historical proximity makes the gauging of relevance more difficult. Yet the decision to end in 1993 – whilst taking the freedom to extend this scope whenever developments did not fall neatly within a strictly defined period – follows the notion of the machine, for it was here, with the general introduction of the Unicode standard, the popularisation of personal computers and the advent of the Internet age, that Arabic type lost the direct link to the machines which had defined its shape for the preceding 85 years. Aspects of the Arabic Script Arabic forms part of the North Semitic script family. It evolved from Nabataean and its first established use dates from the fourth century.32 Today, it is the second most widely used phonemic script in the world, encompassing numerous languages which are linguistically unrelated to Arabic; for example Malay, Urdu and Uighur each belong to different language families.33 This wide dissemination of the script resulted in the rich linguistic and cultural diversity within the Arabic script world, manifest in the wide range of writing styles, calligraphic techniques and manuscript traditions, as tentatively outlined in Figure 1.1 on page 2. 32

Florian Coulmas, The Writing Systems of the World, (Oxford: Basil Blackwell, 1989), 150.

33

Only the Latin script and Chinese characters are used more widely, however the latter is generally considered a logographic script. Malay, Urdu and Uighur languages belong to the Austronesian, Indo-European and Altaic language families respectively. See “Ethnologue. Languages of the World,” n.d. http://www.ethnologue.com (accessed 16 May 2017).

Introduction

15

The Arabic script falls within the class of Semitic writing systems, which also includes Hebrew.34 Through the spread of the Arabic script with Islam, numerous orthographies for other languages evolved. Additional letters – new graphemes that are mostly distinguished through diacritical signs – were introduced to provide for the differences in phonemic repertoires. Persian, for example, requires four additional letters for sounds not found in the Arabic ‫گ‬ language: �‫ پ‬pe ‫ چ‬če �‫ ژ‬že and � gāf, with other languages making use of nu� merous additional graphemes. Despite these additions and variations for different orthographies, the principal characteristics of the script remained identical. For the sake of clarity, its morphology is discussed here only on the basis of the Arabic language.35 The Arabic script is a so-called ‘abjad’ – a consonantal script – which is written from right to left and contains 29 letters; all letters are consonants but the following three, which may also serve as long vowels: ‫ ا‬alif, ‫ و‬wāw, �‫�ي‬ yāʾ ­(Table 1.1). Short vowels are commonly omitted, but may be used for disambiguation and in specific texts. Religious, educational and technical works make ample use of such vocalisation. These short vowels, as well as other orthographic signs such ّ as the ◌ šaddä which indicates the doubling of letters, are denoted by small marks placed above or below the letter they modify. For the beginner of Arabiclanguage studies, texts without vowels can be difficult to read as the meaning of a word may change significantly with different vocalisation (see Table 1.2 over�leaf). Indeed, meaning has often to be derived from the wider context, a process that is facilitated by the language’s root system. In Arabic, as in other Semitic languages, typically three consonants, also called radicals, form a root that provides a general semantic concept. Grammatical transformations of the root in accordance with regular patterns then define the precise meaning of the word. 34

Following the system of linguist Florian Coulmas, here the following categories are distinguished: writing system, script and orthography. According to Coulmas, “every writing system makes a selection of the linguistic units to be graphically represented (not language specific) […] every script makes a specific selection of the possibilities of a given system in accordance with the structural conditions of a given language” and “every orthography makes a specific selection of the possibilities of a script for writing a particular language in a uniform and standardized way.” Coulmas, The Writing Systems of the World, 38–39.

35

The term morphology is borrowed from linguistics, where it is used to describe “the structure, form, or variation in form (including formation, change, and inflection) of a word or words in a language”. “Morphology, N.,” OED Online (Oxford University Press, n.d.), http://www.oed.com/ (accessed 16 May 2017). Here morphology is used to describe the rule-based characteristics of word formation, letter variation, and vocalisation found in the Arabic script.

16

Chapter 1

Letter name ‫َ�هْ زَ �ة‬ hamzä �‫�م‬ َ ‫أف‬ ���‫�ِ�ل‬ َ ‫ب�ا ء‬ َ ‫ت�ا ء‬ َ ‫ث�ا ء‬

�� ‫جِ يم‬ َ ‫ح�ا ء‬ َ ‫�خ �ا ء‬ َ ‫دال‬ َ‫�ذ‬ ‫ال‬ َ ‫را ء‬ َ ‫ز�ا ء‬ ‫ن‬ �‫ِ��س��ي‬ ‫ن‬ �‫���شِ ��ي‬ َ ‫�ص�ا د‬ َ ‫� �ا د‬ ‫�ض‬ َ ‫ط�ا ء‬ َ ‫�ظ �ا ء‬

‫َْ ن‬ ���‫عي‬ ‫�غَْ ن‬ ���‫ي‬ َ‫ف‬ ‫��ا ء‬

‫َق ف‬ �� ‫��ا‬ ‫َ ف‬ � �� ‫كا‬ َ ‫ل‬ ‫ا‬ ‫م‬ �‫ِ�مي‬ ‫م‬ ‫نُ ن‬ �‫�و‬ َ ‫�ه�ا ء‬ َ ‫وا و‬ َ ‫ي�ا ء‬

Transliteration Isolated ‫ء‬ ʾ

alif

a

bāʾ

b

tāʾ

t

ṯāʾ

ṯ

ǧīm

ǧ

ḥāʾ

ḥ

ḫāʾ

ḫ

dāl

d

ḏāl

ḏ

rāʾ

r

zāʾ

z

sīn

s

šīn

š

ṣād

ṣ

ḍād

ḍ

ṭāʾ

ṭ

ẓāʾ

ẓ

ʿayn

ʿ

ġayn

ġ

fāʾ

f

qāf

q

kāf

k

‫ا‬

�‫ب‬ ‫ت‬ � ‫ث‬ � �‫ج‬

‫ح‬ ‫خ‬ � ‫د‬ ‫�ذ‬ ‫ر‬ ‫ز‬ �

‫��س‬ ‫ش‬ ���

‫�ص‬ ‫ض‬ �� ‫ط‬ ‫�ظ‬

‫ع‬ ‫�غ‬

‫ف‬ �� ‫ق‬ �

‍‫ا‬

�� ‫‍ب‬ ‫ت‬ ��‍ ‫‍ث‬ �� �‍� �‫ج‬ �‍� ‫ح‬ ‫�‍� خ‬ � ‫‍�د‬ ‫�ذ‬ �‍ ‫‍�ر‬ ‫ز‬ ��‍

‫‍��س‬ ‫ش‬ ���‍

‫‍���ص‬ ‫ض‬ ����‍ ‫‍��ط‬ ��‍‫�ظ‬

‍ ‫ع‬ ‫‍�غ‬ ‫ف‬ �‍ ‫ق‬ �‍

Medial

‍��‫‍ ب‬ ‍��‫‍ت‬

‍��‫‍ث‬

‍�‫�‍� ج‬

��‍� ‍‫ح‬ ‍�‫�‍� خ‬

Initial

‍�‫ب‬ ‫�‍ت‬ ‍�‫ث‬

‍�‫ج‬

� ‍‫ح‬ ‍�‫خ‬

‍‫‍��س‬

‍ ��‍‫�ش‬

‍‫‍���ص‬

‫‍���ض‬ ‍� ‍‫‍��ط‬ ‍ ��‍‫�ظ‬ ‍‫‍�ع‬ ‍�‍‫�غ‬ ‍��‍‫ف‬ ‍��‍‫ق‬

‍‫��س‬

‍ �‫� �ش‬

‍‫�ص‬

nūn

n

hāʾ

h

wāw

w

yāʾ

y

‫�ه‬ ‫و‬

�‫�ي‬

‫‍�ه‬

‫‍و‬

�‫‍ي‬

062D 062E

‍‫‍�ل‬

‍��‫‍ن‬

‍‫�‍�ه‬ ‍��‫‍ي‬

0632 0633 0634 0635

0638

‍‫ط‬

‍‫�ع‬ ‫�غ‬ ‍� �‍‫ف‬ ‫�ق‬ ‍

‍‫�م‬ ‍�‫ن‬

‫م‬ ‫ن‬ �

062C

‍ ‫�ظ‬

‍‫‍م‬

‫‍�ل‬

062B

0636

�‍ ‫م‬ ‫ن‬ ��‍

‫ل‬

062A

‫�ض‬ ‍�

� ‍‫‍ ك‬

l

0628

0631

‫‍�ك‬

m

0627

0630

‫ك‬

lām

Unicode 0621

062F

� ‍‫ك‬

mīm

Table 1.1

Final

‍‫�ل‬

‍‫�ه‬ ‍�‫ي‬

0637 0639 063A 0641 0642 0643 0644 0645 0646 0647 0648 064A

The Arabic abjad. Adapted from The M.E.C.A.S. Grammar of Modern ­Literary ­Arabic, (Beirut: Khayats, 1965). The transliteration scheme of the Deutsche Morgenländische Gesellschaft is used in this book.

Introduction Spelling ‫�ع��ق���د‬ َْ ‫�ع��ق���د‬ َ ََّ ‫�ع��ق���د‬ َُ ‫�ع��ق���د‬ ْ‫ق‬ ‫ِ�ع�����د‬

17

Transliteration

Translation

ʿqd ʿaqd ʿaqqada ʿuqad ʿiqd

n/a contract, n. to complicate, v. complexes, n. necklace, n.

Table 1.2 Changes in meaning with varying vocalisation.

‫ ت‬t � b which indicates meanings This system can be illustrated by the root ‫ ك‬k � ‫ب‬ َ‫ت‬ َ ‫َ ْ َ�ت‬ ‫ك‬ related to writing. Transformations of this root include ‫ا‬ � � kitāb (book), ‫�م ك‬ � � ‫�� � ب‬ ‫ِ ب‬ َ maktaba (office) and �‫كا تِ� ب‬ � kātib (writer / clerk) (Table 1.3 overleaf). One of the most readily apparent characteristics of the script is the joining of letters when written in sequence. All but six letters connect to preceding and following letters, adopting varying shapes depending on their position; ‫ا‬ ‫�ذ‬ alif, ‫ر‬ rāʾ, �‫ ز‬zāʾ, ‫د‬ dāl, ḏāl and ‫ و‬wāw only connect with the letters that precede them, thus creating breaks within the otherwise continuous words (Figure 1.2 overleaf). This connectedness of the script is sometimes compared to Latin cursive handwriting, often with the note that in Arabic no monumental or ‘printed’ form (here defined as a separation of letters) exists. This comparison, however, lacks persuasiveness. As explained by John Hudson, “one should be careful not to confuse the kind of linking between letters that occurs in Latin script, in which morphographical behaviour is minimal and mostly a byproduct of the ductus, and features of writing systems that are normatively morphographical.”36 In Arabic, which is among the latter, the connectedness of letters is part of the script’s morphology. It is one aspect of the joining behaviour, which equally involves the changing of letterforms in different contexts. Joins and elongations are integral parts of letters, and there is reciprocal influence between letter shape and connection. This contrasts with Latin cursive handwriting, where connections are largely discretionary, and have hardly any noticeable effect on the letterform itself: generally short joining elements are added to the letter, but its basic shape need not change. As Hudson points out, “this is made obvious when one considers the positioning of dots and marks relative to the connections, particularly when those connections are elongated. In Latin cursive writing, one would never displace a diaeresis or other accent over the connecting stroke coming off a letter, but this is exactly what one does when elongating, for example, the connecting stroke of the Arabic �‫ ب‬in an initial or medial position.”37 36

John Hudson to ATypI member discussion list, Email, “Re: [ATypI] a New Arabic Typography Book,” email, 31 March 2011.

37

Ibid.

18

Chapter 1

Translation

Transliteration

to write (he wrote)

kataba

writing (n.), essay

kitābä

book

kitāb

Qurʾān school

kuttāb

bookseller

kutubī

books

kutub

office

maktab

offices

makātib

library, bookstore

maktabä

printer, typewriter

miktāb

correspondence

mukātabä

enrollment

ktitāb

dictation

stiktāb

writer, clerk

kātib

written

maktūb

correspondent

mukātib

subscriber

muktatib

Unvocalised ‫�ت‬ �‫�ك‬ �‫�� ب‬ ‫ت‬ ‫كا �ة‬ �‫�� ب‬

Vocalised

‫ت‬ �� �‫كا ب‬ ‫ت‬ �� �‫كا ب‬ ‫�ت‬ �‫�ك‬ �‫�ب�ي‬ ‫�ت‬ �‫�ك‬ �‫�� ب‬

‫�ت‬ ‫�م ك‬ �‫�� � ب‬

�‫�م ك‬ �‫��ا ت� ب‬ ‫���ت����ة‬ ‫�م ك ب‬ ‫ت‬ ��‫ك‬ �‫�م�ا ب‬ ‫� ت����ة‬ ‫�م ك��ا ب‬

‫ٱ �تت‬ �‫� �ك‬ �‫����ا ب‬ ‫ٱ ت‬ ��‫� ����ست�� ك‬ �‫�ا ب‬ � �‫كا ت� ب‬

‫ت‬ �‫�م��ك�و ب‬

�‫�م ك‬ �‫��ا ت� ب‬ ‫�م� ت �ت‬ �‫�ك��� ب‬

َ ‫ََ�ت‬ �‫�ك‬ �‫�� ب‬ َ‫ت‬ ‫كا َ�ة‬ �‫ِ�� ب‬ َ‫ت‬ ��ِ �‫كا ب‬ َّ‫ُت‬ �� �‫كا ب‬ ُ‫ُ�ت‬ �‫�ك‬ �‫�بِ�ي‬ ُ‫ُ�ت‬ �‫�ك‬ �‫�� ب‬ ‫َ ْ َ�ت‬ ‫�م ك‬ �‫�� � ب‬ َ �‫َ�م ك‬ �‫��ا تِ� ب‬

َْ ‫��َ�ت��َ��ة‬ ‫�م ك ب‬ َ‫ْت‬ ��‫ِ�م ك‬ �‫�ا ب‬ َ ُ ‫� ت�َ�َ��ة‬ ‫�م ك��ا ب‬ ‫ٱ ْ َت‬ �‫����ا ب‬ ِ‫� �ك��ت‬ َ‫ٱ ْ ْت‬ ��‫� ����ستِ�� ك‬ �‫�ا ب‬ َ � �‫كا تِ� ب‬ ُ‫َ ْ ت‬ �‫�م��ك�و ب‬ َ �‫ُ�م ك‬ �‫��ا تِ� ب‬ َْ ُ �‫�م��كت���تِ� ب‬

‫ٱ‬ ‫ٱ‬ ‫ت‬ ‫ت‬ ‫ت‬ ‫ت‬ � � ‫ك‬ ‫ا‬ � ‫ب‬ � �� �‫س � ك � ب‬ � ‫� ����س ��ا‬

Table 1.3

‫ ت‬t � b. Adapted from Blair, Islamic Transformations of the root ‫ ك‬k � ‫ب‬ ­Calligraphy.

Figure 1.2

‫ٱ ت‬ From right to left: the word �‫�ا ب‬ ��‫ � ����ست�� ك‬stiktāb (dictation) in its word-formation, and divided into the individual letters of the word. Note that there are two breaks in the otherwise joined word.

Introduction Rasm Letters Table 1.4

‫ا‬ ‫أآ‬ � ‫� �إ‬

19 ‫ٮ‬ ‫ث‬ ‫ت‬ � � �‫ب‬

‫ح‬ ‫خ‬ �‫ج�ح‬

‫�ذ‬

‫د‬

‫د‬

‫ر‬ ‫ز‬ �‫ر‬

‫��س‬ ‫�� ش‬ �� � ‫س‬

‫�ص‬ ‫� ض‬ � � ‫ص‬

‫ط‬ ‫ط �ظ‬

‫ع‬ ‫�غ‬ ‫ع‬

‫ره‬ ‫ه �ة‬ ‫ر‬

Disambiguating function of diacritical signs for Arabic. Other ­languages ­multiply the number of functions of rasms with different diacritical ­configurations.

Another characteristic of the script is its limited number of distinct letter­ forms. Numerous letters share the same structure, or rasm (drawing/sketch) as it is called in Arabic script terminology, and are only distinguished by diacritical signs (Table 1.4);38 moreover, many parts are identical between two or more lettershapes, making the Arabic script a highly efficient and predictable system. In Thomas Milo’s work this characteristic is investigated systematically, and applied rigorously in describing Arabic.39 Borrowing from linguistics, Milo suggests an analysis of the script from the view of contrastive features of letterforms, introducing a distinction between grapheme, allograph and archigrapheme to describe the elements of the Arabic script.40 In Milo’s analysis a grapheme is a letterform with all its distinguishing features. For example �‫ب‬ ‫ ت‬tāʾ are two distinct graphemes. As Arabic letters may have multibāʾ and � ple forms without affecting their meaning, Milo suggests the term allograph to describe these instantiations.41 Thus, �‫ب �‍  ‍ ب��‍  ‍� ب‬ ‫ � ب‬are four allographs of the grapheme �‫ ب‬bāʾ. But for a full description of the Arabic script’s morphology, a yet finer distinction is required. Because many letters share the same rasm and are only distinguished by diacritical signs, Milo introduces the term archigrapheme, which he describes thus: An archigrapheme occurs when the distinction between two or more letters is neutralized. The archigrapheme is a graphic unit that consists of the shared features of neutralized letters minus the features that

38

Adam Gacek, The Arabic Manuscript Tradition. A Glossary of Technical Terms & Bibliography (­Leiden: Brill, 2001) 55.

39

See also “DecoType” on page 410.

40

Thomas Milo, “Authentic Arabic: A Case Study. Right-to-Left Font Structure, Font Design and Typography.,” Manuscripta Orientalia 8, no. 1 (2002): 49–61.

41

Bigelow and Holmes already used ‘allograph’ earlier, but did not develop the terminology further. See Charles Bigelow and Kris Holmes, “The Design of a Unicode Font,” in Electronic publishing, vol. 6 (Nottingham: University of Nottingham, 1993), 289.

Language Culture Type ter groups. In Arabic, however, connecting letters is not the exception but 20

Chapter 1

the rule. Theoretically each letter can have a different appearance in any combination, something that can only be crudely imitated with ligatures. ifferentiate them. the archigraphemic analysis The use­dof ligatures tends to be In determined by the nature of the typefaceof Arabic script, ­vowels design and and the technical limitations of theof font technology. variable Many con-features.42 (Figure 1.3) dots are different layers additional, temporary fonts contain a multitude of ligatures in order to approximate the desired appearance of authentic Arabic text, but it needs to be underHowever, dominant approach to Arabic generally divided stood that this isthe a technical solution to a technical problem, type-making and not an the Arabic script letterforms withofall their distinguishing feainherent feature of the into Arabiccomplete script. It is not hard to conceive different technologies for typesetting Arabic that would achieve equivalent or recurring elements.43 tures (allographs in Milo’s terminology) as the smallest superior results without using any ligatures at all.

abcd Figure 14a. Skeleton script showing individual archigraphemes.

eafbgchd Figure 14b. The addition of dots establishes the identity of ambiguous graphemes.

emafjnbgkochpd Figure 1.3 Layers of the Arabic script. (a) Rasm (archigraphemes). (b) Diacritical marks disambiguate Figure 14d. Vowels are identified, or their absence is marked. distinct letterforms (allographs). (c) Šaddä marks for consonant doubling (d) Short vowels are marked or absence of vowels is indicated. Figure 14e. Punctuation or cantillation (precise verbalisation) is indicated. (e) Recitation aides indicate punctuation or cantillation, only applies in special contexts. (d) Letterforms which may carry other diacritics are further disambiguated by Figure 14f. The erroneous placement of illegal dots in copying is blocked. the use of miniature forms. (g) Discretionary ornamental elements may be used to fill holes in the text. Adapted from Milo, “Authentic Figure 14g. Ornamental elements are added to aesthetically fill holes in the text. Arabic”. Figure 14c. Consonant enhancers identif y reduplicated consonant graphemes.

emqafjnrbgkoschptd emquafjnrvbgkoswchptxd emquEafjnrvFbgkoswGchptxHd eimquEafjnrvFbgkoswGchlptxHd 42

nguage Culture Type 1.indd

43

[ 124 ]

Based on this understanding, Milo developed an approach to Arabic type-making emancipated from notions that arose from historical practice. It uses characters which are parts of letters and pen strokes, rather than complete letterforms. As Milo points out, “if the repetitive nature of Arabic writing is exploited, only the outline paths of a limited set of sub-letter elements need to be drawn”, resulting in a smaller repertoire of individual shapes that can be assembled automatically by a corresponding computer programme. Milo, “Authentic Arabic: A Case Study. Right-to-Left Font Structure, Font Design and Typography” 49. See “DecoType” on page 410. 124 5/1/2002, 1:42 The notable exceptions to this tendency will be discussed later in the book.

Introduction

21

Thus evolved the basic convention to provide one character per letterform, supplemented by a selection of ligatures in which multiple letterforms were represented by a single character. The issues that arise from this approach ­became decisive for the evolution of Arabic type, and will be introduced in the following section. Beyond the fundamental script characteristics, the range of expression in Arabic writing merits consideration. The joining properties of the Arabic script, as well as the towering renown of the Islamic calligraphic tradition, often lead observers to confuse any Arabic writing with the art of calligraphy. By extension, discussions of printed Arabic frequently evoke a dichotomy of calligraphy and typography, whereby stereotypical notions of tradition, artistic expression and complexity are assigned to calligraphy, and modernity, sobriety and simplicity are assigned to typography. While some of these associations undoubtedly arise from actual features of the two modes of text production, their loose and inconsiderate use tends to blur the contours of the subject. Here, a deliberate distinction is proposed that aims to clarify and facilitate discussions surrounding the Arabic script. The spectrum of Arabic writing extends from casual handwriting, as it most mundane form, to the calligraphic art as the pinnacle of penmanship and artistry. Between these extremes, which are easiest to delineate, one finds a range of writing modes which might collectively be called ‘formal writing’. Handwriting, at the most basic end of the spectrum, describes any decipherable writing by a literate user of a language. While conforming to conventions and exhibiting a vast variety of writing skills, handwriting generally has no aspiration or claim to aesthetic merit. Formal writing, on the other hand, encompasses penmanship in correspondence, trade lettering, and the copying of manuscripts by professional scribes, amongst others. Incidentally, this middle ground of the writing spectrum is the most relevant for Arabic typography, and curiously it is almost entirely absent from discussions of the subject. The various expressions of formal writing are, in fact, the direct precursors of typography, as they all share one key aspect: their primary aim is to convey content through time and space to an anonymous audience. In order to achieve this, the writing needs to be highly conventional (to be decipherable by any reader), formally more accomplished than handwriting (to make the reader want to read it), and visually unobtrusive (for the form not to get in the way of the message). Calligraphy, by contrast, aims primarily for visual effect, with the content becoming a second priority. Calligraphic pieces are by and large not intended to be read, but to be looked at and appreciated for their beauty, in many ways akin to fine art in the Western world. Indeed, it is an art form also in the sense of

22

Chapter 1

being an end in itself: it is not employed to convey new content. Certified professional calligraphers rarely copied any books except for the Qurʾān – which pious people often know by heart anyway – and would consider it ­beneath their dignity to copy a mundane text. This distinction is reflected in the Arabic terms for the two writing professions, the kātib (scribe or clerk) is not to be confused with the ḫaṭṭāṭ (calligrapher).44 In conclusion, it should be noted that despite the substantial differences in these various modes of writing, they all share the same fundamental characteristics of the Arabic script. In principle the morphological rules are identical between a handwritten note, a shop sign and a calligraphic panel, notwithstanding the stark differences of execution. In other words, unless actual calligraphy as an art form is discussed, there is nothing calligraphic about the Arabic script. Furthermore, it follows that the use of the adjective calligraphic is largely meaningless in a typographical context and therefore avoided in the present book unless actual calligraphy is meant. The Typographic Representation of the Arabic Script From the earliest attempts at composing the Arabic script with movable type, adaptations and simplifications of the manuscript form were necessary to accommodate the characteristics of the technology. These adaptations were diverse, reflecting the printer’s and punchcutter’s knowledge of Arabic, their skills as craftspeople, the model the type was based on, what sizes of type had to be cut and crucially, the physical properties of metal type. In this process, the morphology of the Arabic script posed formidable challenges to the (initially only) European printers. When the Arabic script was adapted to the principles of movable type, generally following a static, one-to-one relation between letterform and sort, it lost the efficiency it had had in its dynamic manuscript form. The ease of creating semantic differentiation with minimal means through the use of diacritical signs and optional short vowels could not be replicated: Letters which are distinguished by diacritics had to be duplicated in all their forms. ‫  ت‬tāʾ, � ‫  ث‬ṯāʾ nominally feature only two distinct For example the letters �‫  ب‬bāʾ, � shapes, an identical rasm with different configurations of the rhombic dot. In movable type, three distinct sorts were required for the isolated forms, and a further nine for the representation of the principal joining forms of these let44

Note that the two terms do not even share the same three letter root, suggesting a significant semantic difference.

Introduction

23

Figure 1.4 The rasm of the letters �‫ ب‬bāʾ and �‫ �ي‬yāʾ is shared by numerous other letters. With an identical rasm, semantic differences can be generated easily by a change in the configuration of the diacritic marks. Lettering by the author.

‫ت ث‬ ters: �‫ت ‍� ث‬ ��‍ ‫ �‍ �‍ �‍  ‍ ب��‍ ‍ت��‍ ‍ث��‍ ‍��ب‬.‫ ب‬The rigidity of the one-to-one relation between letterform and type sort thus caused an exponential growth of the number of characters in Arabic founts. This was further exacerbated by the use of ligatures for the representation of context-specific letterforms. For example, the combination of �‫ ب‬bāʾ and �‫ �ي‬yāʾ is typically written with two distinct letterforms which are joined vertically, rather than horizontally (Figure 1.4). When reproduced in type, these two variant forms were commonly made as a ligature sort. However, if a combination of other letters with the same rasm was required, an entirely new sort, rather than a change in the configuration of Dubais was needed (Figure 1.5 overleaf). Thus, for all nominally possible combinations of letters sharing the same rasm as �‫ ب‬bāʾ and �‫�ي‬  yāʾ, a total of twelve ligatures ‫�ت �ث �ن �ئ‬ would have been required: ‫  �بي� ي� ي� ي� ي� ىي� �بى �تى �ثى �نى �ئى ىى‬.45 Consequently, Arabic foundry type became unwieldy, featuring hundreds of sorts, yet failing to render the script in its conventional form.46 This large number of characters was a major obstacle, substantially slowing composition and hindering an efficient use of Arabic type, not to mention the economic deterrent it implied. Moreover, the small size of vowels and other marks further complicated the typesetting process, as every line of type could consist of multiple levels (Figure 1.6 overleaf). Because the manufacturing processes and the physical restrictions of Western type technology were not adapted to the characteristics of the Arabic script, it made their reproduction practically unfeasible.47 45

This number is for Arabic only and would greatly increase if one considered Persian, Urdu and other languages written in the Arabic script.

46

Even the inclusion of hundreds of ligatures could not realistically cover all possible combinations, thus making it a structural problem of approach.

47

An attempt at rendering the Arabic script through typographical means without formal compromise is found in Ohannis Mühendisoğlu’s type from the end of the nineteenth century. Conceived to reproduce the manuscript form of the Ottoman Nasḫ style in

24

Chapter 1 Figure 1.5 Foundry type sort with a ligature of letters ‫ ك‬kāf and �‫ �ي‬yāʾ, e­ nlarged. Photograph by the author. Figure 1.6 Notional arrangement of ­vowel sorts in relation to other c­ haracters. Sketch by the author.

Moreover, the linear arrangement of characters along a static baseline which originated from Western print culture became an overarching principle of typo­graphy per se. The diagonally inclined joins of most Arabic writing styles could not be reproduced in foundry type, and had to be either discarded from the repertoire, or artificially flattened to conform to the horizontal alignment. Although punchcutters and printers tried to imitate manuscript forms, the characteristics of foundry type only provided little margin for manoeuvre, in effect prescribing how Arabic was represented typographically.48 By contrast, early European printing from foundry type successfully imitated the appearance of Latin manuscript forms, using means that had been developed for this very purpose.49 Here it is worth bearing in mind that the technology was not only more suited to the task, but that early European printers employed it in a way which sought utmost fidelity to manuscript conventions. Rather than providing the bare minimum alphabetic characters required for intelligible text, numerous alternate sorts with different letterforms, widths full, the fount attained a complexity that made manual composition too error-prone to be practicable. See Thomas Milo, “Arabic Script and Typography,” in Language, Culture, Type, ed. by John D. Berry (New York: ATypI Graphis, 2002), 112–27. 48

A well-documented example for such attempts is Franciscus Raphaelengius’ cutting of Arabic type. The manuscripts that served Raphaelengius as model exhibit smoke proofs of the punches he cut. Yet, the marked differences between type and manuscript are testimony to the difficulty of the formal translation. See Lane, Breugelmans, and Witkam, The Arabic Type Specimen of Franciscus Raphelengius’s Plantinian Printing Office.

49

Yasin Safadi’s statement that “towards the end of the sixteenth century the basic problems of Arabic typesetting were more or less solved” is grossly misleading and demonstrates the persistence of the erroneous assumption that Gutenberg’s method was suitable for Arabic. Safadi, “Printing in Arabic,” Monotype Recorder no. 2, New Series (October 1981): 4.

Introduction

25

and ligatures were made at a considerable cost in order to comply with readers’ expectations.50 For Arabic, as Ian Proudfoot observed, the situation differed markedly: No such initial concessions were made in Arabic typography. This is not surprising, since all the early typographic presses, in Europe, the Middle East and India, depended upon European technicians already accustomed to a strongly alphabetised type script, who no longer appreciated the need to make the same concessions to Arabic readers that Gutenberg had made to his European customers. In India and Southeast Asia the Christian missions were serious offenders. Early Muslim experiments, like that of Müteferriḳa or the early Būlāq press were more sympathetic, but still depended upon European technicians and faced constraints inherent in the technology.51 Arabic typography began as a compromise in which the method of reproduction dictated the appearance of the script. Arguably, the initial adaptation of manuscript Arabic to movable type was revolutionary, as it entailed the emergence of a typographic form without antetype in the script’s written form. Thus, no parallel to the continuity from manuscript practice to printing, as known from the European evolution of type, is found in the history of Arabic typography. As Shaw pointedly notes, early typographical developments in the Middle East and South and Southeast Asia “did not emerge ‘organically’ through the efforts of mother-tongue speakers from within their own aesthetic and calligraphic traditions or cultural norms of orthography and legibility”, but “were dominated from outside by European powers and their interests, both religious and secular”.52 Indeed, he argues that “this was technology transfer from Europe with very specific purposes in mind: to convert the heathen and impose foreign rule”.53

50

This observation was also made by Ian Proudfoot, an exception in a field that is largely agnostic of the aesthetic dimension of typography. See “Mass producing Houri’s moles, or Aesthetics and Choice of Technology in Early Muslim Book Printing”, in Islam: Essays On Scripture, Thought and Society, A Festschrift in Honour of Anthony H. Johns (Leiden: Brill, 1997), 161–189.

51

Proudfoot, “Mass producing Houri’s moles”, 171–172.

52

Graham Shaw, “Non-Latin Scripts and Printing Technologies: Triumphs and Tribulations,” in Non-Latin scripts, from metal to digital type (London: St Bride Library, 2012), 25.

53

Ibid.

26

Chapter 1

The form given to Arabic type did not change in principle when printing became accepted and spread in the Middle East, as much of this revolution was derived from Western inspiration.54 Although the knowledge of language and writing culture in print shops of the region improved the quality of Arabic type – as demonstrated by the work of the Egyptian Būlāq press – the fundamental principles of typographical composition remained unchanged.55 Typesetting equipment and the methods of movable type composition still imposed assumptions originating from the Latin script on Arabic type, making substantial compromises to the script’s morphology a necessity, and subsequently a characteristic feature.56 Arabic Publishing at the Turn of the Twentieth Century As indicated above, a growing body of research and literature is dedicated to the beginnings of printing with Arabic type in Europe, and its later, gradual adoption by Muslim populations in the eighteenth and nineteenth centuries. The Hungarian émigré and convert to Islam, İbrahim Müteferrika (c. 1670– 1747), is widely recognised for his contributions to printing in the Middle East.57 Müteferrika, considered as an exponent of the so-called Tulip Era – a period of reform and Westernisation in the Ottoman Empire – opened the first Muslim print shop in 1727. Yet, between 1729 and 1742 only 16 titles were published by Müteferrika and it is acknowledged that the small print runs and limited reach of his work meant that it fell short of a ‘media revolution’.58 According to Chris54

Geoffrey Roper’s account of Fāris Al-Shidyāq’s role as a pioneer of print culture in the Middle East during the second half of the nineteenth century demonstrates the influence of the European role-model. See Roper, “Fāris Al-Shidyāq and the Transition from Scribal to Print Culture in the Middle East.”

55

For a brief introduction to the Būlāq printing house see for example M.Y. Hammam, “History of Printing in Egypt”, Gutenberg-Jahrbuch 1951, 156-159.

56

The sustained rejection of typographically composed Arabic in the Middle East was probably related to the differences from its manuscript forms, and the generally inferior quality of type. The early acceptance of lithography, particularly in reproductions of the Qurʾān or Nastaʿlīq calligraphy, where high aesthetic norms are paramount, supports this thesis.

57

Orlin Sabev, “Waiting for Godot: The Formation of Ottoman Print Culture,” in Geoffrey Roper ed. Historical aspects of printing and publishing in languages of the Middle East, 101–120 (Leiden: Brill, 2014).

58

Christoph K. Neumann, “Book and Newspaper Printing in Turkish”, in Middle Eastern languages and the print revolution: a cross-cultural encounter (Westhofen: WVA-Verlag

Introduction

27

toph Neumann, until the 1860s only around 100 books per year were published in the Ottoman Empire and government presses with their limited repertoire of publications dominated the trade. Book printing thus remained a marginal phenomenon.59 Indeed, in view of the recorded 30,375 editions printed from movable type before 1501 in Western Europe, the notion that Müteferrika initiated a comparable shift in the Ottoman Empire seems difficult to maintain.60 The next important steps in the evolution of printing in the Middle East happened in Egypt. There, although nominally a vassal state of the Empire, printing and publishing enjoyed more favourable conditions. Muḥammad ʿAli (1769–1849) rose to power after the withdrawal of the French in 1805 and attained a critical role as moderniser of Egypt during his reign of more than four decades. The steps Muḥammad ʿAli initiated comprised far-reaching land reform, the transformation of the administration towards a modern bureaucracy, as well as the establishment of a more powerful army. Most relevant for the present subject, however, was his patronage of the publishing sector. Only four years into his reign, Muḥammad ʿAli sent an aide to Europe with the mission to collect books on a range of subjects, and in 1815 a student was sent to Milan in order to learn the printing and typefounding trades; at the same time printing presses were purchased from Italy. Upon his return in 1819, the foundation stone for the Egyptian Government Press was laid in the Cairene district Būlāq, which also lent its name to the printing house. By 1822 its first publication, an Arabic-Italian dictionary, was produced, and over the course of the coming decades Būlāq attained a leading role in the nascent printing trade of the region.61 By the end of Muḥammad ʿAli’s reign a few hundred books, often in translation, had been published, as well as countless typical government press productions: public notices, circulars, administrative documents and importantly, the first printed Arabic periodical. This relatively late beginning of serial publications in the Middle East is to be seen against the general state of the society, as pointed out by Ami Ayalon: In the year 1785, the Middle Eastern counterpart of The Daily Universal Register [from 1788 The Times of London] was a combination of the Skulima, 2002), 232. 59

Neumann, “Book and Newspaper Printing in Turkish”, 235.

60

“Incunabula Short Title Catalogue,” British Library, http://www.bl.uk/catalogues/istc/index.html (accessed 16 May 2017).

61

Ahmed Mansur, “The Bulaq Press Museum at the Bibliotheca Alexandrina,” in Geoffrey Roper ed. Historical aspects of printing and publishing in languages of the Middle East, 287–315 (Leiden: Brill, 2014).

28

Chapter 1

o­ fficial munadi [the caller employed by the authorities for announcing news in public spaces] […], the mosque preacher, and the exchange of intelligence in public gatherings. Newspapers did not appear in the Arab countries prior to the nineteenth century; they were unnecessary.62 Thus, the Jūrnāl al-Khidiw (The Khedive’s Jūrnāl), first published in 1821–22, assumed the historical role of the first printed Arabic periodical. Yet apart from its relatively regular publication, it should be noted that it had little in common with a newspaper: The Jurnal, a bilingual Turkish-Arabic bulletin, was little more than a domestic circular intended for official consumption. With a run as small as 100 copies, it was designed for no other purpose than to keep the vali himself and his chief aides informed of state affairs. Handwritten at first, it was subsequently printed lithographically, appearing irregularly for a while before it became a weekly and later a daily publication.63 Its successor, Waqāʾiʿ Miṣriyyä, (Egyptian Events) first published on 3 December 1828, followed a similar concept as the Jūrnāl. Despite its larger print-runs (600 per issue), this publication also remained far removed from the conventional notions of a newspaper: it was not sold to the public but distributed to a select state elite, it was published irregularly, the languages were often deficient, it featured few and often outdated news items, and spelling errors rendered the texts difficult to comprehend.64 In the meantime printing and publishing in the Ottoman Empire, in a slumber since Müteferrika, had resumed against a changed societal backdrop. From the seventeenth century the Ottoman Empire’s power waned and its decline became increasingly apparent with the rise of the European powers.65 A stagnant economy, repeated military defeats with losses of territory, and the increasingly palpable influence of Europe provided the impetus for an am-

62

Ami Ayalon, The Press in the Arab Middle East: a History (New York and Oxford: Oxford University Press, 1995) 6.

63

Ibid., 14. In relation to its title, Ayalon notes: “a name whose two parts accurately reflected the foreign, Euro-Ottoman roots of the idea”.

64

Ibid.

65

See Bernard Lewis, The emergence of modern Turkey (London & New York: Oxford University Press, 1961).

Introduction

29

bitious reform path.66 The Tanzimât era (1839–1876) entailed numerous and d­ iverse efforts of the Ottoman administration to modernise itself and to reform the Empire. Amongst them were the creation of an education system, the establishment of a conscript army, the introduction of an Ottoman citizenship irrespective of religion or ethnicity, and the abolition of slavery. Beyond such substantial changes, the reforms also involved more cosmetic changes such as the conception of a national flag and anthem; yet, what all of them had in common was the marked influence of the European nation state as model.67 At the same time, publishing endeavours throughout the empire increased. In 1830 the first Ottoman bulletin was founded by Sultan Mahmud II (1785– 1839) in apparent imitation of Muḥammad ʿAli’s Waqāʾiʿ Miṣriyyä. Yet, unlike its Egyptian model, Le Moniteur Ottoman was less a control tool than an outwards facing item of representation, conceived to further the image of the Ottoman Empire in the international arena. Its successor, the Taqvim-i Veqayi (Calendar of Events), however, demonstrates the change of attitude that was soon to unfold. As an official state mouthpiece, it aimed to counter the proliferation of serial publications throughout the Ottoman Empire which were perceived as a potential threat.68 By the end of the 1850s, however, it became apparent that a control organ alone could not stem the influx of information and dissenting voices. The continued increase of newspapers published in the languages of the Ottoman subjects, inside the Empire as well as abroad, led to the establishment of printing and publishing regulations in 1857, and an Administration of Press Affairs in 1862. While technically falling within the Tanzimât period, these regulations did not provide any liberalisation and have been described as “disproportionately – or preemptively – limiting for the embryonic Ottoman Press”.69 Indeed, in the last quarter of the century censorship and control of publishing and the Press in particular further intensified in the Ottoman Empire, a trend that only

66

The Crimean peninsula was lost to Russia in the war of 1768–74, the French occupation of Egypt and Syria in 1798 could only be ended with British help, in 1821 Greece gained independence, in 1830 Algeria was conquered by France and Muḥammad ʿAli of Egypt gravely threatened the integrity of the empire during the first Egyptian–Ottoman War (1831–33).

67

Jürgen Osterhammel, The Transformation of the World: A Global History of the Nineteenth Century (Princeton University Press, 2014), 627.

68

Ayalon, The Press in the Arab Middle East, 21–22.

69

İpek K. Yosmaoğlu, “Chasing the Printed Word: Press Censorship in the Ottoman Empire, 1876–1913” in The Turkish Studies Association Journal, 27, no. 1/2 (2003), 17–18.

30

Chapter 1

came to an end with the Young Turk revolution of 1908 and the beginning of the Second Constitutional Era.70 Meanwhile in Egypt, a more liberal political climate provided the ground for a different evolution of publishing in the second half of the nineteenth century. After a hiatus of 15 years under the rulers Khedive ʿAbbās and Khedive Saʿīd, the reforms initiated by Muḥammad ʿAli were resumed from 1863 by Khedive Ismaʿīl (1830–1895) and contributed to the formation of the Arabic cultural ‘awakening’ al-Nahḍa in Egypt. This expansion of literary and journalistic activity revolved around a circle of intellectuals – besides Cairo and Alexandria mainly in Beirut – and was primarily expressed through the medium of print. But as Ayalon points out, the Nahḍa entailed a transformation of the publishing culture that went well beyond the output of the handful of luminaries that most of the literature is concerned with.71 He cites 1,516 works published in Beirut and Mt Lebanon before 1900, as well as 9,538 titles issued in Egypt in the second half of the nineteenth century, and notes that “the appearance of thousands of printed works in millions of copies where nearly none had existed several decades earlier surely signifies a major cultural shift”.72 In addition to the key works which are seen as shaping modern Arabic thought and identity – translations of influential European books, editions of classical Arabic works, grammar and language treatises – countless ephemeral items such as business cards, posters, advertisements, train and theatre tickets contributed to the acquaintance of the people with printed products. Emphasis is often put on the surge in periodicals during this period, and their fundamental novelty for the Middle Eastern public: Newspapers and journals in particular represented a stark innovation: printed in a novel format and disposable after reading, they addressed issues previously beyond the horizon and offered interactive platforms for public exchange.73 Three exponents of the Nahḍa exemplify how the medium of print in general and the Press in particular were taken up as vehicles for public discourse 70

Ibid., 31–32.

71

See Albert Hourani, Arabic Thought in the Liberal Age 1789–1939 (London, New York, Toronto: Oxford University Press, 1962).

72

Ami Ayalon, “Private Publishing in the Nahḍa,” in International Journal of Middle East Studies, 40 (2008), 562–563.

73

Ibid., 563.

Introduction

31

and contributed to the cultural revolution of the Nahḍa. The Lebanese Fāris al-Šidyāq (1805–1887) is regarded as one of the first Arabic journalists for his work at the Waqāʾiʿ Miṣriyyä; in addition to contributions to the English Church Missionary Press in Malta, a Bible translation in Cambridge and independent writing and publishing endeavours in Paris, al-Šidyāq became most famous for the newspaper al-Ǧawāʾib that he published in Constantinople from 1861.74 Similarly, Ḫalīl Sarkis’ (1842–1915) is best known for his Lisān al-Ḥāl newspaper, published in Beirut from 1877, but besides his journalistic work Sarkis was an author and publisher. He established one of Beirut’s principal print shops, which produced numerous journals and hundreds of books and other publications.75 Ǧurǧī Zaīdān (1861–1914) is also considered a key contributor to the Nahḍa. A prolific writer, he published numerous works of history as well as 23 historical novels during his career. His literary journal al-Hilāl, started in 1892 and printed at his own press, became one of the most influential periodicals in the region and established Zaīdān as one of the pioneering journalists of the Arab world.76 For the development of a printing culture in the Middle East, the contribution of the Nahḍa was substantial. It engendered the establishment of institutions and practices which had not previously found their way into the societies of the region. Professions which had not previously existed in the Middle East now took root there: journalists, authors, printer-publishers, typesetters, proof-readers, booksellers, and others in the domain of print were either new occupations, or became established in their modern form during the second half of the nineteenth century. The current view on the Nahḍa, is summarised by Ayalon in unambiguous terms: It entailed massive production of written texts, the emergence of new diffusion channels carrying the printed products to the customers and, not least, the formation of sizeable and ever-expanding reading audience.77 The magnitude assigned to this transformation must, however, be seen primarily in relation to the preceding state of Middle Eastern societies. Although 74

Roper, “Fāris Al-Shidyāq and the Transition from Scribal to Print Culture in the Middle East,” 213.

75

Ayalon, “Private Publishing in the Nahḍa,” 561.

76

Ami Ayalon, “Arab Booksellers and Bookshops in the Age of Printing, 1850–1914,” British Journal of Middle Eastern Studies, 37, no. 1, (April 2010), 83.

77

Ayalon, “Arab Booksellers and Bookshops in the Age of Printing, 1850–1914,” 73.

32

Chapter 1

printed output multiplied during this era, it remained a fraction of that in the European incunabula period, and fell distinctly short of the consumption of printed products by South Asian Muslims in the nineteenth century.78 ­More­over, whilst literature on the printing revolution in the Middle East emphasises that books and newspapers were sometimes read out to more than one person, it is acknowledged that the extent of this consumption cannot be quantified, and one should hasten to add that it did not amount to widespread literacy. The fact remains that literacy rates in the Middle East were minuscule even by the turn of the century: in 1897 in Egypt, the cradle of the Nahḍa and the centre of Arabic publishing, merely 5.8 percent of the population was literate and similar figures are estimated for greater Syria.79 Print and the Press in particular therefore lacked a defining characteristic of the medium: there was no mass readership. Critically, it also means that the transformative processes of the Nahḍa were confined to a demographic sliver of Middle Eastern societies. By definition, it follows that printing was not yet established as a mass medium in the late nineteenth century. When the age of mass communication began in Europe with the first newspapers reaching a daily circulation of a million in the 1890s, the printing industry in the Middle East only just experienced its first birth pangs.80 As Dagmar Glass notes, this was not only because of the societal make-up: […] for technical reasons alone, the publication of daily papers was hardly possible around 1880, either in Beirut or in Cairo, let alone other cities

78

Francis Robinson’s figures about printing by Muslim populations on the Indian subcontinent tempers the significance of the Middle Eastern print revolution in the nineteenth century. “Technology and Religious Change: Islam and the Impact of Print.” ­Modern Asian Studies 27, no. 1 (February 1993): 233.

79

Egyptian literacy rate from Ayalon, The Press in the Arab Middle East, 143. Ayalon notes that male literacy rates were higher in cities with 21 percent in Cairo and 19 percent in Alexandria. Female literacy rates were negligible throughout the region. Syrian literacy rate estimate from Donald J. Cioeta, “Ottoman Censorship in Lebanon and Syria, 18761908,” International Journal of Middle East Studies, 10, no. 2 (May, 1979), 180.

80

The Daily Mail was the first paper that reached a circulation of one million in the United Kingdom, with a French paper following soon after. Michael Twyman, Printing 1770–1970. An Illustrated History of Its Development and Uses in England (London: Eyre & Spottiswoode, 1970), 51; Eric Hobsbawm, The Age of Empire: 1875–1914 (London: Vintage, 1989), 53.

Introduction

33

of the Arab world. The printing presses first installed there could usually produce, at the most, one to three copies per minute.81 Contemporaries, it appears, were keenly aware of the infancy of the trade. The eminent Zaīdān observed in 1895 that “the Arab press, compared with its Western counterpart, is but a newborn baby just opening his eyes before a very old, worldly-wise and experienced man,” wondering whether “this baby should aspire to attain such goals as the old man had not achieved except after decades of trial and error?”82 The Egyptian al-Ahram newspaper, one of the pioneering and most advanced publications in the region, is a case in point. It was founded as a weekly in 1875 (first edition from 5 August 1876) by the Lebanese brothers Salīm (1849– 1892) and Bishāra (1853–1902) Taqlā, and published daily from January 1881.83 Yet, despite its leading role, the reach of this newspaper was modest until well into the twentieth century. Its circulation numbers remained a fraction of large European newspapers, reaching approximately 2,775 in 1892, 3,600 in 1898 and 13,000 in the early years of the twentieth century.84 According to some estimates the total circulation of all newspapers in Egypt, then the country with the most established publishing industry, reached only between 80,000 and 100,000 copies by the eve of the First World War.85 Only with the arrival of typesetting machines and powered printing presses could the circulation of the Arabic Press aspire to attain levels worthy of a mass medium. Even despite such technological advances and profound changes in the societies of the Middle East, the next leap in the evolution of the printing and publishing trades only happened three to four decades later, by the end of the colonial empires.86 81

Dagmar Glass, “Arabic Newspapers and Periodicals in the Arab World (1828–1928),” in Middle Eastern languages and the print revolution: a cross-cultural encounter: a catalogue and companion to the exhibition (Westhofen: WVA-Verlag Skulima, 2002), 211.

82

Ayalon, The Press in the Arab Middle East, 243.

83

Ibid., 42–43. Al-Ahrām, (the pyramids).

84

Glass, “Arabic Newspapers and Periodicals in the Arab World (1828–1928),” 215.

85

Ayalon, The Press in the Arab Middle East, 152. Ayalon emphasises the difficulty in obtaining reliable circulation numbers for this period.

86

Safadi relates changes in the printing trade to the “growing independence of the Arab countries in the early decades of the 20th century”. Safadi, “Printing in Arabic,” 6. But as the early mechanisation of Arabic typography predated decolonisation by decades, this does not seem plausible. In the early twentieth century European colonial powers remained well established in the Middle East, and only after the Second World War with the dismantling of the colonial empires such a causal link appears more sustainable.

34

Chapter 1

As a new institution, the Arab press played a modest role in the region’s history during this period [1895–1945], although at certain times and places it had an important impact on political and social developments, and at least in one sphere, the modernization of literary norms and of the language, its contribution was pivotal.87 Although the printing and publishing trades were pioneered in the nineteenth century, at this point the potential readership, the number of publications and their reach were still limited. Particularly newspapers, which often served as catalysts for developments in typographical practice, had remained embryonic at best. Whereas printing and typography had been practised in Europe during four hundred years and evolved into a number of specialised professions which wholly displaced scribal culture, the written Arabic word was generally still read in its manuscript form in the nineteenth century.88 Even when taking the century’s second half as the Arabic incunabula period, maturity of the printing and related trades was only reached in the twentieth century. This is a crucial realisation, for the infancy of the trade would have made it particularly vulnerable to the novel concepts of machine composition. One can surmise that the lack of conventions and traditions of an established trade, coupled with a reading public unaccustomed to the specificities of foundry type, provided printers and publishers with no grounds for resisting mechanically composed Arabic. Arguably, then, printing with Arabic type achieved its lasting and far-reaching significance at a time when mechanical typesetting came to fruition and replaced both the preeminent scribal culture of text production and manual composition of type. Arabic typography thus only became a domain in its own right when mechanisation redefined all processes related to it. Consequently, typesetting machinery attained unparalleled influence on the visual form of the Arabic script, an influence that went well beyond the impact that is recognised in the history of Latin type-making. Based on this understanding, the present book seeks to shed some light on the evolution of Arabic type in the machine age.

87

Ibid., 243.

88

Note that also a significant proportion of Arabic script books printed in the nineteenth century were lithographic copies of manuscripts.

Chapter 2

Mechanical Composition of Arabic The Industrialisation of Printing Until the beginning of the nineteenth century, the techniques of printing had substantially remained unchanged since the days of Gutenberg. In 1800, the primary processes employed in European print shops differed little from those used in the preceding 350 years. Paper was still made by hand and the processes of type-making – punchcutting, striking of matrices and casting type – as well as those of typesetting and printing had changed little since the fifteenth century (Figure 2.1).1

Figure 2.1 Punch and corresponding matrix, as used in manual typemaking. From de Vinne, The Practice of Typography.

But throughout the nineteenth century, industrialisation affected the printing and related trades of industrialised countries. What was a craft at the beginning of the century transformed into an industry at its closing. This gradual mechanisation was often causally related, with one improvement opening the door to innovation in another domain as “technical progress, rationalised organisation and compulsory education interacted one on another”.2 Mechanical paper-making was a prerequisite for powered printing presses, improved distribution networks a necessity for increasing newspaper circulation, whilst general electrification facilitated all the tasks involved. The printing trade, a field that had been remarkably reticent to change, transformed fundamentally 1 Michael Twyman, Printing 1770–1970. An Illustrated History of Its Development and Uses in England (London: Eyre & Spottiswoode, 1970), 48. 2 S. H. Steinberg, Five Hundred Years of Printing, rev. ed. by John Trevitt, (London & New Castle, DE: The British Library & Oak Knoll Press, 1996), 136.

36

Chapter 2

within a few generations.3 Its principal catalyst was the newspaper industry: The Times of London became a leading sponsor of improvements in printing machinery in response to its expanding readership.4 In 1814 it adopted the steam-powered Koenig press, increasing the number of sheets printed per hour from 300 to 1,100.5 In lockstep with improvements in machinery and mechanisation of labour came numerous inventions that were aimed at accelerating printing and reducing costs. Stereotyping, pioneered in the eighteenth century, became commercially viable through the Frenchman Jean-Baptiste Genoux’s use of papier mâché for the creation of flongs in 1829.6 Electrotyping, an electro-chemical process for replicating a model in metal, was invented in Russia in 1839 by the German engineer Moritz von Jacobi.7 It soon entered the printing trade and was put to use for copying metal type without recourse to punchcutting and striking of matrices.8 The process lent itself to producing matrices that had to be exact reproductions of existing type where the original punch had become defective or gone missing. Also, as Richard Southall points out, this technique

3 “As late as 1772 the Basel printers caused the city authorities to ban the use of an improved handpress because its inventor was not a professional printer” Ibid., 137. 4 Daily circulation rose from around “1,500 in the eighteenth century to 11,000 in 1830, 38,000 in 1850, and 70,000 in 1870. Before this last date it had been overtaken by the Daily Telegraph which had a circulation of over 100,000 in the 1860s and twice that number in the 1870s”. Twyman, Printing 1770–1970, 51. 5 This number was gradually increased to 4,000 in 1828, 8,000 in 1848 and 20,000 by 1868. Steinberg, Five Hundred Years of Printing, 192. 6 Twyman, Printing 1770–1970, 54. 7 Herbert Heinrich, “The Discovery of Galvanoplasty and Electrotyping,” Journal of Chemical Education 15, no. 12 (December 1938): 566–75. 8 “The model letters are cut on type-metal, and, after preparation, are suspended in a porcelain-lined jar containing a solution of sulphate of copper in connection with a voltaic battery. The chemical action created in the battery cells causes an electric current that liberates atoms of copper from the solution, which adhere to the suspended model letters. When the desired deposit is obtained the letters are taken from the solution and their thick shells of copper removed. These shells are reinforced by brass and are converted by the fitter-up into movable matrices. Matrices can be made by the electrotype process from models in type-metal or from cast type as readily as from punches of steel”. Theodore Low De Vinne, The Practice of Typography. A Treatise on the Processes of Typemaking, the Point System, the Names, Sizes, Styles and Prices of Plain Printing Types (New York: The Century Co., 1902), 18–19.

Mechanical Composition of Arabic

37

was the preferred method employed by type founders “when they had occasion to copy or steal one another’s designs”.9 The casting of type, which had remained virtually unchanged since Gutenberg’s day, saw the first tentative steps towards automation in the nineteenth century. For the first half of the century traditional hand-casting of type, with its output of between 400 and 500 sorts an hour, remained the norm, but in 1838 the first successful type-casting machine was conceived and put to use.10 David Bruce’s Pivotal had an output of 6,000 sorts per hour and was rapidly taken up and employed in the United States and Germany, and arrived ten years later in the United Kingdom.11 Demand for large quantities of type came mainly from the expanding newspapers, and improvements in casters led to the most potent machine of its kind. The rotary caster patented in 1881 by Frederick Wicks boasted 100 moulds and recorded up to 60,000 sorts an hour.12 The one process that withstood numerous attempts at mechanisation for the longest part of the nineteenth century was the composition of text. Although the earliest recorded invention of a device to mechanically set type dates from 1822, it was “not until 1885 that a genuine labor-saving composing machine was offered to the printing world” – the Linotype.13 During the course of the century hundreds of patents were taken out for machines that sought to facilitate this most complex of processes employed in the print shop, but few were ever developed and successfully put into practice. As late as 1874, eminent printer and historian Theodore Low de Vinne is on record as concluding a review of typesetting machines with this cautious assessment: A careful review of what has been done, and of what is promised in the way of typesetting machinery, gives us no reason to believe that compositors will ever go out of fashion, or that automatic machinery will supplant brains. The man must be the master of the machine.14 9

Richard Southall, Printer’s Type in the Twentieth Century: Manufacturing and Design Methods (London & New Castle, DE: The British Library & Oak Knoll Press, 2005), 27.

10

The earliest known caster was patented in 1822 by William Church, UK Patent 4,664. John Smith Thompson, History of Composing Machines (Chicago: The Inland Printer Company, 1904).

11

Twyman, Printing 1770–1970.

12

Thompson, History of Composing Machines, 77. See his account for a comprehensive history of composing machines in the nineteenth century.

13

Ibid., 1.

14

Quoted in Ibid., 3.

38

Chapter 2

Robert Hattersley’s machine was the earliest successful typesetting machine, first installed in 1866 at the Eastern Morning News in Hull, Yorkshire.15 It continued to be used to some degree in newspaper composing rooms in the United Kingdom up to 1915. In 1872, The Times of London installed a Kastenbein typesetting machine which remained in use until 1909.16 However, two main disadvantages of these early typesetting machines prevented their widespread success and their acceptance outside the newspaper industry. Justification of lines and redistribution of sorts were not mechanised and constituted expensive and time-consuming tasks. To accomplish the latter, low-paid juvenile labour was the only economical solution, and “the opposition of the London Society of Compositors to the employment of unskilled workers explains the restriction of the Hattersley machines to the provinces as well as the exclusive use of the Kastenbein machine by The Times, for The Times was the only nonunion newspaper in London”.17 With mechanised type-casting successfully implemented towards the end of the century, the question of justification of lines became the most challenging hurdle for further increases in the speed of composition. In the 1880s, two distinct and initially unrelated inventions resulted from the quest to fully mechanise all aspects of typesetting; their combination eventually made possible a machine that revolutionised the composing room. Linn Boyd Benton, a type-founder in Milwaukee, Wisconsin, commenced work on an automatic justifying machine in 1882.18 Part of the solution he proposed was the first unit-width system employed to limit and rationalise the widths of type sorts. Hitherto, the width of foundry type sorts could be freely chosen, as punches, matrices and type bodies were individually adjusted to provide the best possible proportions of a letter. Consequently, the thickness of different sorts was irregular, making the compositor’s task of justifying a line a difficult endeavour.19 Benton’s invention came to be called Self-Spacing 15

Twyman, Printing 1770–1970, 61.

16

Lawrence W. Wallis, Typomania (Upton-upon-Severn: Severnside Printers Limited, 1993), 67.

17

Steinberg, Five Hundred Years of Printing, 145. Note that the London Society of Compositors union was also strongly opposed to female labour and prevented the employment of women for work on the distributing machines at the Daily News. Ellic Howe & Harold E. Waite, The London Society of Compositors (re-established 1848): a centenary history (London, Toronto, Sydney, Melbourne: Cassell & Co., Ltd, 1948) 229–230.

18

Patricia A. Cost, The Bentons: How an American Father and Son Changed the Printing Industry (Rochester: Cary Graphic Arts Press, 2011), 47.

19

“Justifying, which like many printer’s terms carries different meanings in different con-

Mechanical Composition of Arabic

39

Type, and it was granted a patent in 1883.20 It was based on the principle of assigning multiples of standard width-units to all sorts of the fount. As with related inventions discussed above, Benton’s type was primarily targeted at the newspaper industry, which had the most demanding requirements of speed in the printing trade. It was claimed that the higher price of such founts quickly proved economical for the newspapers, as the increases in composing speeds saved many times the investment in the type. For the successful commercialisation of this new type, Benton required too large a number of punches to be manufactured manually.21 He therefore set out to develop a mechanical device to assist him with the task. By 1885, he was granted a patent for the pantographic Punch Cutting Machine (Figure 2.2 overleaf).22 A pantograph is an instrument for reproducing patterns at different scales, featuring a tracing point which is linked to a writing arm through pivoted levers. The tracer’s movement is thus conveyed to the writing arm, mechanically scaling the pattern it describes. Conceptually the pantograph was not new and similar instruments had been used before, notably to cut wood type. But Benton’s device was revolutionary in that it allowed the precision necessary for cutting steel punches at text sizes, and their exact multiplication in unlimited numbers.23 The rapid reproduction of identical punches was texts, means in this context the filling out of the line of type to a fixed length with spacing material, so that the composed type can be handled without becoming jumbled or ‘pied’, and printed without individual types being drawn out of the forme”. Richard Southall, “A Survey of Type Design Techniques before 1978,” Typography Papers no. 2 (1997): 42. 20

Benton, Linn Boyd. Printing-Type. US Patent 290,201, filed 8 May 1883, and issued 18 December 1883. The majority of sorts employed four units (a, b, d, h, k, n, o, p, q, u, v, x, y), the narrowest letters were on two units (f, i, j, l and punctuation), the widest were given five (w), six (m), or seven units (W), and the remainder had a three unit width. As Cost points out, these types of Benton attracted sustained criticism from various quarters for their compromised aesthetics and poor readability. Cost, The Bentons.

21

Henry Lewis Bullen estimates that there was an immediate need for more than 3,000 punches. “Linn Boyd Benton – The Man and His Work,” The Inland Printer 70, no. 1 (October 1922).

22

Benton, Linn Boyd. Punch Cutting Machine. US Patent 332,990, filed 29 February 1884, and issued 22 December 1885.

23

As Southall remarked, “it is impossible to exaggerate the importance of Benton’s invention for the development of type design techniques.” Southall, “A Survey of Type Design Techniques before 1978,” 43. For an account of historic precursors of the pantograph see Linotype & Machinery Ltd, “From the Stork’s Beak to Benton’s Punchcutter,” Linotype Matrix II, no. 33 (May 1960): 2.

40

Figure 2.2

Chapter 2

Punch Cutting Machine, patent drawing, reduced to 75% linear. From Benton, Linn Boyd. Punch Cutting Machine. US Patent 332,990.

Mechanical Composition of Arabic

41

novel, as manual copying of punches could not provide exact replicas and took significantly longer.24 Initially conceived for the making of electrotyped matrices for the Benton, Waldo & Co. type-foundry, the machine broke new ground when it was employed to cut punches for the multiplication of matrices. In that role Benton’s machine paved the way for the success of another invention that emerged at about the same time in Baltimore, Maryland. In the 1870s, Ottmar Mergenthaler, a young German émigré working as a watchmaker and model builder for the firm of August Hahl in Baltimore, became involved with attempts at developing new type writer machinery.25 Although no commercially viable proposition developed from these early experiments, they led the young engineer to discover the needs of the evolving printing trade. In 1883 Mergenthaler established his own machine shop and began investigating mechanical typesetting and justification. Backed by investors, he developed a series of devices with varying degrees of success. By 1885 Mergenthaler’s experiments had eliminated most conceptual problems and he was able to demonstrate a machine that upon depression of keys selected corresponding matrices, justified them automatically, and cast a solid line of fresh type from them – thus the principle of the linecasting machine was born. Whilst attracting further investors, Mergenthaler continued improving his invention until the first machine, the so-called Blower Linotype, was ready to be deployed in practice at the New York Tribune in summer 1886 (Figure 2.3 overleaf). On 3 July, the first page composed on a Linotype linecaster was printed, heralding the era of mechanised typesetting. Yet, necessary improvements on the machine, combined with quarrels between the inventor and the investors caused another delay of four years until the principles of the Linotype were established. Among the complications Mergenthaler experienced was the issue of matrix manufacture, as told in his own account: One great difficulty in the matrix department was found to be the expense and trouble of maintaining the original steel stamps [punches] which produced the matrix proper. There was no machine existing at that time by means of which these stamps could be engraved at a small 24

For a detailed description of the pantographic punchcutter’s effect on the design and manufacture of printing types see Southall, Printer’s Type in the Twentieth Century, 19–34.

25

For accounts of Mergenthaler’s invention and life see Carl Schlesinger, The Biography of Ottmar Mergenthaler (New Castle, Delaware: Oak Knoll Books, 1989); Basil Kahan, Ottmar Mergenthaler: The Man and His Machine: a Biographical Appreciation of the Inventor on His Centennial (New Castle, Delaware: Oak Knoll Press, 2000) and recently Frank Romano, History of the Linotype Company (Rochester Institute of Technology, RIT Press, 2014).

42

Chapter 2

Figure 2.3

Left: The first commercially available Linotype, as used in the New York Tribune. From Legros and Grant, Typographical Printing-Surfaces.

Figure 2.4

Right: Schematic illustration of a composed line of single letter matrices, demonstrating the justification by means of ‘spacebands’. From Legros and Grant, Typographical Printing-Surfaces.

cost with an absolute certainty of maintaining the same size and shape, but they had to be engraved by hand at a cost of $ 5 per piece.26 It was at this point that Benton’s pantographic punchcutting machine attained a role for which it had not initially been conceived, providing “the final link in the chain of processes which made the composing machine a practical possibility”.27 At the heart of the Linotype mechanism for composition were 1,500 brass matrices which had to be struck from punches – a colossal task when done manually. The pantograph not only enabled the exact and fast reproduction of punches, but the resulting matrices were also of a superior technical quality to those made from hand-cut punches. The pivotal role of the pantographic punchcutting machine for Mergenthaler’s invention was already recognised by de Vinne in 1902: 26

Quoted in Schlesinger, The Biography of Ottmar Mergenthaler, 29.

27

Linotype & Machinery Ltd, “From the Stork’s Beak to Benton’s Punchcutter,” 2.

Mechanical Composition of Arabic

43

The success of the Linotype [...] machine is largely due to the accuracy of the matrices made from Benton machine punches. As the counters are deeper and the bevels truer, the types do not show distortion when they have been flattened by wear.28 Indeed, when a mature version of the Linotype was released in 1892 (the Model 1 which formed the basis for successive machine generations), it was swiftly taken up by the trade, and within the decade it had revolutionised the publishing sector. The Linotype presented in a single machine the solutions to various aspects of typographical composition that had hitherto been treated separately. Text could be rapidly composed by means of a keyboard which released individual matrices from their compartments in a magazine and assembled them in the order of typing. Whereas experienced hand compositors were able to set around 2,000 characters per hour, a Linotype could master an average 6,000.29 The operator saw the matrices assemble in view and controlled their automatic and continuously variable justification by means of wedgeshaped “spacebands” (Figure 2.4). Once completed, the obtained mould was secured and entire lines of type were freshly cast from liquid type metal – from which the name for the entire process, hot-metal, derives. The Linotype’s mechanism then distributed the matrices back to their positions within the magazine, thus eliminating the laborious task of type distribution (Figure 2.5 overleaf). The cast lines (or “slugs”) were trimmed by knives and assembled in a galley, ready for page make-up. After the printing was done the type could be melted down and recycled for further use. It was a complete system, and continuous technical improvements contributed to establishing the Linotype as a dominant force in the trade (Figure 2.6 overleaf). By 1895, 2,549 linecasters were in operation in the United States and had already replaced approximately 12,000 hand compositors; and whilst at this stage only 250 Linotypes had been installed in the United Kingdom, their potential was sufficiently apparent for the trade. Whereas the London Society of Compositors initially mounted resistance against the introduction of typesetting machinery and the employment of unskilled or non-unionised labour, by 1895 its Executive Committee advised its members not to fight the

28

De Vinne, The Practice of Typography, 353.

29

John W. Seybold, Fundamentals of Modern Composition (Media, Pennsylvania: Seybold Publications Inc., 1977), 35. The average speed of operators was approximately 5,000 ems per hour, with recorded peaks of 14,000 ems. Thompson, History of Composing Machines.

44

Chapter 2

Figure 2.5 Diagram of the progressive positions of the matrices travelling through the machine. From The Linotype Manual, (London: Linotype & Machinery Ltd, 1964).

i­ ntroduction of machine composition as it was recognised that the mechanisation of the composing room could no longer be held up: In voting upon this question, however, the members will do well to bear in mind that an adverse vote will probably place them­—so far as public sympathy is concerned—in the undesirable position of attempting to fight machinery. Time after time charges of hostility towards the machines have been levelled against the executive and the members, and as often refuted; but the experience of other industries has clearly demonstrated that wherever machinery has been introduced in which was embodied the elements tending towards success, time has eventually proved its success, despite the opposition levelled against it. In the case

Figure 2.6 Illustration of a duplexed matrix showing the regular and the italic ‘A’ of a typeface. From Linotype Machine Principles, (Brooklyn, New York: Mergenthaler Linotype Company, 1940).

Mechanical Composition of Arabic

45

of composing machines, none are perfect, although the strides towards perfection during the past few years cannot and must not be ignored.30 In 1895 alone, 987 new machines were manufactured by Mergenthaler in Brooklyn, half of which were immediately sold, and at the turn of the century Linotypes were produced in factories in Canada, the United Kingdom, and Germany.31 With the Industrial Revolution incrementally spreading beyond its initial heartlands, the Linotype followed in close succession and found use across the world. As a consequence demand arose for the mechanical composition of languages and scripts Mergenthaler could not have foreseen when inventing the device. As Ross pointed out, “there existed considerable disadvantages for non-Latin typesetting, for which purpose the composing machines had never been designed”, and the results of adapting those scripts to the limitations of the mechanical typesetters “ultimately redefined what was to become acceptable as legible typography”.32 Arguably, for many scripts the mechanical typesetting revolution was thus two-fold. Beyond the dramatic increase in speed and efficiency, and the causally related upheaval of the printing and publishing trades (which in the West are associated with the rise of the Linotype), for many writing cultures mechanisation also entailed the invention of a printed form without antecedent in its manuscript form. The Beginning of Arabic Mechanical Composition Following the establishment of the basic design of the Linotype, Arabic was one of the first non-Latin scripts to be adopted for Linotype composition – a widely acknowledged fact which has, however, received little scholarly attention (­Figure 2.7 overleaf). Although of considerable import, ushering in the era of mass production of the printed Arabic word, the precise date of the first mechanical composition of Arabic has remained elusive. Until recently, no study had sought to precisely date and locate the beginnings of Arabic mechanical 30

The pragmatic attitude of the Executive Committee proved to be appropriate, as “although the widespread introduction of composing machinery, and particularly of the linotype [sic], caused the displacement of hand-compositors, the effect was only temporary. The constant expansion of the printing trade, both in and out of London, speedily absorbed any surplus of men, and many more in addition.” Howe & Waite, The London Society of Compositors (re-established 1848), 234–235.

31

Romano, History of the Linotype Company, 309.

32

Ross, The Printed Bengali Character and Its Evolution, 135.

46

Chapter 2

Magazines

Side-case

Keyboard

Figure 2.7

Arabic Linotype machine as built in the 1910s. The keyboard of this model had 180 keys to give access to the two magazines that were required for a single Arabic fount. From Legros and Grant, Typographical Printing-Surfaces, 529.

typography, and in the few cases where literature provides dates and origins it frequently appears to repeat, rather than scrutinise claims.33 This hazy historical picture is illustrated in the following discussion, summarising the conflicting dates found in literature and primary sources. At the beginning of the twentieth century there were three successful typesetting machines that were widely used: the Linotype, the Intertype, and the Monotype system. Seeking to identify the first Arabic composing machine, the latter can be excluded, as Monotype only began to manufacture Arabic 33

The author’s PhD research covered significant ground in establishing a plausible account for this aspect of Arabic typographical history, but did not provide conclusive evidence. Titus Nemeth, “Arabic type-making in the machine age: the influence of technology on the form of printed Arabic type 1908–1993,” PhD thesis, University of Reading, 2013.

Mechanical Composition of Arabic

47

type from the late 1930s.34 Mergenthaler Linotype’s main competitor was the I­ ntertype Corporation, and could in principle have developed the first Arabic linecaster.35 Indeed, a relatively recent study locates the beginning of Arabic “machine setting” in 1910 at the Egyptian newspaper al-Ahram.36 If al-Ahram really pioneered Arabic “machine setting”, this would imply Intertype as the originator, for the newspaper primarily used Intertype machinery until the late 1960s. However, Hišām Baḥarī, al-Ahram’s long-time technical manager recorded in his history of the newspaper that it adopted Arabic typesetting machinery (sourced from Intertype) only in 1936.37 According to Baḥarī, al-Ahram initially used composing machines only for English and French matter, and developed an Arabic machine with Intertype from 1932. This date is supported by evidence of earlier printed issues of the newspaper being set with foundry type, and by Intertype Ltd’s trade journal Interludes, which reported in 1933: Al-Ahram, the most important Arabic newspaper in the Near East, was also the first Arabic newspaper to use composing machines. The fact that further Intertypes equipped for Arabic composition are now being installed is another tribute to their adaptability for the most diversified forms of composition.38 34

The Monotype Corporation only began to produce Arabic founts in 1938, when ample evidence of earlier mechanically composed Arabic typography exists.

35

The Mergenthaler Linotype Company (Mergenthaler) was located in the United States. Its subsidiary Linotype & Machinery Ltd (Linotype) was a British company formed from the Linotype Company and the Machinery Trust Ltd in 1903. It operated largely independently of Mergenthaler. The Intertype company was founded in 1911 (as the International Typesetting Machine Company) and began distributing its linecasting machines commercially from 1913. Yet, in 1913 only one Intertype machine was sold, and before 1917 the total number of units sold (mostly in the US) did not exceed 1000. In 1916 the company was discontinued and sold to new owners who changed the name to Intertype Corporation. Lawrence Wallis, A Concise Chronology of Typesetting Developments 1886–1986 (London: The Wynkyn de Worde Society in association with Lund Humphries, 1988), 10–11. The Intertype Company Ltd, the British subsidiary of the US Corporation, was founded in 1921. Romano, History of the Linotype Company, 129.

36

Dagmar Glass, ‘Arabic Newspapers and Periodicals in the Arab World (1828–1928)’, in Middle Eastern Languages and the Print Revolution, 212. According to Romano, the first Linotype was imported to Egypt in 1906 for the French language newspaper Bourse Egyptiènne, which thus introduced mechanical composition to the country. Romano, History of the Linotype Company, 72.

37

Hišām Baḥarī, Ṣaḥāfa Al-Ġadu (Cairo: Dār al-Maʿārif Bimiṣr, 1968), 138.

38

Intertype Ltd, ‘With the Editor’, Interludes III, no. 2 (April 1933): 9.

48

Chapter 2

It follows that any Arabic machine composition before the 1930s could only have been done with a Linotype machine, as suggested by the account of ­Walter Tracy, the long-time typographical adviser of Linotype. According to Tracy, who was instrumental in developing Linotype’s Arabic library after the Second World War, the company “began to manufacture matrices for the Arabic language in 1911”, referring to “records of supplying Altrincham-­built Linotype machines to Turkey [sic] during the period 1911 to 1914”.39 Corroborating evidence for the earliest use of Arabic typesetting machines is found in the autobiography of Reginald Orcutt, a travelling representative of Mergenthaler, who reported from his visit to Constantinople in 1928: […] Composing machines were strange and exotic novelties in Turkey. To be truthful, the only two in the country at the time (they had long been in disuse and were beyond repair) were early English Arabic Linotypes which a predecessor of mine named Jack Banks had sold to the then progressive public printer before World War I.40 Although confirming that Arabic Linotypes were found in Turkey in 1928, the anecdotal nature of Orcutt’s account does not make it a reliable source.41 Meanwhile, the accounts cited above are contradicted by Chauncey Griffith, Mergenthaler’s long-time director of typographical development. In an internal memorandum to his successor Jackson Burke, Griffith wrote in 1954: Prior to (circa) 1940 all Oriental, and particularly Arabic, typographic development was centered in Brooklyn, which made it possible to organize a group of experts, recognized scholars in their respective fields, for advice and consultation on every phase of this branch of our development. 39

Walter Tracy to Merryl Cook, 23 October 1972, WT correspondence, folder 18, DTGC. Tracy may be the source of a frequently quoted date: the year 1911 can be found in later specimens by the company and articles by Hrant Gabeyan, ‘Modern Developments in Arabic Typesetting’, in Middle Eastern Languages and the Print Revolution: a Cross-Cultural Encounter: a Catalogue and Companion to the Exhibition (Westhofen: WVA-Verlag Skulima, 2002), 216–21; and Yasin H. Safadi, ‘Printing in Arabic’, Monotype Recorder, New Series, no. 2 (October 1981): 2–7. According to Romano, the first Linotype was shipped to Constantinople in October 1910 for a French newspaper. Romano, History of the Linotype Company, 72.

40

Reginald Orcutt, Merchant of Alphabets (London: Jarrolds Ltd, 1947), 71.

41

Orcutt’s assertion that the machines were particularly old is curious, as 15 to 20 years was well within the life-span of a Linotype machine.

Mechanical Composition of Arabic

49

In later years L. & M. [Linotype] considered it more advantageous to a­ ssume the initiative in Arabic and Hebrew development, since the bulk of Linotype business in these languages originated in countries served directly by them through their local representatives.42 From the same document it emerges that “the first [Arabic] Linotype font was developed C. 1912 to be accommodated in the two-magazine Model 4 mixer Linotype”; and in another article Griffith dated the “introduction of Arabic” to 1913, “followed by Syriac, Armenian, Hebrew, Turkish, and other related scripts.”43 Both accounts suggest, however, that Arabic Linotype composition was first achieved in the United States. Although the above sources provide numerous clues, hitherto their contradictions precluded the development of an unambiguous, historically accurate image. New evidence found for this book, however, reconciles some of the accounts, rectifies the erroneous portrayals, and establishes the origin of Arabic mechanical typesetting in the United States. In October 1911, The Linotype Bulletin, a trade journal published by Mergenthaler, reported that “the first machine in the world to set Arabic characters has been installed in the office of Al-Hoda (The Guidance), a Syrian daily published in New York City.”44 Moreover, Mergenthaler company records and a 1929 specimen book identify the source for its first Arabic typeface as foundry type that had been provided in 1908 by a certain S. A. Mokarzel.45 While largely unfamiliar to historians of printing and typography, Salloum Antoun Mokarzel (1881–1952) is well known among scholars of the Arab-American ­diaspora (Figure 2.8 overleaf).46 He was the brother of Naoum Antoun ­Mokarzel (1864– 42

Chauncey Griffith to Jackson Burke, ‘Re: Arabic L.&M. Letter of September 3, 1954’, Internal Memorandum, (14 September 1954), Box P3640, NMAH. Ross also demonstrated that the British company did not regard itself qualified to undertake any conceptual work on ‘Oriental’ language developments as late as the 1930s, relying entirely on the expertise of various consultants of Mergenthaler. Ross, The Printed Bengali Character and Its Evolution, 139.

43

Chauncey Griffith, ‘The Linotype Development of Type Faces’, in The Annual of Book Making 1927 · 1937 (New York: Colophon, 1938), no page numbers.

44

Mergenthaler Linotype Co., “Arabic on the Linotype,” The Linotype Bulletin VII, no. 10, (October 1911): 81.

45

Mergenthaler-Linotype Co., ‘Origin of the Designs of the Arabic Faces’, 20 August 1943), Box 3614, NMAH. A ‬‬ l-Munaḍḍada al-ʿarabiyya (New York, 1929).‬The author is indebted to Frank Romano for sharing this valuable source.‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬

46

See for example Linda Jacobs, Strangers in the West: The Syrian Colony of New York City,

50

Chapter 2

Figure 2.8 Salloum Mokarzel, circa 1940. Courtesy of The Khayrallah Center for Lebanese Diaspora Studies, North Carolina State University.

1932), the founder and proprietor of al-Hoda.47 A Maronite Lebanese immigrant, Naoum established the weekly Arabic language publication in Philadelphia, Pennsylvania, where the first issue appeared on 22 February 1898. In 1903 the newspaper moved to New York and became a daily. Al-Hoda, unlike other Arab-American newspapers from this era, established itself and turned into the longest-published, and thus most influential Arabic newspaper in the US.48 Although the claimed representatives in 40 countries appear to be hyperbole, its impact went beyond the Arab-American diaspora.49 Given Naoum’s explicit political stance as a propagator of Lebanese nationalism and independence from the Ottoman Empire, al-Hoda’s influence undoubtedly reached back to the homeland.50 The founding of al-Hoda at the peak of Lebanese migration is 1880–1900 (New York: Kalimah Press, 2015), 271–274. Mokarzel saw himself as Lebanese, and although the country did not then exist, ‘Lebanon’ is used here as shorthand for the origin of this migrant community. 47

Al-Hudā (the guidance).

48

Between 1892 and 1930 some forty-four Arabic language newspapers were published in New York alone. Jacobs, Strangers in the West, 262. Al-Hoda ceased publication in 1971, reflecting the demographic changes in the Arab-American community where few people still read Arabic. Linda Charlton, ‘End Comes To Al-Hoda, Arab Paper’, The New York Times, 21 September 1971, 39.

49

Jacobs, Strangers in the West, 272.

50

“Naoum rejoiced in the fruits of his journalistic enterprise, Al-Hoda, as it aroused Lebanese nationalism and loyalty among the immigrants. After all, this was his primary purpose in founding the newspaper.” Mary Mokarzel, Al-Hoda, 1898–1968: The Story of

Mechanical Composition of Arabic

51

not a coincidence: between 1890 and 1920, about a third of the peasant population of Lebanon emigrated, making the relation between diaspora and homeland current and active.51 For the first years of al-Hoda’s publication the printing equipment was modest, “consisting of a few sets of type imported from Egypt, a small old fashioned press and a commercial press”.52 But after its relocation to New York, the newspaper’s readership expanded, and the Mokarzels soon sought to increase its print capacity. According to Mary Mokarzel, Salloum’s daughter who later also became editor of al-Hoda, to this end the newspaper proprietors collaborated with Mergenthaler and developed the first Arabic Linotype: […] in 1910, Naoum’s brother, Salloum Mokarzel, conceived the idea of adapting the Linotype to Arabic, which he did for the Mergenthaler Linotype Company. Al-Hoda became the first newspaper to acquire such a machine. Consequently, he was then able to expand Al-Hoda to eight standard-size pages.53 Salloum Mokarzel recognised the significance of this feat, as is apparent from an article in The New York Times published at the occasion of al-Hoda’s fiftieth anniversary: Looking back over the last decades, Salloum A. Mokarzel, present editor and publisher of Al-Hoda […] said yesterday that the most significant milestone in the paper’s long history was the installation in its old offices at 81 West Street of the first Arabic Linotype. Mr. Mokarzel himself had adapted the Linotype to Arabic characters in the cellar of 55 Washington Street. Al-Hoda adopted the machine in 1912 and was the first Arabic paper to use it commercially. This invention made possible and immeasurably stimulated the growth of Arabic journalism in the Middle East.54 L­ ebanon and Its Emigrants Taken from the Newspaper Al-Hoda (New York: Al-Hoda, 1968), 8. 51

Jürgen Osterhammel, The Transformation of the World: A Global History of the Nineteenth Century (Princeton University Press, 2014), 117.

52

Mokarzel, Al-Hoda, 1898–1968, 1.

53

Ibid., 1–2.

54

‘Arabic Paper Here Now 50 Years Old; Editor of Al-Hoda Recalls First Linotype Installed in 1912 as Significant Milestone’, The New York Times, 9 August 1948.

52

Chapter 2

In conclusion, it appears that following Salloum’s initiative, Mergenthaler worked from 1908 in collaboration with al-Hoda to develop the first Arabic Linotype machine and completed a first installation at the newspaper’s premises in Manhattan in 1911 – as the announcement in The Linotype Bulletin noted, “to make the Linotype practical for the voluminous [sic] Arabic language required years of study and hard work.”55 Indeed, such collaboration between client and manufacturer would have been a prerequisite for a successful Arabic typeface as Mergenthaler neither had the necessary expertise, nor the incentive to develop type without a prospective customer. The geographical proximity of Mergenthaler and al-Hoda in New York facilitated their successful cooperation: the combined knowledge of Arabic script and language, with technological know-how in one place was a rare, probably unique situation at this point in time. Mary Mokarzel’s anecdote that Mergenthaler offered Salloum $1,500 for a trip to Lebanon as an expression of gratitude for the sale of “thousands of Linotype machines to Arabic presses throughout the Middle East” underlines that their collaboration had been fruitful and satisfactory for all involved.56 Based on the work done in New York, British Linotype was able to manufacture similar Arabic machines, and may have shipped them to Constantinople before the outbreak of the First World War. By 1915 Mergenthaler’s Specimen Book of Type Styles featured a sample of an Arabic 22 pt fount: the first Arabic typeface ever made for mechanical composition (Figure 2.9). ∵ In the 1910s, the adaption of the Linotype for Arabic composition undoubtedly appeared as an impressive achievement, demonstrating to contemporaries the rapid advancement of the machine age. Only some 20 years after the commercial break-through of Mergenthaler’s machine, the mechanical composition of a script perceived as exotic and complicated seems to have contributed to the view that the Linotype embodied the pinnacle of printing technology. For example, Lucien Legros and John Grant’s seminal book Typographical PrintingSurfaces demonstrates the belief that the then current technology could adapt

55

Mergenthaler Linotype Co., “Arabic on the Linotype”, 81.

56

Mokarzel, Al-Hoda, 1898–1968, 52.

Figure 2.9

Mergenthaler 22 pt Arabic, Specimen Book of Type Styles, (New York: Mergenthaler Linotype, 1915), p. 483, reduced to 70% linear.

Mechanical Composition of Arabic 53

54

Chapter 2

Figure 2.10

Detail of the opening page of the chapter about linecasters in Legros and Grant, Typographical Printing-Surfaces, 421. Arabic composed on a Linotype, reduced to 90% linear.

to any typographical challenge.57 Three lines of Arabic set on a Linotype open the chapter on linecasting machines (Figure 2.10);58 and in the ensuing discussion of the intricacies of Arabic Linotype composition a stance emerges that became a recurring view on Arabic typography, and an influential factor in the evolution of the script’s printed form: The form of character is perforce more upright than would be the case were kerning permissible, but the result attained testifies to the skill of those workers at the subject, who have overcome the really great and exceedingly complicated typographical difficulties involved in adapting the arabic character to the restrictions of the composing machine.59 57

Lucien Legros and John Grant, Typographical Printing-Surfaces: The Technology and Mechanism of Their Production (London: Longmans, Green and Co., 1916).

58

The accompanying caption reads “composed and cast in Cairo on a Linotype machine in 18 pt and 14 pt Arabic”. No further evidence about Linotype composition in Egypt at this time has been found.

59

Legros and Grant, Typographical Printing-Surfaces, 542.

Mechanical Composition of Arabic

55

Mechanisation and its resulting advantages were considered so important that compromises were sought with the script, rather than the already existing equipment. The reasoning by Legros and Grant suggests that the difficulties in adapting Arabic for mechanical composition justified the drawbacks that were recognised in the resulting typography. This view, although entailing problematic features for the history of Arabic typography, should not necessarily be read as an expression of carelessness or Western ignorance. Indeed, the urge for progress through technological advance was a leitmotif for Arab publishers, often belonging to the educated elite with progressive ideas, as illustrated in the case of the Mokarzels in New York. The latest technology also stood for the more general aspirations of self-determination and advances in society as can be gathered from the proud description of the printing equipment employed by the newspaper: Al-Hoda now had the best Arabic pressroom in the world. No Arabic newspaper had comparable equipment, for Naoum was extravagant in equipping his New York pressroom. He bought set after set of type from Mergenthaler, and machine after machine of the best quality. When Model 14 was newly brought out, he immediately bought two more machines; this model allowed the use of three different sets of type at one time.60 Pride was not only taken in modern equipment, but also in the emulation of Western ideals such as the freedom of speech and the independence of publishers. Unlike in the homeland, the Press that the diaspora encountered was a well established institution and its principles were recognised, valued and adopted. With their explicitly nationalistic stance, the editors of al-Hoda aspired to blaze a trail for a Lebanese state and used their position in the US to articulate and spread their ideology. To this end the Western Press was emulated where possible, from the editorial policy to the design of the newspaper. With hindsight, Mokarzel emphasised this influence on al-Hoda’s development: ‘One of the reasons why Al-Hoda’s influence has been so great through the years […] is that it brought the American tradition of objectivity to Arabic journalism. In the East no distinction used to be made between news and comment. Everything was, and often still is, editorialized. Al-

60

Mokarzel, Al-Hoda, 1898–1968, 6.

56

Chapter 2

Hoda was the first Arabic newspaper to treat news strictly as news, to adopt the American system of headlines and regularly to run pictures’.61 The beginning of mechanical composition of Arabic thus happened alongside the establishment of contemporary Arabic journalism. The Linotype, although just a tool for the advancement of publishing, attained a central role as the first, and most widely employed typesetting machine, shaping the appearance of printed Arabic for decades.. Mergenthaler Linotype’s First Arabic Linecaster Founts The model for Mergenthaler’s first Arabic Linotype fount was probably one of al-Hoda’s foundry types, but documentation about its adaptation is lacking. The methods employed would, however, have been similar to those described in the authoritative literature on type-making.62 The process differed substantially from the crafting of foundry type: with the advent of mechanised typesetting, also type-making became a rationalised process of mass production. Whereas every manually cut punch was unique, in the mechanical production of type all matrices of a given character had to be identical, irrespective of the number of copies required. But because a frequently used punch would gradually wear down, changing the face it carried, the character configuration had to be stored differently. For this problem pantographic punchcutting offered a solution as it allowed for exact copies of a master configuration, making it possible to replace worn punches by identical replicas. In the type-making process this master configuration became the socalled pattern, an element unknown in the trade before the Linotype machine came to dominate it. Mechanical type-making began with the choice of artwork from either of two sources, (1) existing type or (2) new letter drawings, often provided by external designers.63 Where type served as a model, an optical device was used 61

The New York Times, ‘Arabic Paper Here Now 50 Years Old; Editor of Al-Hoda Recalls First Linotype Installed in 1912 as Significant Milestone’.

62

For general discussions of type-making and design methods see Tracy, Letters of Credit, and Southall, Printer’s Type in the Twentieth Century.

63

William Addisson Dwiggins’ description of his collaboration with Mergenthaler’s manufacturing department is a rare and lucid account of methods employed in the making of type for mechanical composition. WAD to RR, A Letter About Designing Type (Cambridge, Massachusetts: Harvard College Library, Department of Printing and Graphic Arts, 1940).

Mechanical Composition of Arabic

57

to project its enlarged face on to a drawing board. The enlargement was then copied in a process that gave ample room for interpretation of shape. Legros and Grant describe the results as being “of necessity neither exact nor sharp in outline” and requiring “further correction by skilful manipulators to standardize the resulting drawing for gauge, thickness of stroke, form of character and position to be ultimately occupied on the body of the type to be produced”.64 Similarly, if the artwork was provided by external designers the drawings could not be directly used for manufacture, but had to be adapted to the technical requirements of the type-making processes. Here the drawing office held a crucial role. Its draughtspeople were in charge of adapting the source, whether type or artwork, and translate it to the large, high precision drawings that were required for pattern making.65 In the next step the character outline guided a pantograph that cut a reduction of the drawing into a brass plate, creating a relief character shape, the pattern (Figure 2.11 overleaf).66 Finally, this pattern was used to guide the punchcutting machine which cut a reduced copy of the character in steel (Figure 2.12 overleaf). The resulting punch was then used in matrix manufacture, a laborious process including up to 60 different operations.67 ∵

64

Legros and Grant, Typographical Printing-Surfaces, 214.

65

“The pattern drawings for Linotype designs were made to a nominal size of 768 pt (270 mm; 10.6 in). […] strokeweight measurements were on a scale whose minimum increment was 0.00025 in (0.00625 mm). Suppose we want to hold dimensional errors on the type itself to less than half of this amount, with a maximum of, say, 0.0001 in (0.00254 mm). The reduction from drawing to type is 128 times, so that an error of this size on the type corresponds to an error on the drawing of 0.00128 in (0.325 mm). However, each of the two stages of reduction (drawing to pattern; pattern to punch) introduces its own inaccuracies, making the maximum permissible error on the drawing less than this.” Southall, Printer’s Type in the Twentieth Century, 26–27.

66

The process described here was followed at Mergenthaler Linotype, the Monotype Corporation had slightly different procedures. For a comprehensive account of the technical aspects of Monotype’s system see Richard Southall, ‘Technical History of Monotype Composing Machines’, in History of the Monotype Corporation, ed. Andrew Boag and Christopher Burke (London & Woodstock: Printing Historical Society, Vanbrugh Press, 2014), 350–370.

67

On average a matrix was subject to 35–40 operations. Linotype & Machinery Ltd, “Many Stages, Chapter IV in The Matrix Story,” Linotype Matrix II, no. 6 (Spring 1950): 7.

58

Chapter 2

Figure 2.11

Stages of pattern making, from left to right: plain brass plates; after cutting of the outline; finished pattern, ready to guide the pantographic punchcutting machine. From Legros and Grant, Typographical Printing-Surfaces, 208.

Figure 2.12

Schematic drawing of the pattern’s function in guiding the punchcutting machine. From Beatrice Warde, “Cutting Types for the Machines: A Layman’s Account”, in The Dolphin, no. 2, (New York: The Limited Editions Club, 1935), 65.

Mechanical Composition of Arabic

59

The 22 pt Arabic was the first of a number of founts developed by Mergenthaler according to models and requirements specified by Mokarzel, and it demonstrates recurring design issues of Arabic linecaster founts.68 In the 1915 specimen (Figure 2.9 on page 53) an accompanying paragraph explains the setup of the first Arabic Linotypes: The complete Linotype font of the 22 point consists of 181 characters and the usual quads. In order to accommodate the font it is necessary to use a machine carrying at least two magazines. Eighty-nine of the most frequently used characters and en quad are distributed in one magazine, and the other 89 characters and en quad are notched so that they will run in another magazine. The 3 remaining characters run as sorts. The number of matrices in a font of 22 point is 2,526.69 This description summarises one of the primary difficulties of mechanical Arabic composition – the character set limitation of composing machines.70 Traditionally, the typographical representation of the Arabic script required large character sets: for example the popular and successful Egyptian 68

By 1943 Mergenthaler had manufactured thirteen Arabic founts, four of which were derived from models provided by Mokarzel. Mergenthaler Linotype Co., “Origin of the Designs of the Arabic Faces”. These founts ranked among the first non-Latin typefaces available for machine composition and set a precedent for the mechanical composition of other scripts. This new possibility soon attracted further interest and in 1914 a Persian immigrant newspaper instigated the manufacture of a Syriac fount. J. F. Coakley, The Typography of Syriac. A Historical Catalogue of Printing Types, 1537–1958 (New Castle and London: Oak Knoll Press and The British Library, 2006), 251.

69

Specimen Book of Type Styles (New York: Mergenthaler Linotype Company, 1915), 464. Although only the 22 pt size was shown, also a 26 pt fount was advertised as being “ready by the time this book is in your hands”.

70

Right-to-left composition was not mechanically challenging. An article in the Linotype Matrix described it thus: “Arabic and Hebrew, which are read from right to left, are easily dealt with by the use of the special device which reverses lines after casting.” Linotype & Machinery Ltd, “Linotype Organizations Are Making Valuable Contributions to the Typographic Resources of the East,” Linotype Matrix II, no. 14 (1953): 1. The exact date of the device’s invention could not be established. By 1913 a patent was claimed for a mechanism “enabling a composing machine arranged to compose in languages which read from left to right, to compose in languages which read from right to left”. Previously Linotypes could only compose in either of the directions. Walker, Arthur Henry. Improvements in Typographical Composing Machines. UK Patent 22,300, filed 3 October 1913, and issued 14 May 1914, 1.

60

Figure 2.13

Chapter 2

Comparison of Arabic fount extent. (a) Case arrangement of an Arabic fount containing 470 characters, as used by the Egyptian Government Press in the early twentieth century. MT, correspondence folder Arabic (Egyptian 2). (b) Notional case arrangement of the first Linotype Arabic fount; characters without a direct equivalent to those shown in (a) are placed in the box at right. The reduction from 470 to 181 characters is achieved mainly through the removal of ligatures.

Mechanical Composition of Arabic

61

­ overnment Press type from the same period contained around 470 sorts (FigG ure 2.13).71 But compared to such foundry type where the size of the character set was discretionary and in practice only limited by expense, the Linotype only held 90 characters per magazine. This meant that for Arabic a machine with a secondary magazine was required to compose text in one size and one type style only. Basic typesetting therefore necessitated a significantly higher initial investment than was the case for Latin composition. Moreover, the composing speed was slower as the operator had to shift between two magazines and occasionally resort to the manual insertion of matrices.72 As explained by Ross in relation to the adaptation of Bengali to the Linotype, for complex non-Latin scripts the operation of the machine imposed its logic on the very concept of the fount: The keying method indubitably governed the design of the characters. Its size determined the number of sorts, and thereby the fount conspectus. Its manner of composition, in this case linear, affected the actual shape of the letter forms, as well as their spacing which was also governed by the channel sizes.73 In addition to the reduction of characters, the specimen setting of the 22 pt type demonstrates other deficiencies that Arabic linecaster founts were to retain for the entire period of hot-metal typesetting. Because whole lines were cast from rows of matrices, the characters of linecaster founts could not kern (Figure 2.14 overleaf). And while the lack of kerning was an aesthetic shortcoming in Latin type, for the Arabic script this limitation amounted to a severe design and legibility problem. Arabic letterforms frequently protrude into the space above or below adjacent letters, and therefore require kerning for an appropriate typographical representation. In that first Linotype fount the lack 71

A foundry type fount cut by the Ottoman printer Ohannis Mühendisoğlu is on record as having contained an approximate 1,500 sorts. Thomas Milo, “Arabic Script and Typography,” in Language, Culture, Type, ed. by John D. Berry (New York: ATypI Graphis, 2002), 122.

72

Matrices for manual insertion were called ‘sorts’ in a different use of the term to the conventional meaning of a character of foundry type. Such matrices had a tooth-combination that prevented them from falling into a magazine, making them travel the full length of the distributor bar until they dropped back into the so-called ‘sorts stacker’ for further use. Linotype Machine Principles (Brooklyn, New York: Mergenthaler Linotype Company, 1940).

73

Ross, The Printed Bengali Character and Its Evolution, 144.

62

Chapter 2

Figure 2.14

Schematic view of the casting edge of a composed line of Arabic linecaster matrices. Note how the linear arrangement of matrices prevents the casting of kerning characters. From Al-munaḍḍadä al-ʿarabiyyä, (Brooklyn, New York: Mergenthaler Linotype Company, 1929), 7.

Figure 2.15

Illustration of kerning problems. On the left the notional widths of individual matrices of Mergenthaler’s 22 pt fount are indicated. No character elements can protrude beyond the body of the matrix. Note that the word space (grey) is as wide as inner-word spaces. For comparison the desirable interlocking of character elements is demonstrated on the right with DecoType’s digital Emiri typeface.

‫ذ� �ف‬ �‫� كر�ه�ا ي‬

‫أ‬ � � ‫ر� س‬

Mechanical Composition of Arabic

63

Figure 2.16 Illustration of inadequate mark positioning, same configuration as above. In the Mergenthaler fount the word image is compromised because of the lack of kerning.

of kerning was exacerbated by character shapes that largely maintained conventional, poorly adapted letterforms and proportions. Deliberately designed characters could have minimised the problem, but here wide letterforms contributed to the erratic spacing (Figure 2.15).74 The fount exhibits another critical weakness inherent in the linecaster’s mechanics. In the Arabic script vowels and other discretionary marks should be positioned above or below the letter that they are meant to modify. Analogous to the lack of kerning, here the mechanism prevented the casting of such marks in their conventional position. Rather than taking the correct and unambiguous position above or below the modified character, matrices carrying ‫أ‬ marks had to precede or trail them as illustrated by the word ‫( ر� ��س‬Figure 2.16). The mark ‫ ء‬hamzä, which is supposed to be centred on its carrying ‫ ا‬alif follows the letter, and causes a large white space within the word. As the principal market for Linotype machines was the Press, and since it was customary for newspaper composition to rarely employ vowels, these could be side-lined in the development of the first Arabic hot-metal founts.75 Stylistically, Mergenthaler’s 22 pt Arabic fount had similarities with foundry type widely used in the late nineteenth and early twentieth century. Basing a fount on an existing typeface was not unusual, as the copying of an established model was the safest and easiest route to an acceptable rendering, and would have been sought by customer and manufacturer alike. Similar type was used by contemporary Arabic newspapers and the quality of some examples would have commended them as models for Mergenthaler (Figure 2.17 over�leaf). With features and proportions closely following the standardised Nasḫ manuscript hand, readability and elegance of such late Arabic foundry type make them exemplary for pre-industrial Arabic typography. Their principal characteristics were large character sets, conventional and well executed letterforms retaining proportions of great similarity to their manuscript forms, 74

The same shortcoming applied to Indian text composed on hot-metal Linotype casters.

75

The Linotype’s limitations for vocalisation became a more pressing problem once the machine found use in other publishing sectors. For a discussion see “Monotype Series 589, Full Vocalisation” on page 170.

64

Figure 2.17

Chapter 2

Detail of al-Mustaqbal, an Arabic newspaper published in Paris, France, 12 May 1916, 1, reduced to 90% linear. Courtesy of Internationales Zeitungsmuseum Aachen.

and numerous, often widely projecting kerns. Yet as has been discussed above, on a linecaster all of these characteristics were either unfavourable, as they increased cost and reduced speed, or simply impossible due to its mechanics. The appearance of Mergenthaler’s first Arabic fount suggests that it aspired to reproduce the foundry type model without substantial divergence from the original. Indeed, it seems improbable that at this point a deliberate design was sought; the adaptation of Arabic to the Linotype was undoubtedly the primary and dominating concern. As the 1915 specimen shows, excepting the compromises imposed by mechanical constraints, some aspects of the fount suggest that a deliberate effort was made to reproduce the model as faithfully as possible (Figure 2.9 on page 53). A number of letters are represented with more than the bare minimum of characters, indicating an attempt to maintain the standards of pre-mechanical Arabic type. For example, the letter �‫ ب‬bāʾ has five, instead of the minimum four characters, and ‫ ��س‬sīn and ‫ �ص‬ṣād feature

Mechanical Composition of Arabic

65

six distinct characters for use in different contexts. 35 ligatures of letter combinations were also included in the fount, an apparent attempt to lend a more familiar appearance to composed text. Overall, the 22 pt fount avoids too mechanical an impression through a slight slant and a generally rounded appearance. This is emphasised through variety in the height and vertical alignment of individual characters, that lend the type a certain liveliness. In addition to the noticeable effect of kerning restrictions, spacing of the fount is uneven. Where some character combinations create noticeable white gaps, other characters are almost colliding, as illustrated by the final ‫ د‬dāl followed by a ‫ و‬wāw. These spacing irregularities impede fluidity and open the potential for confusion where spaces within words appear as big as those between words. Overall the design of the letterforms appears unassuming, yet individual characters such as the initial and medial ‫ ك‬kāf look idiosyncratic and stand out on the page. Notwithstanding its shortcomings and the marked quality difference to foundry type, vivacity and proximity to conventional letterforms made the 22 pt fount a viable proposition. Combined with the significant gains in efficiency and speed that mechanical composition offered, the first Arabic Linotype fount must have been considered satisfactory for its intended use in al-Hoda. During the following years a range of related sizes and styles was developed by Mergenthaler. Of the thirteen founts mentioned in an internal memorandum of 1943, five were based on sources provided by Mokarzel, six followed other specimens submitted to Mergenthaler, and a further two were based on punches probably obtained from the British Linotype company.76 In the 1940 Mergenthaler Linotype faces specimen book, twelve of these earliest Arabic Linotype founts are shown (Figure 2.18 overleaf). Although presented in groups that suggest familiarity (Arabic, Arabic Bold, Arabic Old Style, Arabic Bold Old Style), the founts betray their heterogeneous origins.77 There is little consistency in terms of form or execution. The 14 pt fount was based on the 22 pt, and is among the more successful sizes. A specimen from 1921 shows a face with reduced liveliness, more homogenous vertical alignment and improvements in details from the earlier fount, such as the size reduction of the ‫ ك‬kāf (Figure 2.19 overleaf). Its character repertoire has been greatly reduced, limiting the number of characters for the rendering of most letters to their respective minimum. Moreover, the specimen presents only a fraction of the 76

Mergenthaler-Linotype Co., ‘Origin of the Designs of the Arabic Faces’.

77

The denomination of some of the typefaces as ‘Old Style’ is an idiosyncratic choice. Neither does it correspond to any established Arabic script classification, nor is the intended reference apparent from their design.

66

Figure 2.18

Chapter 2

Specimen of Mergenthaler’s Arabic founts. From Linotype faces, (New York: Mergenthaler Linotype Company, 1940), 879, reduced to 85% linear.

Mechanical Composition of Arabic

Figure 2.19

67

“14 point Arabic”, specimen, 19 January 1921, reduced to 85% linear. NMAH, box 3614.

ligatures available in the earlier 22 pt size, further reducing the number of matrices required for Arabic composition. The potentially most significant change is the introduction of half-forms. It meant that letterforms were divided into multiple characters which could be re-used in different contexts. In the 14 pt fount this renewed decrease of characters was only employed in a single instance, perhaps to test its viability. ‫ف‬ The letter �� fāʾ is commonly rendered with a minimum of four distinct characters for all positions within a word. Instead of this typical breakdown, the fount provides only three characters that may be used in more than one function, creating the four letterforms through their respective combinations. In the specimen’s presentation of individual characters, this method is shown at its best, rendering the initial and final letterforms joined together, constructed from the three separate characters (Figure 2.20 a overleaf). Yet, its use in the text setting betrays the limitations of the concept. Whereas the initial, medial and final shape of the letter are acceptable, the composition of the isolated shape from two characters results in a deformed, unwieldy letterform (Figure 2.20 c overleaf). The noticeable nick in the main horizontal stroke and the added width make the shape and proportion of the letter awkward and in-

68

Chapter 2

Figure 2.20

Details of Figure 2.19, illustrating the principle and use of half-forms for the con�‫ف‬ struction of the letter �� fāʾ. From left to right: (a) optimal setting (b) composition error (c) correct but unwieldy rendering.

consistent with similar letters such as the �‫ ب‬bāʾ. This direct influence of the fount synopsis on the design of the type is typical for the hot-metal era and representative of the often detrimental influence of mechanical constraints ‫ف‬ on aesthetics and legibility. The use of the half-form for the letter �� fāʾ must have been deemed acceptable by Mergenthaler, as the principle was adopted in later founts. Its deficiencies are most flagrant in the 24 pt Arabic Old Style, where misalignment between characters reaches unacceptable levels.78 Strikingly, the 14 pt fount under the same heading exhibits individual characters (see final ‫ ر‬rāʾ and mīm) that belong to the Arabic Bold fount of the same size. ‫م‬ Such laxity in an official specimen calls into question the company’s quality control mechanisms and may also suggest an underdeveloped appreciation of Arabic type. It follows that the continued addition of sizes and styles to Mergenthaler’s library did not entail a concurrent improvement of quality. The most successful founts were those first cut by the company, or those based directly on the original designs. The manufacturing issues observable in the founts under the heading Arabic Old Style are indicative of a lack of competence and care, and the diversity of sources and styles suggest little vision to develop a typographical programme that merits this description. Rather than a coherent, systematic selection of related faces and sizes, the Arabic library resembles a pot-pourri in which circumstance and convenience governed the addition of new founts. From the aforementioned range of sources, one may deduce that no strict selection criteria were applied, and that founts were developed as requested, and according to sources provided by clients. An Arabic typographical programme by Mergenthaler Linotype that featured considered planning and demonstrated initiatives informed by a firm grasp of the subject emerged later. Only in the middle of the twentieth century, when independence from colonial rule, advancing industrialisation and societal changes provided the context for increased indigenous publishing ventures, demand for typographical equipment increased and a veritable market 78

Despite the issues one can observe in these early Arabic linecaster founts, half-forms remained in use and were applied more successfully in later designs.

Mechanical Composition of Arabic

69

emerged. Against this background, and after a hiatus of some decades, Western typesetting machine manufacturers renewed their interest in the Arabic script and began the active development of Arabic type libraries – but now for clients in the Middle East. The Beginning of Arabic Type-Making at Linotype & Machinery From the 1930s the British Linotype & Machinery Ltd company commenced a separate Arabic programme of modest proportions that mainly provided equipment to Egypt, Palestine and Iraq. Its earliest typefaces strongly resemble the designs of Mergenthaler and have probably been based on models provided by the US company. An undated Linotype catalogue features six founts, three of which are duplexed with bold versions (Figure 2.21 overleaf).79 They feature similar characteristics and the same structural shortcomings as the earlier Mergenthaler Arabic faces, and do not merit detailed review. It is noteworthy that the catalogue concentrates on various kinds of printing equipment, another branch of Linotype business. The prominence of machinery indicates that type served as a means to the end of selling other equipment: type was an essential component of typesetting machines, but did not, in these early stages, constitute a product in itself. Whereas founts had to comply with ­elementary demands of legibility and aesthetic norms, their design was hardly pushed beyond a basic degree of functionality. It appears that these early Arabic line­ casting typefaces followed a largely utilitarian mind set in which decipherable and acceptable were often substitutes for readable and appealing. Overall, the extent and importance of the British Arabic type programme before the Second World War appear marginal, but began to supersede the US programme in the 1940s. In 1943 Mergenthaler’s works were still largely used for armament manufacture, but foreseeing an Allied victory it commenced to set out its post-war strategies.80 The then president of the company, Joseph T. Mackey tasked his deputy Chauncey Griffith to assess potential typographical developments across the globe after the cessation of hostilities. The resulting six-page report illustrates the company’s realisation of a changing world-order and the related prospect of new markets for typographical equip79

The equipment advertised in this catalogue suggests a publication date in the 1940s. Linotype & Machinery Ltd, Mākīnāt al-laynūtayb mākīnāt aṭ-ṭibāʿä mākīnāt ūtūblayt liṣabbi aṣ-ṣafaḥāt (Altrincham: Linotype & Machinery Ltd, n.d.).

80

For an account of the war effort of Linotype & Machinery Ltd see It Can Now Be Revealed (Altrincham: Linotype & Machinery Ltd, n.d. [1946]).

70

Chapter 2

Figure 2.21

Type specimen from Mākīnāt al-Laynūtayb mākīnāt aṭ-ṭibāʿä mākīnāt Ūtūblayt liṣabbi aṣ-ṣafaḥāt, (Altrincham: Linotype & Machinery Ltd, n.d.), 7, reduced to 50% linear. DTGC.

ment. Articulated in the language of Western hegemony, Griffith’s introduction provides valuable insights into the world view of an Atlantic enterprise with a global market in the middle of the twentieth century: This phase of our business must necessarily be considered from a broad international viewpoint in all its aspects and detail. Liberal as has been our policy in this respect, the future will require a more lively conception of the potentialities of cultural development amongst people hitherto regarded as submerged illiterates. There is bound to be an awakened appreciation of the advantages of the printed word among the vast populations of Africa, the backward nationalities of the Far East and in the Southwest Pacific as a result of their contacts with our English-speaking armed forces. It is not possible for us to translate this anticipation of future events into terms of physical preparation at this time. It is important, however, to recognize the situation as an almost certain possibility,

Mechanical Composition of Arabic

71

develop a sympathetic appreciation for it, and lay such ground work to meet it as may be possible.81 Griffith realised the particular potential of the Arabic script world, noting that “the Post-war period will unquestionably bring about a tremendous revival of printing among Arabic speaking peoples”.82 Whilst he assessed the company’s facilities for the script as “fairly well developed”, he also appreciated the need for a further expansion of the typeface range to address the typographical needs of the region. Conversely, this expansion began in the United Kingdom, not the United States. In December 1945, only months after the cessation of hostilities in Europe, Linotype had begun the cutting of the 14 pt size of the duplexed Arabic 2 with 3 fount, as documented in a punchcutting report to Mergenthaler.83 Bearing in mind that it took almost two years for the Linotype works to regain their full capacity after the war, and that during this period manufacture was significantly impeded by material and labour shortages, the immediate development of Arabic type may indicate a shift in priorities.84 It also marks a change of responsibilities and initiative from Mergenthaler to Linotype; a situation also noticed by Griffith in his reply: It has been our understanding for some years that development work in connection with Arabic and other Oriental or non-roman faces was to be considered a joint affair to prevent unnecessary and expensive duplication of effort. In fact it is my recollection that all development work in connection with Oriental faces was to be handled by the American company. In view of the extremely limited market and enormous amount of expense involved in the development and cutting of Oriental types such as Arabic, it would seem reasonable, from the standpoint of economy

81

Chauncey Griffith to Joseph. T. Mackey, Report, “Typographic Program,” 22 June 1943, 1, Box N9911, NMAH.

82

Ibid., 4.

83

Linotype & Machinery Ltd, “Monthly Punch-Cutting Report,” 30 December 1945, Box P3640, NMAH.

84

The difficulties of the company in the immediate aftermath of the war are set out in the lead article of the Linotype Matrix in early 1949. Victor E. Walker, “Mr V. E. Walker, Chairman and Managing Director of Linotype & Machinery Ltd, Writes to Friends in the Trade and Outlines the Company’s Aims and Prospects,” Linotype Matrix II, no. 3 (Spring 1949): 1.

72

Chapter 2

and efficiency, to concentrate work of this kinds in the hands of one organization.85 Griffith further mentioned a new Arabic development in Brooklyn aimed at scientific publications, referring to the company’s “expert consultants on Oriental and other foreign languages who are recognized internationally as scholarly authorities in their respective fields”.86 He thus hinted at the clients of Mergenthaler’s non-Latin faces, with the gradual demise of Arabic newspapers in the United States, becoming primarily academic institutions.87 Mergenthaler had a continuous working relation with Princeton University Press since the 1930s when the “first series of Oriental texts ever to be actually printed in America” were published by the institution.88 In numerous cases Mergenthaler sought the advice of its scholars in regard to foreign language developments and continued this practice after the Second World War.89 This particularity of the US company may account for Griffith’s assessment of an “extremely limited market” for non-Latin type, a situation which no longer applied for Linotype. As explained by Victor Walker, Linotype’s chairman and managing director, the situation in the Middle East, one of the areas served by the British company, was distinctly more competitive. Outlining the reasoning for the new Arabic type development, he noted that “unless we were to lose a considerable amount of business to the Intertype, it was essential we modified our 14 Pt. Arabic – in other words thicken it up – in order that it could be duplexed with a Black Face.”90 Walker added that although “a very fine script”, the new typeface under development in the United States was useless for newspaper production, hinting at the focus of Linotype in the Middle East. Whereas Mergenthaler at this time considered it pertinent to produce a fount that provided Arabic 85

Chauncey Griffith to Victor E. Walker, Letter, “Punch Cutting Report for December, 1945,” 21 January 1946, 1, Box P3640, NMAH.

86

Ibid., 2.

87

For an account of the demise of the US-based Arabic language Press, see William Geerhold, “The Rear Guard Press,” Saudi Aramco World, February 1967, 2–7.

88

“Princeton to Print Textbooks in Arabic; University First in America to Be Equipped to Turn Out Such Work on Its Own Press.,” The New York Times, 19 October 1930. http://select.nytimes.com/gst/abstract.html?res=F70E1EFA3A5C11738DDDA00994D8415B808FF1 D3&scp=2&sq=linotype+arabic&st=p

89

Mergenthaler was assisted in its Indian script developments by Norman Brown, Professor of Sanskrit at the University of Pennsylvania. Ross, The Printed Bengali Character and Its Evolution, 140.

90

Victor E. Walker to Chauncey Griffith, Letter, 7 February 1946, 1, Box P3640, NMAH.

Mechanical Composition of Arabic

73

vocalisation, a characteristic only found in religious, didactic or scholarly texts, the demands of the region where biased towards cheap and fast production. Walker thus concluded that “in the opinion of our sales force” the Arabic typeface developed by Mergenthaler “will have a very limited sale” in the Middle East.91 Although some Arabic and other non-Latin types continued to be made in the United States after the war, the majority of such developments gradually moved to the United Kingdom. By the 1960s, Mergenthaler’s Arabic fount sales were negligible, as gathered from the reaction of its typographical director, Michael Parker, to the request to provide samples for a joint specimen book: I would prefer not to advertise any of our Arabics [sic]. Disregarding small sorts orders, we have sold one font out of fourteen triangle numbers in five years – and that was five years ago. We do not have enough stock on hand in any face to make advertising attractive. All the designs are old-fashioned; I doubt that they appeal to anyone except the owners of Linotype in our few small American Arabic newspapers. We keep them on the list for these customers only.92 This shift was formally sealed soon afterwards when in 1965 the two companies agreed to give the responsibility for all non-Latin developments to Linotype.93 The Arabic Typewriter The numerous attempts and inventions dedicated to developing a writing machine have been well documented.94 Christopher Latham Sholes is generally credited as the inventor of the earliest commercially successful writing machine, first patented in 1868 and manufactured by the E. Remington & Sons company from 1873.95 The large number of competing writing machine 91

Ibid., 2.

92

Michael Parker to Walter Tracy, Letter, “Arabics,” 27 January 1965, WT correspondence, folder 18 Al Ahram, DTGC.

93

Walter Tracy to Michael Parker, Letter, “Non-Roman Manufacture,” 21 May 1965, WT correspondence, folder 17 Non-Roman General, DTGC.

94

For an account and a catalogue of early writing machines see Adler, Writing Machine, for a summary of the most relevant developments in typewriter history see Wilfred A. Beeching, Century of the Typewriter (London: William Heinemann Ltd, 1974).

95

Sholes, Christopher Latham, Carlos Glidden, and Samuel Soulé. Type Writing Machine.

74

Chapter 2

d­ esigns from the late nineteenth century gradually converged in the early twentieth century towards the standardised device that came to be the archetypal typewriter.96 It featured a keyboard of 45 keys, giving access to 90 characters through the shift mechanism. The keys were arranged in four rows, developed in a plethora of arrangements for different languages and character sets. For a machine capable of writing the Arabic script with its large number of letterforms, a scheme of substantial shape reduction had to be conceived. The history of the Arabic typewriter has yet to be exhaustively documented. Literature on the subject is sparse and largely inconclusive, and the numerous competing claims to its invention have to be scrutinised.97 In the present context suffice it to say that the inventive and entrepreneurial spirit of the late nineteenth century – in particular mechanisation and the nascent consumer market – probably gave rise to parallel developments of such a machine by numerous people and companies, without knowledge of each other’s work.98 Here only a brief sketch, casting some tentative spotlights about the particularities of the Arabic typewriter, is attempted given that the relevance of the typewriter extended well into the domain of Arabic typesetting. Two early patents illustrate some of the diverse techniques that were conceived to accommodate the technical constraints of a typewriter for the Arabic script. Selim S. Haddad from Cairo filed a patent for the simplification of US Patent 79,265, issued 23 June 1868. 96

For an historical account see for example George Tilghman Richards, The History and Development of Typewriters, 2nd rev. ed. (London: Her Majesty’s Stationery Office, 1964).

97

A recent article by Uri M. Kupferschmidt provides some pointers, yet remains limited by the apparent lack of primary sources. See “On the diffusion of ‘small’ Western technologies and consumer goods in the Middle East during the era of the first modern globalization,’ in Kozma, Schayegh, Wishnitzer (eds.), A Global Middle East: Mobility, Materiality and Culture in the Modern Age, 1880–1950 (New York: I. B Tauris, 2015): 229–262. Amongst others, Salloum Mokarzel is credited for adapting the Royal and Remington typewriters to the Arabic script. By contrast, the eminent literary theorist Edward Said claims that his own parents ‘invented’ the Arabic typewriter in Cairo in collaboration with the American Royal Company. Both accounts mention this invention in passing and lack corroborating evidence, but Mokarzel is a more plausible originator as his ties with Mergenthaler are documented and the geographic vicinity of al-Hoda to the Royal and Remington companies in New York make such collaboration a realistic proposition. A definitive history of this device would merit further research. See Mokarzel, Al-Hoda, 1898–1968; and Edward W. Said, Out of Place : a Memoir (New York: Vintage Books, 2000).

98

The history of numerous claimants to the invention of the first Latin-script typewriter is critically appraised by Michael Adler, Writing Machine: History of the Typewriter (London: George Allen & Unwin Ltd, 1973).

Mechanical Composition of Arabic

75

Figure 2.22 Patent drawings, from Selim S. Haddad, Types for Type-Writers or PrintingPresses, US patent 637,109, reduced to 40% linear.

­ rabic letterforms at the US Patent Office in 1899 (Figure 2.22).99 In his specifiA cation no specific claims to any implementation or particular machine can be found; rather, it articulated a simplification principle which was based on the use of characters for more than one position within words. Thus, letterforms were designed in such a way that one character could be employed at the beginning, and in the middle of a word, and one character could be used in final and isolated positions. The characterset could thereby be drastically reduced, and made to fit on a typewriter. The Hammond Type-Writer Company of New York patented a different approach at around the same time.100 The company was advertising its prowess in multi-lingual text production with the slogan “for all Nations and Tongues”, a selling point that was based on the device’s mechanics: the Hammond 99

Selim S. Haddad, Types for Type-Writers or Printing-Presses. US patent 637,109, filed August 12, 1899 and issued November 14, 1899.

100

George Croydon Marks, Chartered Patent Agent on behalf of the Hammond Type-Writer Company. Improvements in Type-writers. UK Patent GB190110460 filed 20 May 1901, and issued 22 June 1901. This British patent was filed by a chartered agent for the company, suggesting that an earlier patent had been filed in the United States.

76

Chapter 2

Figure 2.23 Keyboard layout of Hammond Arabic typewriter, patent drawing, reduced to 35% linear. From George Croydon Marks, Chartered Patent Agent on behalf of the Hammond Type-Writer Company. Improvements in Type-writers. UK Patent GB190110460.

t­ ypewriter allowed for a rapid change of type, facilitating the composition of languages with varying character sets.101 Hammond’s patent for an Arabic machine was also based on the need to reduce the number of characters, but its approach was perhaps a little more sympathetic to the script’s integrity. The inventors sought to devise a technique which maintained the conventional spacing mechanism in which all characters share the same width, but realised how poorly suited this was to Arabic. Thus, and in order to avoid undue stretching or compression of letterforms, the patent described a system in which narrow half-form characters were combined with suffix characters for use in final position.102 The patent’s primary claim was to a keyboard, and a corresponding division of the script; exact mechanics were not covered in the specification, and its influence on other Arabic typewriters is not known (Figure 2.23). Given the pronounced mechanical differences to what became standard machines, and its limited commercial success, the Arabic Hammond typewriter probably remained a niche product.103 More representative for the majority of Arabic typewriters that came to be used in practice is the Olympia SM9.104 It was a German-built (large) portable 101

Hammond Type-Writer Company, Advertisement, The American Journal of Semitic Languages and Literatures, 16, no. 1 (October 1899).

102

Improvements in Type-writers. UK Patent 10,460, at 2.

103

“While never an outstanding commercial success, the unique qualities of the Hammond machine endeared it to a small but devoted following”. Beeching, Century of the Typewriter, 111. A device called VariTyper was developed from the Hammond machine in the 1930s and featured among numerous other scripts also three Arabic founts. See De ­Baerdemaeker, “Tibetan Typeforms: An Historical and Visual Analysis of Tibetan Typefaces from Their Inception in 1738 up to 2009”.

104

The IBM Selectric was an exception in terms of its mechanics and was widely acclaimed

Mechanical Composition of Arabic

Figure 2.24

77

Arabic Olympia SM9 typewriter. Photograph by the author.

model of the last generation of mechanical typewriters, manufactured in the late 1960s and 70s (Figure 2.24).105 The typeface it employed remained virtually identical to one found in models from the 1940s and 50s, indicating its well-established position in the market. In terms of design there appears to be little variation between manufacturers, but in all Arabic typewriters similar principles of character reduction were employed. Generally speaking, Arabic typewriters had only two characters to represent letters which commonly have for its typographical advantages over conventional typewriters. But although the famous ‘golf ball’ writing head could be rapidly exchanged, allowing the user to choose different typefaces, it could only hold 88 characters. Given that Arabic typewriters employed proportional founts long before the Selectric was introduced, and no larger character set was offered, the device remained as constraining for the Arabic script as the earlier machines. 105

The serial number of this machine (3837854) suggests being manufactured in 1969 or 1970, see Beeching, Century of the Typewriter, 155.

78

Chapter 2

Figure 2.25 Examples of character design principles as employed in typewriter founts. Two characters are conceived in such a way that they can be used in four positions. Sample created by the author on the Olympia SM9, enlarged to 200% linear.

Figure 2.26

Typewriter composition sample demonstrating the character design principles in actual words, sample created by the author on the Olympia SM9, enlarged to 220% linear.

a minimum of four forms, and a single character for letters which usually have two forms (Figure 2.25). The reduced number of characters for this rudimentary rendering of Arabic could thus be accommodated on the keyboard of a typewriter. The design’s ingenuity lay in the conception of individual characters which allowed their use in multiple positions. Initial and medial characters were thus positioned to the very right of the body, giving them the required joining appearance (Figure 2.26). As consecutive sequences of such characters would result in tightly spaced words, extended connecting strokes projected to the left, counterbalancing the overall impression. Similarly, characters which were intended to serve as isolated and final forms had sufficient space to the left to create the distinct demarcation within a word. Indeed, spacing of the type was paramount for the success of the system, and explains why the Arabic typewriter fount had to employ a proportional width design. In this respect it was more advanced than its Latin-script counterpart, for which monospaced

Mechanical Composition of Arabic

Figure 2.27

79

Diagrammatic keyboard layout of the Arabic Model 5 Remington Portable typewriter. DTGC, box Simplified Arabic.

founts remained the norm until the introduction of the IBM Executive machine in 1944.106 A feature the Arabic typewriter shared with Indian script typewriters were characters whose body had zero-width, thereby not ٓ advancing theٔ  carriage when struck.107 In the layout at hand the marks ◌ (maddä) and ◌ (hamzä) occupied one key and could be employed above other characters to obtain ‫آ‬ ‫ة‬ correct spellings.108 The first word in Figure 2.26 above, ��‫ ال� �ل‬al-āla (the instrument/machine), ّillustrates use of the former. In addition to these floating marks, also the ◌ (waṣla), but none of the short vowels of the Arabic abjad were fitted on to the machine. 106

The fount of the Executive typewriter employed a basic unit system that assigned from two to five units per character. Beeching’s assessment that “after only eight years IBM had successfully solved a problem which had baffled typewriter inventors and manufacturers for nearly eighty years” seems to be an overstatement in light of the earlier Arabic proportional type. Beeching, Century of the Typewriter, 124.

107

Fiona Ross, email message to the author, “Re: Arabic Typewriter,” 28 March 2012.

108

To use these marks one had to resort to the return and space key for correct positioning above letters, a constraint that certainly discouraged their frequent use.

80

Chapter 2

Figure 2.28

Diagrammatic keyboard layout of an Arabic Continental typewriter. DTGC, box Simplified Arabic.

Similar layouts were employed with slight variations by other typewriter manufacturers, exemplified by a US-made Model 5 Remington Portable (Figure 2.27 overleaf) and a German Continental machine (Figure 2.28). Keyboard layouts for other languages using the Arabic script were also developed within the strict character set limitations shared by all typewriters (Figure 2.29).109 Internal mechanics and dimensions of the keyboard strictly prescribed the number of keys, and consequently characters, available for the representation of the script, leaving little margin for manoeuvre or typographical finesse.110

109

According to Mohammed Aman, Arabic typewriters featured between 65 and 75 [alphabetic] characters. Mohammed M. Aman, “Use of Arabic in Computerized Information Interchange,” Journal of the American Society for Information Science 35, no. 4 (1984): 206.

110

This constraint was not exclusive to Arabic typewriters, but a feature of typewriters generally. Seybold points out that although it was conceivable to customise the character repertoire for individual machines this was impractical, as “the available characters are chosen to provide as much universality as possible and at the same time to offer as many different characters as possible, even to the point where the user may be required to strike a lowercase l (ell) in place of the number 1 (one)”. John W. Seybold, The World of Digital Typesetting (Media, Pennsylvania: Seybold Publications Inc., 1984), 19.

Mechanical Composition of Arabic

Figure 2.29

81

From top to bottom, diagrammatic Olympia keyboard layouts for Pashto, Persian and Urdu. Adapted from Wilfred A. Beeching, Century of the Typewriter.

The typewriter became one of the most successful consumer goods of the twentieth century. When Beeching published his Century of the Typewriter on the occasion of the centenary of the earliest Remington machine in 1974, 25 manufacturers produced a total of eight to nine million typewriters per year.111 The typewriter dominated the office and its related output, where correspondence, accounting and the drafting of typescripts were all done on a device whose principles had been pioneered 100 years earlier. Yet, its application generally steered clear of the field of typesetting, the composing room and publishing.112 Typesetting, as a specialised craft with demands that could not be met by the simplicity of the typewriter, existed in its own right, but only as far as the Latin script world was concerned. For printing and publishing in languages using the Arabic script, the invention of the typewriter had a profound influence beyond the confines of the office desk. 111

Beeching, Century of the Typewriter.

112

The VariTyper and the IBM Selectric Composer blurred this distinction and presaged desktop publishing, which revolutionised the printing trade in the 1980s and 1990s.

82

Chapter 2

Arabic Script Reform The notion of modernisation through technological progress, as demonstrated with the example of al-Hoda, was a recurring feature in the evolution of Arabic typography. Reflecting wider societal trends in the first half of the twentieth century, paired with increased demand for printed communication, the idea of planned, rational modernisation was extended to the Arabic script.113 A particular expression of this search for modernity is found in the history of the script reform investigations of the Academy of the Arabic Language in Cairo. Born from the perceived need of language modernisation, a reverberation of the Nahḍa, the Academy aimed to modernise and standardise Arabic in order to make it more suitable for technical and scientific usage.114 Various committees were formed and considered at length how translations and specialised vocabulary could be conceived and systematised; their success, however, was limited: They contributed little […] to the effective creation of technical vocabulary, and language reform got bogged down in what looks like a parody of Egyptian bureaucracy.115 While not all judgements of the Academy’s work on language reform have been as unambiguous, suffice it to say that its aspirations were rarely matched by the realities of a living language. In the present context, its attempts at script reform merit more detailed consideration.116 Script reform projects were not new, nor exclusive to the Arabic script world: the nineteenth century saw many experimental proposals for a reform 113

Hobsbawm argues that most such ‘modernisations’ were “attempts by the elites of nonbourgeois societies to imitate the model pioneered in the West, which was essentially seen as that of societies generating progress, the form of wealth, power and culture, by economic and techno-scientific ‘development’ in a capitalist or socialist variant”. The Age of Extremes (London: Michael Joseph, 1994), 200.

114

Founded by decree of King Fuʾād I of Egypt as the Royal Academy of the Arabic Language in 1932, its name changed to Fuʾād I Academy for the Language in 1938; from the revolution of 1952 it became known as Academy for the Arabic language. Anwar G. Chejne, Arabic Language: Its Role in History, (University of Minnesota Press, 1969) 104–105.

115

Marwa Elshakry, Reading Darwin in Arabic, 1860–1950, (Chicago & London: University of Chicago Press: 2014), 302.

116

Chejne takes a more sympathetic view on the Academy’s work. Arabic Language, 105– 106. For ­Chejne’s perspective on script reform see pp. 157–160.

Mechanical Composition of Arabic

83

of the Latin script. Some of these attempts were driven by the hope of improving legibility, whereas others aspired to establish “a greater correspondence between written and spoken language”, and yet others proposed simplifications of shape for the sake of rational, modernist ideals.117 For the Arabic world, however, none of these would have been as arresting as the script reform executed by the young Turkish Republic in 1928. There, the radical resolve of President Mustafa Kemal Atatürk prescribed a shift from the Arabic script to the Latin alphabet within months.118 The rationales for this shift were the increase of literacy and general modernisation, shaped after an idealised European model. In August 1928, The New York Times cited Atatürk thus: Within two years […] the Latin alphabet will be universally used in Turkey and we shall prove that our intelligence is equal to the civilized world after banishing the incomprehensible Arab characters, which are responsible for 80 per cent. of Turk illiteracy. It is no longer a time for speeches, it is a time for action, and we shall institute this important reform, which is not our last.119 The Turkish script reform succeeded, bringing about the sudden and unmediated break with hundreds of years of highly developed writing culture.120 ­Ottoman calligraphy had just reached its peak at the turn of the century, and

117

Herbert Spencer, The Visible Word, 2nd rev. ed. (London: Lund Humphries in association with the Royal College of Art, 1969), 57.

118

Christoph K. Neumann, “Book and Newspaper Printing in Turkish,” in Middle Eastern languages and the print revolution: a cross-cultural encounter: a catalogue and companion to the exhibition (Westhofen: WVA-Verlag Skulima, 2002), 241–42. According to Neumann, newspapers had to be published in the Latin script within one month from 1 November 1928, books within two months.

119

“Kemal Says Latin Alphabet Will Oust Arabic in Turkey,” The New York Times, 11 August 1928, http://select.nytimes.com/gst/abstract.html?res=F20615F83E58127A93C3A81783D8 5F4C8285F9&scp=4&sq=arabic+alphabet&st=p (accessed 16 May 2017).

120

Atatürk intended this break, as he sought to distance Turkey from its Ottoman past and distinguish it from fellow Muslim cultures using the Arabic script. According to Frank Tachau, for Atatürk the break was necessary, and “if the proposed reform meant the loss of a great historic and religious tradition, this was simply one of the costs of modernization.” Frank Tachau, “Language and Politics: Turkish Language Reform,” The Review of Politics 26, no. 2 (1 April 1964): 194.

84

Chapter 2

the script reform proposal caused the swift demise of this sophisticated art form.121 ∵ Against this background, in 1936 the topic of script reform was raised for the first time at the Academy of the Arabic Language in Cairo with a discussion of the transliteration of European names in Arabic.122 By January 1938 the debate broadened when the difficulty of the typographical composition of vowels was identified and led to a general discussion of a reform of the Arabic script.123 Subsequently, a commission was formed and tasked to study ways of simplification within a loosely defined brief: L’objet de la recherche sera de travailler, par tous les moyens recevables, à simplifier l’écriture des lettres arabes, en inventant, pour faciliter une lecture arabe correcte, mais sans que cette amélioration ni cette invention ne sortent de ses principes fondamentaux.124 Over the following six years discussions ensued in which various degrees of reform were proposed by different members of the commission. In 1944 two competing projects were presented, yet neither was endorsed by the Academy. To engage a wider audience, the proceedings of the discussions were 121

Uğur Derman, Siegel des Sultans. Osmanische Kalligrafie aus dem Sakıp Sabancı Museum, Sabancı Universität Istanbul (Berlin: Deutsche Guggenheim, 2001), 44. Renewed interest and activities in Islamic calligraphy have been supported by IRCICA, the Research Centre for Islamic History, Art and Culture, since its foundation in 1980. “About Us,” n.d., http://www.ircica.org/about-us/irc380.aspx (accessed 16 May 2017).

122

Roland Meynet, L’écriture Arabe En Question, Les Projets de l’Académie de Langue Arabe Du Caire de 1938 à 1968, Hommes et Sociétés du Proche-Orient 3 (Beirut: Dar El-Machreq, 1971), 37. The account given here is based on this source.

123

In his account of the reform proposals Hunziker likens the complexity of vocalised composition to “the setting of a mathematical formula”. Hans Jürg Hunziker, “Aspects of Arabic Script Reform,” Typographische Monatsblätter 4 (1985): 15–16.

124

The purpose of the research will be to work, by all means admissible, to simplify the writing of Arabic letters, by inventing, to facilitate the correct reading of Arabic, but without the improvement or the invention going beyond its basic principles. Approximate translation by the author. Meynet, L’écriture Arabe En Question, 41.

Mechanical Composition of Arabic

85

published, and in 1945 a competition endowed with 1,000 Egyptian pounds of prize-money was announced.125 Having called upon the general public for proposals, the Academy had received more than 200 ideas by 1947 and formed a new commission to study the suggestions. In 1951 the subject resurfaced with a proposal by another member of the Academy, Mahmoud Taymour, yet it was referred back to the commission for review and comparison with the results of the competition.126 Finally, in 1952, the Academy announced that none of the projects was endorsed, causing further postponement of a decision on script reform. Only in 1958 pressure from the Egyptian Ministry of Education spurred a decision by the Academy. After renewed commissions and sub-commissions assessed the various proposals, in January 1960 the final decisions by the Academy confirmed that no meaningful reform could be arrived at.127 For a few more years the topic of script reform arose occasionally, and further projects were submitted to the Academy until this peculiar project eventually ceased in 1968. Tangible achievements of this prolonged process were minimal: it was agreed that educational books should be composed with vocalisation and conventions for the notation of sounds not found in the Arabic language were defined. However, the latter was also limited to two letters, the hard �‫ پ‬pe and ‫ڤ‬ soft �� ve, which was merely a codification of established practise.128 The proposals for reform exhibited a wide range of principles, from a complete shift to the Latin alphabet as in Turkey, to the use of disconnected letters in a single form and the reduction of contextual variants. Irrespective of their merits and shortcomings, the underlying assumptions that had initiated the debate were ill-informed. Problems and difficulties of existing typesetting techniques – notably character set limitations, the complexity of vocalised composition and the resulting compromises in speed – were taken as permanent facts, and their resolution delegated to a reform of the script. The inflexibility of the then current typographical technology gained supremacy over the evolved form of the writing system; the Arabic script, rather than its ­reproduction in 125

Ibid., 53.

126

Mahmoud Taymour, “A New Arabic Script to Facilitate the Use of the Diacritical Points Essential for Correct Speech and Writing,” The Islamic Review (October 1951): 16–20.

127

The commission proposed a reduction of the Arabic case lay from 300 to 169, suggesting ignorance of typesetting practise. The apparent assumption that there was a single case lay with a precisely defined number of characters was unrealistic.

128

Meynet, L’écriture Arabe En Question, 61–62. The third letter definition, the ‫ چ‬, proposed � to denote a soft g as in the “Genève”, was contested and had to be returned for further consideration.

86

Chapter 2

type, was identified as the source of the problem. This analysis, while questionable in 1936, conclusively lost its raison d’être as methods of composing Arabic evolved that equal to those of the Latin script in terms of ease and speed of production. By the time the project was abandoned, Arabic typography had advanced well beyond solutions to the initially identified problems. With hindsight the efforts towards script reform appear merely as a contained historical event of little lasting impact. In addition to questionable principles that were rendered obsolete by technological progress, the project also failed due to an apparent lack of subject-specific expertise. The open call for reform proposals primarily attracted amateurs, resulting in an expectable lack of knowledge and in consequence, impractical and ill-informed ideas. Despite the failure of the search for a convincing reform of the script, some of the contributors managed to promote their reform ideas outside the original framework, and without the appraisal of the Academy. The two most widely cited projects were those of the American-Lebanese architect Nasri Khattar, and of the Moroccan Ahmed Lakhdar Ghazal.129 Nasri Khattar’s Endeavours in Arabic Type Design In early 1932 Nasri Khattar (1911–1998), a recent graduate in Business Administration from the American University of Beirut, recognised in an apparent chance discovery that he could still decipher an Arabic word he had incorrectly typed on a typewriter. Instead of the appropriate medial form of ‫ �ه‬heh, he had keyed the initial variant and could still read the standard greeting for‫ن‬ ‫ن‬ mula � ‫ ا �ه�لا � و���س�ه�لا‬ahlān wa sahlān. Based on this realisation, he began his attempts to simplify Arabic printing type by reducing its various letter shapes to a set of 30 characters.130 The system he developed emulated Latin printing type with its individual, non-joining letterforms, thus breaking radically with Arabic script morphology (Figure 2.30).131 By April of the same year – in 129

Aḥmad al-Aḫḍar Ġazāl.

130

Yara Khoury Nammour, “Fighting Illiteracy With Typography,” Works That Work, no. 6, https://worksthatwork.com/6/unified-arabic (accessed 16 May 2017).

131

Sakkal cites Khattar as expressly drawing inspiration from the Latin alphabet: “Arabic, like English, will now have a method of writing in which the letters of a word are attached to each other, and a different method for printing in which the letters of a word appear disconnected.” Mamoun Sakkal, “A Brief Survey of Proposals to Simplify Arabic Script,” Sakkal Design, 2008, http://sakkal.com/articles/simplified_arabic/survey.html (accessed 16 May 2017).

Mechanical Composition of Arabic

Figure 2.30

87

Illustration of the design principles of Unified Arabic. From top to bottom, conventional type, then three styles of Khattar’s system: Neo-Naskhi, Neo-Kufic and Modern. From Form, no. 9, (New York: Kurt H. Volk Inc, n.d.). DTGC, box Simplified Arabic.

a ­remarkably fast process – Khattar had completed his first design. It came to be known as Unified Arabic. Subsequently, Khattar attempted to interest typewriter companies for his proposal and a prototype was built by the Remington Rand company, but its development appears not to have been pursued further.132 Fifteen years later, by then Khattar had moved to New York and worked there as an architect, a collaboration with the IBM company and the American missionary Frank Laubach led to the development of six typewriters which were presented to King Fārūq of Egypt. Whilst their collaboration was intended as part of Laubach’s fight against illiteracy and was thus embedded in a highly successful campaign, also this collaboration did not proceed further.133 Seeking other avenues for the continuation of his simplification ideas, the same year Nasri Khattar also submitted his proposal to the commission of the Academy of the Arabic Language; but like all others it was rejected.134 When Linotype began its collaboration with Kamel Mrowa in the 1950s (see page 109 and following), Khattar’s system was also discussed as a potential can�didate for adaptation to mechanical composition.135 Linotype, as a company, 132

Khoury Nammour, “Fighting Illiteracy With Typography.”

133

Ibid.

134

He filed a patent for the design the same year. Khattar, Nasri Najib. Arabic Alphabet. US Patent D.157,983, filed 3 October 1947, and issued 4 April 1950.

135

In December 1956 Mergenthaler’s Director of Typographic Development Jackson Burke

88

Chapter 2

experienced in gauging the potential of typographical developments, was sceptical, but Khattar was able to convince the Ford Foundation that there was potential in his method. A letter from Carl Gauerke, supervisor of Mergenthaler’s Overseas Sales Department, briefed Jackson Burke about the agreement between inventor and sponsor: I have just received a letter from Mr. Arthur Clark, Davidson representative, Washington, D.C., advising that Mr. Marshall Roethe, Managing Director, Middle East Institute, told him that the Institute has just approved an expenditure of $3,000.00 for Mr. Khattar to open an office in Beirut, Lebanon. It is anticipated by Mr. Roethe that Mr. Khattar will be given about $75,000.00 per year for at least ten years to exploit his alphabet. Also, Mr. Khattar is currently drawing a salary of $800.00 per month.136 Despite this sponsorship, Linotype remained cautious about the system’s merits. In a letter of June 1957, Tracy outlined his assessment of Khattar’s Unified Arabic: Nasri Khattar’s proposals have been known about since before the war. My personal view is that it is a system which might possibly be accepted (with persuasion) amongst Arab University intelligentsia; but I very much doubt whether it would be at all acceptable amongst ordinary readers of Arabic. In practice, the Khattar characters produce a result comparable in impact to what we would experience if we were asked to accept a single roman alphabet made up partly of capitals, partly of lowercase and partly of letters which were neither one nor the other.137 This unfavourable assessment and the company’s fruitful collaboration with Mrowa determined that Khattar was not approached by Mergenthaler on this occasion.138 Supported by the Ford Foundation’s grant, Khattar continued wrote to Linotype’s typographical adviser, Walter Tracy: “Since receiving your letter of November 27, I have attempted to develop whether or not Mr. Nasri Chattar [sic] has been able to generate any local enthusiasm for his revised or simplified Arabic. As far as I can determine, there has been no further support of Mr. Chattar’s ideas.” Jackson Burke to Walter Tracy, Letter, “Simplified Arabic,” 19 December 1956, Box P3640, NMAH. 136

Carl G. Gauerke to Jackson Burke, Letter, “Re: Mr. Nassi Khattar’s Unified Arabic Alphabet,” 3 May 1957, Box P3640, NMAH.

137

Walter Tracy to Raleigh Christie, Letter, 18 June 1957, Box P3640, NMAH.

138

Pattern drawings of characters resembling Khattar’s design, dated June 1939, suggest

Mechanical Composition of Arabic

89

the work on his method and designed a number of variations on the theme. Around 1959 – in Walter Tracy’s assessment probably because of the financial support through the Ford Foundation – one of Khattar’s designs was manufactured and released by the Fonderie Typographique Orientale of Beirut (Figure 2.31).139 However, despite the substantial economic backing, the method and resulting typefaces found little use in practice. In 1967 al-Ahram experimented with headline lettering based on similar principles for a short period of time, but its limited use suggests that it was not considered successful (Figure 2.32 overleaf).140

Figure 2.31

Type specimen from Al-Ḥurūf Al-Mūḥḥadä (Beirut: Fonderie Typographique Orientale, 1960) 2, ­reduced to 30% linear. DTGC, box Simplified Arabic.

that there had been contact with the company before the war. No evidence of a finished fount, or any initiative by the inventor to have his design manufactured by Mergenthaler Linotype or Monotype has been found. 139

Walter Tracy to Hrant Gabeyan, Letter, “Unified Arabic,” 2 August 1966, WT correspondence, folder 18a, DTGC. According to Tracy, Intertype also published two of Khattar’s design as a single duplexed matrix fount for their machine.

140

Hrant Gabeyan to Walter Tracy, Letter, “Unified Arabic,” 24 January 1968, WT correspondence, folder 18a, DTGC.

90

Chapter 2

Figure 2.32 Hand-lettered headline alphabet inspired by Unified Arabic, drawn by the al-Ahram’s inhouse lettering artist Samir Aziz. Al-Ahram, 24 January 1968, reduced to 50% linear. DTGC, WT correspondence, folder 18a.

Indeed, it appears that Khattar’s ideas failed to attract much attention within the domain of Arabic typography and the printing trade. It may be symptomatic that funding for his activities originated exclusively from a US institutions with little involvement in or knowledge of the field, whereas the system was considered unusable by the general trade. One can imagine that the amateurish approach to Arabic composition was appealing to other non-experts due to the apparent novelty embodied by Khattar’s system, but that it failed to convince informed practitioners.141 Ahmed Lakhdar Ghazal and the Asv Codar Type Ahmed Lakhdar Ghazal’s (1917–2008) proposal for script reform had similar origins as Khattar’s. A patent for the system was filed in France in 1954, and it was submitted to the Academy of the Arabic Language in Cairo in 1958.142 Failing to convince the commission, Lakhdar Ghazal pursued his ideas independently in Morocco. In its later published form, the system was called ASV Codar (Arabe Standard Voyellé – Codage Arabe), and was characterised thus by the inventor: […] un ensemble de caractères normalisés, destinés à la composition et à la transmission des textes en langue arabe selon les procédés et les 141

Recent interest in the subject confirms this distinction. Reappraisals of Khattar and related script reform proposals appear to be mainly driven by the desire for formal experimentation, rather than concern for typography as a means of text production. Frequently, they are also marked by a lack of historical understanding and appreciation of the context script reform originated from. See for example Abou Rjeily Cultural Connectives (New York: Mark Batty Publisher, 2011).

142

Lakhdar, Ahmed. Perfectionnements aux caractères d’imprimerie d’écriture arabe. French Patent 1.145.687, filed 29 December, 1954, and issued 29 October 1957.

Mechanical Composition of Arabic

91

t­ echniques existants dans les domaines de l’imprimerie, de la dactylographie, de l’informatique et des télécommunications.143 The system was expressly conceived for diverse technological environments, namely foundry type, hot-metal composition, typewriting, and later photocomposition. It comprised two versions which catered to different technological limitations. The ‘pure’ system featured only one character per letter, figures, vowels and basic punctuation, totalling 84 characters (Figure 2.33 overleaf). By means of three suffix characters, used in combination with medial characters, final and isolated forms of the letters could be created.144 Vowels were treated as independent characters, rather than floating marks. They were inserted trailing the character that they modified, in effect emulating the linear arrangement of letters found in the Latin script.145 The ‘total’ system included an additional 23 characters: 16 of them were referred to as ‘lettres d’esthétique’ and ّ replaced the constructed final and isolated forms, two catered for the ◌ šaddä and five for Persian and Moroccan spellings. Since 1955 Linotype had been informed about Lakhdar Ghazal’s developments. The company’s North African agent Gérard Sintès conveyed that a certain Paul Chaignaud “wished to interest [Linotype] in the Simplified Arabic scheme invented by his nephew Professor Ahmed (not Mohammed) Lakhdar”.146 But similar to Khattar’s case, the company’s experience suggested that the proposed changes were too radical, and did not promise success – as Tracy explained in a letter to Burke: After a good deal of consideration and consultation with Herbert Ellis, and others, we concluded that the Lakhdar scheme (at least so far as

143

A set of standardised characters, intended for the composition and transmission of texts in Arabic according to existing processes and techniques in the fields of printing, typing, computing and telecommunications, approximate translation by the author. “ASV Codar, Arabe Standard Voyellé, Codage Arabe” (lnstitut d’Etudes et de Recherches pour l’Arabisation, Rabat, n.d.), Box Simplified Arabic, DTGC.

144

The same principle had been invented much earlier in the Hammond typewriter, see “The Arabic Typewriter” on page 73.

145

As vowels could be placed within, as well as after letter groups, they had to be provided with and without a joining base line. This inefficient solution was responsible for the surprisingly large number of vowel characters (22 out of 84).

146

Walter Tracy to Jackson Burke, Letter, “Simplified Arabic,” 11 November 1957, 1, Box P3640, NMAH.

92

Figure 2.33

Chapter 2

“Systeme ASV-Codar Naskhi-2”, undated specimen, reduced to 45% linear. DTGC, box Simplified Arabic.

Mechanical Composition of Arabic

93

printing is concerned) is inferior to the Mrowa system because the result, though intelligible, is too great a visual departure from traditional. […] the Arab world seems to be developing acceptance for the idea of changing the traditional form of script for printing purposes, but the two-form letter system […] is likely to be the best bet from our point of view. When that has had twenty years of acceptance, it may be possible for the one-form system to be introduced into printing.147 Linotype, despite its interest in new Arabic type, was not convinced by the system, yet another body lent its support to the inventor. In 1958 the system was adopted by the Moroccan Government as part of its first national campaign against illiteracy and “studied with interest by National Commissions for Unesco from eight Arab States”.148 At this time Lakhdar Ghazal was the general secretary of Morocco’s National Commission for Unesco and the head of the Fundamental Education Section of the Ministry of Education – undoubtedly a position that favoured the promotion of his ideas. It was suggested that the application of the Lakhdar Ghazal method would promote the vocalisation of texts, and thus facilitate language learning. Purportedly, “when applied to the printing of newspapers and periodicals, the new system will provide an immense stimulus, by encouraging new readers to continue their studies, and persuading illiterate adults to join in literacy campaigns.”149 From Lakhdar Ghazal’s influential position a first practical implementation of his system emerged the following year. In late 1959, Monotype’s Continental Office received an order from the Ministry of Education in Morocco for the installation of one keyboard, one composition caster and one super caster each;150 bound to this order was the manufacture of three matrix founts designed by Lakhdar Ghazal. Monotype agreed to manufacture the founts, and assigned Series numbers 640, 641 and 642 to the various designs, first proofed in May 1960 (Figure 2.34 overleaf). With only one revision, the founts were approved and delivered one year later.151 However, despite the ease of manufacture, the 147

Ibid., 2. Today Tracy’s forecast appears overly optimistic, for some 60 years after the introduction of Simplified Arabic nothing suggests that further simplification is gaining more traction.

148

“Morocco Answers a Royal Call to the Classroom,” The UNESCO Courier, March 1958, 14.

149

Ibid. According to this article, around 1950 between 85 and 90% of Morocco’s adult population was illiterate.

150

Bogdanksi to The Secretary, Typographical Committee, Letter, “Lakhdar,” 13 November 1959, correspondence folder Lakhdar, MT.

151

Works to Stanley Morison, Letter, 5 May 1961, 1, correspondence folder Lakhdar, MT.

94

Figure 2.34

Chapter 2

“Trial No. 1” of Monotype Series, from top to bottom : (a) No. 640, (b) No. 641, and (c) No. 642, May 1960, reduced to 75% linear. MT, correspondence folder Lakhdar.

Mechanical Composition of Arabic

95

typeface turned into a failure for the Corporation. By August 1961, Monotype discovered that a clause in the contract granting full exclusivity to the customer had been overlooked, and that the Corporation was unable to promote or sell the founts without the explicit agreement of the Moroccan Ministry of Education.152 This was unfavourable for Monotype as the development cost for three series could not be offset through matrix sales. Writing in 1963 to L. Collier, the Overseas Manager E. A. Firmage summarised the project in unambiguous terms: I am enclosing as a matter of interest an Arabic face known as ‘Lakhdar’ Roca which Mr. Lequint had cut for a customer in Morocco. After this face was cut we learned that it was a copyright [sic], a situation which as you know we have always avoided in the past. In any case to my eye it is a monstrosity and certainly would not be taken up by any Arabic speaking countries of the Middle East.153 In 1960, the Moroccan government established the Institut d’Études et de Recherches pour l’Arabisation at the Mohammed V University, and Lakhdar Ghazal became its director.154 The institute’s express purpose was to make the Arabic language a modern business language, aptly suited to education, sciences and technology, “on a par with the languages of the most advanced civilisations”.155 This platform, together with the backing of the Moroccan government’s appraisal, provided the inventor with an ideal framework for 152

P. Lequint to The Secretary, Typographical Committee, Letter, “Re: Lakhdar,” 11 August 1961, correspondence folder Lakhdar, MT.

153

E. A. Firmage to L. A. Collier, Letter, 3 May 1963, 1, correspondence folder Urdu, MT. By 1969, Firmage’s assessment of the typeface’s economic potential was confirmed to be correct as only the original set of matrices had been sold, despite the customer’s initial claim that this design would be universally adopted in Morocco. G. C. Barrett to John Goulding, Letter, “Ecole Du Livre Rabat,” 30 June 1969, correspondence folder Lakhdar, MT.

154

Hunziker, “Aspects of Arabic Script Reform,” 18.

155

The Institute’s current website defines its goals thus: “Il a pour mission de faire de l’arabe une langue de travail moderne, une langue d’enseignement dotée d’outils pédagogiques appropriés et une langue des sciences et techniques, à l’instar des langues des civilisations les plus avancées.” Its mission is to make Arabic a modern working language, a language of teaching equipped with appropriate pedagogic tools, and a language of science and technology, on a par with the languages of the most advanced civilisations, approximate translation by the author. Université Mohammed V, “Institut d’Études et de Recherches pour l’Arabisation,” 2011, http://www.um5.ac.ma/um5r/fr/content/institut-desetudes-et-des-recherches-pour-l%E2%80%99arabisation (accessed 16 May 2017).

96

Chapter 2

the ­further promotion of his reform.156 Indeed, a few years later this support brought about the desired outcome, when other manufacturers were also convinced to produce type according to the Lakhdar Ghazal system. In the early 1970s an agreement with Photon International in France was made, and by 1975 the company had implemented the Lakhdar Naskh typeface on its photocomposition machines.157 In May 1975, Linotype’s agent renewed the attempts to interest the company in the system, underlining that development contracts had been agreed on with Photon and Harris Intertype.158 Expressing his personal opinion, the agent noted: For my part, I am persuaded by the influence and future that this product will have and I emphatically deplore the complete lack of interest on the part of Linotype […] if our competitors have given his project the attention it deserves, it is because they certainly realise the possibilities offered by the invention of Lakhdar-Ghazal.159 Here, Tracy was quick to correct Sintès’ suggestion, noting that it was primarily due to a lack of sales forecasts that the company had not taken a more active interest in the matter.160 Linotype’s overseas sales manager Richard Caesar replied to Sintès with the following observation: 156

Tracy learnt from Keith Panter-Brick, a senior lecturer at the London School of Economics in London who had met Lakhdar Ghazal in Morocco, that “though applications to import typesetting machinery have to be submitted to L-G by the Ministry, he can only give an opinion, not a veto. His advice can be set aside if other opinions carry more weight. This is a little different from what Peter Watkins learnt when he met L-G in the summer.” Walter Tracy to Arthur Henry Walker, Letter, “Lakhdar Ghazal,” 23 September 1974, WT correspondence, folder 18J-2 Trad Arabic VIP & 505 1974-1977, DTGC. Although without veto power, this position cannot be underestimated in view of Lakhdar Ghazal’s vested interest in the subject.

157

Ladislas Mandel, Photon’s Arabic Typography (Paris, 4 February 1975), 1, Fonds Mandel, folder Jawi News 55-65, MI. No detailed documentation of the agreement and manufacture has been discovered.

158

Gérard Sintès, “Letter from Graphic Systems to Melo. Characters Lakhdar - Ghazal,” 2 May 1975, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC.

159

Ibid.

160

“It is really absurd of Sintes [sic] to talk about our indifference to the scheme. Even though we have been told by our representatives in Beirut that the scheme is unlikely to have any effect in the eastern part of the Arab world, it has always been in our minds that it might have some value in the Maghreb. The only question has been to obtain from the agent some idea of what that value is. He has never given it, and I really think

Mechanical Composition of Arabic

97

[…] the fact that Photon and Harris Intertype have already exchanged contracts with the Arabisation Institute is not altogether surprising and indeed only to be expected in view of the fact that neither company can actually provide the full Traditional Arabic face.161 In reference to Linotype’s new designs for photocomposition and the advances it had made in software developments, Caesar argued: Linotype is a very long way ahead of Intertype and Photon in the setting of proper Arabic, thus the only course of action that can be taken by our competitors is to try and obtain permission to use certain ideas for very simplified Arabic in order to try and get themselves into the market. We believe that against our capabilities (computer programs) and full Traditional Arabic design, that [sic] Photon and Intertype do not stand a chance. […] I do not understand in any way why you are convinced by the success and the future of Lakhdar-Ghazal’s invention as opposed to our own enormous success with our own scheme (for which Al Ahram alone have placed an order worth over £400,000!!).162 Despite Caesar’s confidence in Linotype’s position, by September 1975 the influence Lakhdar Ghazal exerted on the trade in Morocco eventually provided the incentive to develop the typefaces for Linotype equipment. Here, an order of a V-I-P photocomposition machine from the Imprimerie Ideale in Casablanca was subject to the provision of a typeface by Lakhdar Ghazal.163 As Watkins explained in a letter to Tracy, the insistence on the type was not a choice of the printer: The two largest printers in Morocco are this customer and the Edition Marocaines, the latter being a Moroccan owned company who have that until he does we can hardly undertake any form of development.” Walter Tracy to Richard J. Caesar, Memorandum, “Lakhdar-Ghazal,” Memorandum, 13 May 1975, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC. 161

Richard J. Caesar to Gérard Sintès, Letter, “Lakhdar-Ghazal Typeface,” 23 May 1975, WT correspondence, folder 18J-2 Trad Arabic VIP & 505 1974-1977, DTGC.

162

Ibid.

163

For a discussion of photocomposition machinery see “Photocomposition: Towards Im� material Type” on page 205.

98

Chapter 2

r­ ecently ordered a Harris 600 with Ghazal Arabic. Ideale, who are French owned, fear that due to the other press having Moroccan connections the situation may arise where they will not be allowed to compose Traditional Arabic for the school-book market, which is their main reason for ordering a V-I-P. Consequently our customer does not feel he can take the risk of ordering a V-I-P unless we can give him a firm commitment on Lakhdar Ghazal Arabic.164 After consultation with agents and representatives, it emerged that throughout Morocco five V-I-P machines could be sold and Linotype finally agreed to implement Lakhdar Ghazal’s typeface called Naskhi 2. Subsequently a contract with the same conditions as those established with Harris Intertype and Photon was offered to Linotype. The five year agreement contained the following provisions: the institute granted a non-exclusive usage license to the manufacturer against a 15 % royalty payment per fount sale, whilst the fount’s development costs had to be covered by the company.165 The conditions, it appears, were accepted by Linotype and the development began.166 As the design was going to be implemented on the V-I-P, which provided more advanced capabilities than hot-metal linecasters, Tracy proposed to use these new functions to their best effect. In a letter to the agent Tracy suggested a few changes, including the use of a single weight of vowels for light and bold weights, the elimination of appendix forms (characters for the final position could be included in the fount), and a modification of the vocalisation principles: […] the vowel signs in the L-G system were evidently designed for matrices in the Linotype method of composition: that is, they are designed to appear after the letters. This has the effect of making the text occupy more

164

Peter Watkins to Walter Tracy, Letter, “Imprimerie Ideale, Casablanca. Lakhdar Ghazal,” 5 September 1975, WT correspondence, folder 18H1 Lakhdar Ghazal, DTGC.

165

Gérard Sintès to Linotype-Paul Ltd, Letter, “Caracteres arabes Lakhdar Ghazal - Photocomposition,” 13 September 1975, WT correspondence, folder 18H1 Lakhdar Ghazal, DTGC.

166

Two months later it transpired that the Lakhdar Ghazal typefaces were to become “obligatory” for printers in Morocco. Gérard Sintès to Walter Tracy, Letter, “Imprimerie Ideale - Casablanca,” 6 November 1975, WT correspondence, folder 18H1 Lakhdar Ghazal, DTGC. Testimony to the apparent vested interests the inventor of the system held; for Linotype it confirmed the decision to manufacture the founts.

Mechanical Composition of Arabic

99

space than is desirable. In photo-composition we can easily ‘program’ the signs to place themselves above the letters in the normal traditional way.167 Lakhdar Ghazal, however, only agreed to the use of a single weight of vowels, rejecting Tracy’s proposals to update the system to the newest means available.168 Insisting on the uniformity of the method across different technologies, the V-I-P version thus had to simulate the defects of earlier machinery. Consequently, the artwork for the typeface was provided by the Institute and Linotype’s Type Drawing Office made the necessary adjustments to match the 18-unit system of the V-I-P. By April 1976, it emerged that the Swiss Haas foundry had also been commissioned to develop foundry type according to the Lakhdar Ghazal system.169 In parallel Linotype & Machinery, the remaining part of the company responsible for hot-metal equipment, was asked to produce linecaster matrices of the Naskhi 2 typeface.170 Once the V-I-P founts and program had been finished, Tracy summarised the situation of the Lakhdar Ghazal project in a letter to Linotype’s representative Hrant Gabeyan in Egypt: The author of this scheme of simplified characters has evidently acquired financial support for it. A great deal is being done to ensure that the scheme is available to users of all kinds of typesetting. The Haas foundry has made foundry type, L & M have agreed to make trial matrices, it is on the Monotype and possibly Intertype. We have put the faces on V-IP and written a justification program. It is possible that Compugraphic have made the faces. Professor Lakhdar-Ghazal is actively campaigning for the adoption of the scheme by Unesco. Whether we admire it or not we should be ready to meet a demand for it.171 167

Walter Tracy to Graphic Systems SA, Letter, “Characters Lakhdar - Ghazal in V-I-P,” 2 December 1975, WT correspondence, folder 18H1 Lakhdar Ghazal, DTGC.

168

Gérard Sintès to Walter Tracy, Letter, “Caractères Lakhdar-Ghazal sur VIP,” 17 December 1975, WT correspondence, folder 18H1 Lakhdar Ghazal, DTGC.

169

Alfred Hoffmann to Walter Tracy, Letter, “Arabe Standard Voyellé A.S.V. Naskhi 2,” 6 April 1976, WT correspondence, folder 18H1 Lakhdar Ghazal, DTGC.

170

By this time the small branch of the company which retained the name Linotype & Machinery, continued to provide hot-metal equipment, whereas the main British business was Linotype-Paul Ltd.

171

Walter Tracy to Hrant Gabeyan, Letter, “Lakhdar-Ghazal Arabic,” 17 August 1976, WT correspondence, folder 18H1 Lakhdar Ghazal, DTGC. In addition to the environments

100

Chapter 2

Although Tracy had reservations about the merits of the Lakhdar Ghazal system, a pragmatic approach to the trade is evident in his assessment. Also, his assumption of financial support was sound – from 1975 an agreement of assistance to the Institute, signed by Morocco’s Minister for Higher Education and Unesco, ensured funding for the advance of Lakhdar Ghazal’s ideas. The UN organisation contributed $256,628, with national counterpart funds estimated at $423,000, to the project (MOR/73/024) “whose immediate objective was to enable the Institute to equip itself with the basic technical and administrative structures necessary for the performance of its mission in the field of communication techniques and linguistic development.”172 The project consisted of two components: the “definition of the ASV-Codar system” and a linguistic study entitled Lexar based on a “card file of over 600,000 documented semantic relations”, which may have been a dictionary corpus assembled by Lakhdar Ghazal.173 This first Unesco project lasted until 1977 and upon its closing a further programme of international assistance was launched. Its goals were the continuation of “graphic reform […] and the extension of its application to all fields affecting communication techniques with the object of achieving national and international officialization [sic] of the ASV-Codar system in its entirety;”174 In relation to Lexar, it was planned to create a “European/ Arabic lexicographical data base […] and the installation of a computer system at IERA comprising the hardware and software necessary for ­processing and

mentioned by Tracy, implementation on the Photon photocomposition machine and as Letraset dry-transfer lettering can be listed. The Letraset Type Lettering System was a product of the British Letraset Company, founded in 1959. It consisted of printed, laminated sheets of typographical symbols. Through the application of pressure, a strip-off layer of gummed tissue holding the symbol could be transferred to paper, providing an inexpensive and simple means to achieve high-quality letterforms. 172

Assistance to the Institute for Studies and Research on Arabization (IERA), Terminal Report (Paris, France, 25 November 1985), 1–7, UNESDOC, Paris, France, http://unesdoc. unesco.org/ulis/cgi-bin/ulis.pl?catno=67225&set=51692385_2_339&gp=1&lin=1&ll=1 (accessed 16 May 2017). From the report it emerges that the Moroccan national authorities requested Unesco assistance in 1973. At this time the Institute could only refer to the original patent specification and a “lexicographical card file containing an estimated 600,000 European-Arabic semantic relations” to secure funding.

173

Ibid., 8. The Unesco report appears oblivious to the existence of the system for roughly 20 years, claiming its “definition” as a new achievement.

174

Ibid., 10.

Mechanical Composition of Arabic

101

managing this data base.”175 To these ends, the considerable sum of $1,484,668 was provided by Unesco.176 Among the goals of the project (MOR/77/001), spanning from January 1978 to December 1982, was the training of Moroccan type designers. For this purpose the Swiss designer Hans-Jürg Hunziker participated in the project from January 1981 to June 1983.177 Lacking prior knowledge in Arabic type design, Hunziker familiarised himself with Lakhdar Ghazal’s system and general notions of Arabic writing before embarking on his new post in Rabat.178 As a consulting expert, Hunziker established a training course and began to instruct five aspiring designers. Building on substantial practical experience from previous work with Mergenthaler and Adrian Frutiger, Hunziker conceived a programme “in which theoretical and practical sides had to go hand in hand.”179 In his training course methods of letter-drawing developed for Latin script typography were complemented by Arabic calligraphy lessons, paying critical attention to cultural differences. Moreover, emphasis was laid on “the connection between material, instrument and form”, important to Hunziker “particularly in our search for the best possible letter-forms for new techniques, because these techniques have a tendency to dictate their own laws at the expense of letterforms.”180 By the end of Hunziker’s term in Morocco the students had jointly developed a new typeface. Its design had to be conceived according to Lakhdar Ghazal’s system in apparent contradiction of technological progress, a condition to which Hunziker objected but could not change.181 Within these constraints, the new typeface was loosely based on the Ruqʿä style (Figure 2.35 overleaf).182 According to Hunziker, the drawings were finished and made 175

Ibid.

176

According to the report the Moroccan Government’s counterpart funding was worth four million US dollars.

177

Hunziker, “Aspects of Arabic Script Reform,” 20. A further grant of $800,000, established in 1980 by the King of Morocco, provided additional financial support until 1985.

178

Hunziker recalls that because the position had already been open for some time and no candidates were found, he had to decide and depart to Morocco within only one month, leaving little time for preparation. Hans Jürg Hunziker, interview by the author, 14 April 2011, DTGC.

179

See also Hunziker, “Aspects of Arabic Script Reform,” 21.

180

Ibid.

181

Hunziker, interview.

182

For a discussion of the style see page 352 and following.

102

Chapter 2

Figure 2.35

Illustrations of the design process for the Ruqʿä typeface the students were developing under the supervision of Hunziker, reduced to 75% linear. From ­Hunziker, “Aspects of Arabic Script Reform”.

ready for dispatch to the Swiss photocomposition equipment manufacturer Bobst Graphic, which had agreed to produce and market the typeface.183 However, Hunziker’s Moroccan successor on the training programme unilaterally decided against the typeface’s manufacture. In a decision Hunziker called with hindsight “very sad” and a “scandal”, the outcome of two and a half years work was thus obliterated.184 Contrary to Tracy’s assumption that Linotype had to be prepared for a demand for the typeface, and despite the substantial financial and governmental support the system received, Lakhdar Ghazal’s design only attained a 183

Hunziker, interview.

184

Ibid. The Unesco report lists this as an accomplishment among other project activities. This and other typefaces developed with UN funds are thus described: “Done. Work partly completed but not finally accepted by IERA Directorate.” Assistance to the Institute for Studies and Research on Arabization (IERA), 53.

Mechanical Composition of Arabic

103

­ arginal position in practice. The only notable uses of the type were confined m to Morocco where, as demonstrated, typographical choice was manipulated by government decree, veiling the vested interests of Lakhdar Ghazal. During the ten-year period of Unesco support the following ASV-Codar implementations were imported to Morocco: 28,525 typewriters, 2,740 kg of foundry type made by Haas (the equivalent of 278 complete founts), 30 sets of Linotype matrix founts, and 32 photo-typesetting systems of different makes.185 Assuming that these systems were based on similar contracts as the ones signed with Linotype, Harris Intertype and Photon, the royalties these imports would have yielded for Lakhdar Ghazal without any development costs were substantial. In what is undoubtedly among the best financed typographical projects in history, the final Unesco report estimated that from 1976 to 1985 the ASV-Codar programme benefitted from 3.68 million US dollars of funding.186 Use of the typeface remained limited despite its availability across the widest range of technologies in use at the time, and must be seen as testimony to its limited merits. Based on obsolete assumptions, the Lakhdar Ghazal system reiterated rather than challenged technical constraints. The inventor’s refusal to modify his scheme suggests a stance that was at best ignorant, and at worst wilful in its lack of critical analysis and rejection of evolution. With hindsight such politically driven Arabic script reform proposals are historical curiosities. In view of the failures such proposals yielded over a fifty year period the promises advanced by Arabic script reform initiatives have remained unrealised.187 Indeed, through the evolution of Arabic type a more lasting change of the script has been initiated than by that of prescriptive script reform. Typographical simplification as a by-product of practical requirements, economic considerations, and the various technical constraints of changing technologies had a wider and longer lasting influence on the form of 185

Ibid., 31–32.

186

Ibid., 42. A further 4.42 million went into the Lexar programme.

187

One of the most important arguments in favour of script reform had been the alleged contribution to alphabetisation. No studies of the effects of the ASV-Codar system are known, but Morocco’s illiteracy rate did not improve faster than those in Arab countries with similar literacy levels. From 1985 to 1994 the following adult literacy rates were estimated for Algeria, Morocco and Yemen: 50%, 42%, 37%. From 2005 to 2010, they had respectively increased to 73%, 56% and 64 %. Thus, Algeria’s rate increased by 23 %, Yemen’s by 27%, and Morocco’s only by 14%. The faster increase of literacy in Algeria and Yemen was achieved despite years of civil war and turmoil in both countries. EFA Global Monitoring Report. Youth and Skills: Putting Education to Work (Paris, 2012), 320, http:// www.efareport.unesco.org/.

104

Chapter 2

printed Arabic. Among the most significant evolutionary steps was the invention of Simplified Arabic, a new typographic form of the script that originated in the context of the post-war Middle East. The Printing Trade and the Middle East after the Second World War The Middle East changed markedly between the development of the earliest Arabic Linotype and the 1950s. Whereas the initial impetus to adapt the Arabic script for mechanical composition came from within the Arab diaspora in the United States, 40 years later the centre of demand for Arabic equipment shifted to the Middle East due to increased publishing activity there. This, in turn, reflected changes in geopolitics. Two world wars had shaken the European order and thus the order it had imposed on large parts of the world. In the aftermath of the Second World War and with the emergence of the United States and the Soviet Union as superpowers, European colonial power was waning. The war had been costly for Britain and France especially, and throughout their territories they were compelled to give way to increasingly articulate national aspirations voiced by those they governed. The Middle East was fundamentally affected by this changing world order: here the retreating colonial powers left a fragmented region where often artificially defined entities struggled to become nation states on a European model. Nevertheless, between the withdrawal of French troops from Lebanon and Syria in 1946 and the end of France’s war in Algeria in 1962, all the former colonies of the Middle East had gained their independence.188 The European printing industry also felt the repercussions of the Second World War. The factories of the leading typesetting manufacturers in the United Kingdom, Linotype and Monotype, had been requisitioned by government for armament production (Figure 2.36).189 In 1941 Linotype lost a depot 188

The collapse of the European empires was global. In 1945 the British Empire still governed approximately 700 million people, but by 1965 this number had shrunk to five million. Andrew Porter, The Nineteenth Century, The Oxford History of the British Empire vol. III (Oxford University Press, 1998), 330.

189

The machine manufacturers’ workflows were well adapted to the precision required for weapons manufacture: “between 1939 and 1945 we [Linotype] manufactured over one and a half million complete units and parts for all kinds of guns, aeroplanes, tanks etc., in addition to over 3,000 jigs and fixtures for use in Government ordnance factories. That class of work came to us because in our own line we were accustomed to work to almost microscopic tolerances”. Victor Walker, “Mr V. E. Walker, Chairman and Managing

Mechanical Composition of Arabic

Figure 2.36

105

Front cover and spread of It can now be revealed, published by L&M as a record of the company’s contributions to the war effort. Altrincham: Linotype & ­Machinery Ltd, n.d. [1946].

in ­Endell Street to bombardment and Monotype’s London office in Fetter Lane was destroyed during a raid the same year.190 Further difficulties set in after 1945 with limited supplies of resources, both human and material, and government restrictions imposed on the manufacturers. Because of slow reconstruction efforts Linotype was only able to fully resume the production of their own products after approximately two years. The lead article of Linotype Matrix’s first issue after the war in spring 1948 portrayed the situation of the trade in sombre language: Most printers nowadays are not only short of material and skilled labour, but find themselves facing the gloomy prospect of being unable to undertake building and capital developments during the next few years. And so far as new plant goes, the prospect is no brighter; three out of every four of the machines produced by Linotype & Machinery are being directed abroad […] It is no part of the responsibility of Linotype & Machinery to do other than produce machines. The home market should come first; Director of Linotype & Machinery Ltd, Writes to Friends in the Trade and Outlines the Company’s Aims and Prospects.” See also Linotype & Machinery Ltd, It Can Now Be Revealed. 190

Sebastian Carter, “The Morison Years and Beyond: 1923–1965,” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 14–25.

106

Chapter 2

unfortunately, it must now come last. And not from any choice on our part, but from grim national necessity.191 This recovering industry adapting to a changed world, and the post-war geopolitical situation in the Middle East provided the background for subsequent developments of Arabic typography. The transformation of former colonies into nation states provided new opportunities and markets for the manufacturers of typesetting equipment. Nationalist sentiments increased the demand for publications in local scripts and languages. Independence and late industrialisation coincided to create a need for printing equipment able to support growing volumes of communication. Without indigenous industries to supply it, printers and publishers in the Middle East looked to the former colonial powers for the equipment and technology they required.192 The makers of typesetting equipment, primarily Intertype, Linotype and Monotype, sensed the urgency of the situation – and the economic opportunity – and directed an unprecedented level of resource towards this growing market. A growing awareness of a world of writing systems beyond those of European cultures manifested itself also in increased coverage in trade journals. In 1953, the lead article of Linotype Matrix boasted about the contributions of the company to the “typographic resources of the East”, and acknowledged the recent changes in demand: […] it was not until the eighteenth and early nineteenth centuries that any serious attempts were made to produce printed books for the native peoples of the East, and only in recent times that they have demanded the means to produce printed matter themselves.193

191

Linotype & Machinery Ltd, “Printers Must Face the Present, Here There, Everywhere – Except for the Printer at Home,” Linotype Matrix II, no. 1 (Spring 1948): 1. The same year Walter Tracy, who had worked part-time for Linotype from 1947, responsible for “seeing to press the journal called Linotype Matrix”, was hired as a full-time staff member. Tracy, “Composing Room Days (and After).”

192

In 1960 the traditional centres of industrialization in Western Europe and North America still accounted for over 70 % of gross world output and produced almost 80 % of global industrial output. Nigel Harris, The End of the Third World: Newly Industrializing Countries and the Decline of an Ideology (London: Penguin Books Ltd, 1990), 102–3.

193

Linotype & Machinery Ltd, “Linotype Organizations Are Making Valuable Contributions to the Typographic Resources of the East,” 5.

Mechanical Composition of Arabic

Figure 2.37

107

Photograph of the Linotypes in the machine room of the Government Press of ­Pakistan, August 1951, reduced to 50% linear. NMAH, box P3640.

The Monotype Corporation, too, felt the pressure to produce non-Latin founts and machinery. Post-war shortages of material and personnel were exacerbated by the “large amount of manufacturing capacity […] devoted to cutting non-Latin faces for the newly independent countries of the British Commonwealth, particularly for India and Pakistan, and for other parts of the Far East”.194 This wave of new non-Latin typographical development is exemplified by the inauguration of the new Government Press in Pakistan in 1951, three years after the country had gained independence. The plant was the largest in the country “with a machine room area of 27,500 sq. ft”, equipped with offset printing presses, type-casting equipment, and ten bilingual Linotype machines for English and Urdu composition – a substantial capital investment for the new nation (Figure 2.36).195 In 1953 Linotype Matrix reported on the “efforts which have been made in the past three years to provide Pakistan with the means of producing printed matter in its national language by typographic means”. The article’s account of the collaboration between Mergenthaler, the Government Press and the Dawn newspaper, for which “a ‘pilot’ fount of Urdu was produced”, sheds light on the state of typography in the country. It discussed the 194

“The Morison Years and Beyond: 1923–1965,” 23.

195

Linotype & Machinery Ltd, “Government Press of Pakistan, A Valuable Asset to a Developing Nation,” Linotype Matrix II, no. 11 (Autumn 1951): 6.

108

Chapter 2

ongoing practice of writing entire daily newspapers by hand, identifying it as an impediment to the development of the new nation. However, the attempt to develop a Nastaʿlīq fount for the Linotype was hindered by its mechanics, and the resulting typeface was visually poor and far removed from acceptable standards (Figure 2.38). Linotype Matrix acknowledged the inadequacy of the design, noting that: It was agreed by everyone concerned in this important undertaking that the fount as used had certain imperfections, but it was also understood that it represented merely the beginning of the immense task of performing for Pakistan what had been done for Europe in the fifteenth century – the changing over from the laborious work of the calligraphers to the mass-production typographic method (the only possible system for a modern nation of eighty million people).196

Figure 2.38

“12 point Urdu No. 2”, specimen, March 1952, enlarged to 200% linear. NMAH, box P3640.

Mergenthaler’s Nastaʿlīq was unsuccessful and hindered the shift to mechanical composition of Urdu; it was only 30 years later that new technologies enabled an acceptable reproduction of this writing style.197 Despite its failure, this example is representative of a recurring theme in the period marking the 196

Linotype & Machinery Ltd, “Linotype Organizations Are Making Valuable Contributions to the Typographic Resources of the East,” 5.

197

See “Digital Tools in the Service of the ‘Hanging’ Styles Nastaʿlīq and Ruqʿä” on page 340.

Mechanical Composition of Arabic

109

end of the European colonial empires. Here, the efforts of many emerging nations to modernise, often by adopting Western technology, were also accompanied by attempts to shape progress in appropriate and authentic ways. Industrialisation and imported expertise were necessary and pragmatic but needed to be balanced with expressions of national, cultural and linguistic identity, and the requirements of education. These dynamics were not always or easily compatible, and so their co-existence required compromise and often a re-ordering of priorities. A few years after Mergenthaler’s failed Nastaʿlīq project in Pakistan, a different context provided the background for a successful example of post-war Arabic type-making. Here the initiative came from the Lebanon, and the implementation was effected by the British Linotype & Machinery company. The resulting typeface – Simplified Arabic – exemplifies the tension between indigenous printing requirements and technological progress, and became a milestone in the evolution of Arabic type. Simplified Arabic, Kamel Mrowa, and Al-Hayat 198 The origin of Simplified Arabic can be traced to 1954 when Kamel Mrowa, a Shiah Muslim, asked Linotype to produce a new typeface for the Lebanese daily newspaper al-Hayat (Figure 2.39).199 Mrowa was the owner and ­editor-in-chief

Figure 2.39 Kamel Mrowa, probably in 1966. Courtesy of Dar Al-Hayat Information Center, Beirut, Lebanon.

198

A version of this chapter was published as “Simplified Arabic, a New Form of Arabic Type for Hot Metal Composition,” Typography Papers no. 9 (2013) 173–189.

199

Kāmil Mruwwah. Another current transliteration of his surname is Mroueh. The newspaper’s name is al-Ḥayāt (the life).

110

Chapter 2

Figure 2.40 Sample produced on the Arabic German Continental typewriter sent by Mrowa to Linotype, reduced to 50% linear. See also Figure 2.28 for the corresponding keyboard layout. DTGC, box Simplified Arabic.

of al-Hayat, which he had founded in 1946, the year of Lebanon’s independence. The newspaper typified the entrepreneurial spirit of Lebanon’s early years, as the country developed into one of the Middle East’s economic hubs, with Beirut an important Arab capital. Lebanon’s advantageous geographic position had fostered a long mercantile tradition, attracting foreign investment and turning the country into one of the most prosperous in the region. Its literacy rate, the highest in the Arab world, testified to the country’s wealth and its place as a centre of Middle East publishing. On-going development in the 1950s, encouraged by liberal economic policies, made Lebanon fertile ground for entrepreneurial activities.200 Against this background, Mrowa approached Linotype with a proposal for a new Arabic typeface. In a meeting with the company representative Herbert Ellis, Mrowa “described with enthusiasm the satisfactory and simple form of Arabic script available from a German typewriter he is acquainted with. Mr. Mrowa’s belief is that the principle embodied in it could be applied to Linotype composition” (Figure 2.40).201 This scheme of simplification was based on 200 David Gilmour, Lebanon: The Fractured Country, 2nd rev. ed. (London: Sphere Books, 1987). 201

Walter Tracy to Jackson Burke, Letter, “Arabic,” 17 September 1954, Box P3640, NMAH. Tracy is quoting information that Ellis had gathered and passed on to him.

Mechanical Composition of Arabic

111

a reduction of those letterforms required to produce the joined appearance of Arabic; broadly speaking, where typographical founts employed a minimum of four shapes per letter, the typewriter had only two.202 Typewriters based on this scheme had been in use for decades, and despite the simplified appearance of their founts, the text they produced had proved acceptable for certain kinds of documents such as business correspondence and accounts. For typeset documents however, any equivalent to the typewriter’s simplification of Arabic would have appeared too unconventional.203 Yet, Mrowa had precisely this in mind. The simplification he was proposing, which would reduce the number of characters to the 90 available on the typewriter, was intended to increase composition speeds while producing text that was still aesthetically acceptable to a typical newspaper readership. His aim was a pragmatic solution that could be implemented within the constraints of typesetting machines but without upsetting reading habits too dramatically. Simplified Arabic: Proof of Concept When Kamel Mrowa proposed the simplification concept to Linotype, its typographical adviser, Walter Tracy (14 February 1914–28 April 1995), had been with the company for seven years.204 Tracy’s role was to assess the commercial viability and potential success of new typeface projects. By the 1950s Linotype (rather than Mergenthaler) handled all machinery and equipment sales and servicing contracts in the Middle East, so Mrowa’s proposal came under Tracy’s scrutiny. Tracy’s interest and support for the simplifying of Arabic is evident in his correspondence with Jackson Burke, director of typographical development at Mergenthaler. In a letter of early September 1954 describing Mrowa’s proposal, Tracy reasoned: To wonder why something of this sort was not done long ago is to conclude that the importance of calligraphy as an Arab [sic] art, the ­complicated 202 See “The Arabic Typewriter” on page 73. 203 An approximate analogy in the context of Latin script might, for example, be the use of only capital letters to typeset literary texts. 204 Coincidentally, the period of early mechanical composition of Arabic corresponded with the first half of Walter Tracy’s life. He was born in 1914 into a “poor but respectable” London working-class family. In the 30 years after the Second World War, he was going to leave a lasting mark on Linotype’s typographical programme and the company’s involvement with Arabic typography. Walter Tracy, “Composing Room Days (and After),” 3.

112

Chapter 2

nature of a running script, and the necessary ‘pointing’, have made typefounders anxious to provide printers with everything necessary for acceptance by critical readers, regardless of expense. Mechanical composition is comparatively recent in this field; its ‘limitations’ are still in process of gaining acceptance (though they could probably be turned to definite advantage if prejudices can be removed).205 Tracy then outlined Linotype’s plan to develop a trial design “closely based on our 14 pt double-letter Arabic face” and photographically compose a specimen to assess the interest of potential customers.206 Mergenthaler’s reaction was muted: in an internal memorandum, Chauncey Griffith, Mergenthaler’s former director of typographical development, discouraged attempts at further simplification of Arabic (i.e. beyond that which had already been implemented in existing typefaces for the Linotype), claiming that it was “exceedingly difficult, if not impossible, to overcome the inherent prejudice of Arabic thought respecting any material form in the traditional Arabic script”.207 Griffith’s assessment was probably a reflection of Mergenthaler’s experience with its customers who used Arabic founts in the United States – mainly small, immigrant newspapers and academic publishers. Where Mrowa primarily needed faster composition speeds for a large and growing newspaper, the Arabic Press in the United States probably valued the authenticity of conventional Arabic typesetting. For the academic publishers, linguistic precision was a priority, something simplification could not achieve; and in any case, given their small print runs, they would have little to gain from increases in efficiency. Despite the doubts raised by Griffith, Tracy went ahead with the project, as instigated by Mrowa.208 As mentioned, Mrowa’s proposed simplification 205 Walter Tracy to Jackson Burke, Letter, “Arabic,” 3 September 1954, Box P3640, NMAH. Tracy’s reference to an “Arab” art is a common error which confuses the ethnicity of the Arab people with cultural achievements of the Arabic culture. 206 Ibid., 3. In an interesting note, Tracy remarked that “if it came to the point of making a new Arabic fount according to a simplified scheme, I think we should give thought to the creation of a type face which would meet with greater approval than L & M’s present series – widely used though it is”. This frank assessment illustrates that widespread use did (and does not) necessarily mean popularity, but only attests to a paucity of choice. 207 Griffith to Burke, “Re: Arabic L.&M. Letter of September 3, 1954,” 2, Box P3640, NMAH. Griffith was thus conscious that Arabic as it was composed on a Linotype machine was already severely simplified even before the notion of a deliberate simplification was put forward. 208 That the drive for Arabic simplification originated in Lebanon rather than in Europe or

‫�ب�ا �ب� ب� ب ب� ب ب�ر �ب��د‬ ‫ج‬ � ‫ب‬ ‫�ب� ��ط ب��� �ب��� ب� ��ق� �ب��ك‬ ‫ع‬ ‫ب��م ب� ب� بس ��ه�ه ب��ه�ه �ب��ه�ه‬ ‫ب‬

Mechanical Composition of Arabic

‫ب���س �ب���س‬ ��‫ب��م ب‬ ‫م‬ � �‫ب�ق� بب� ��ق‬

‫�بر‬

‫ب���ل‬

‫ب� ��و‬

Figure 2.41

Forms of �‫ ب‬bāʾ in different contexts. Typeset in Tasmeem Naskh (DecoType), ­­ a typographical model of Ottoman calligraphic practice.

Figure 2.42

The characters for �‫ ب‬bāʾ, “­Linotype Arabic Light with Bold,” specimen (detail), n.d., ­enlarged to 200% linear. DTGC, box Arabic general.

113

scheme was tailored to the requirements of a standard Linotype with a single keyboard and without a side magazine.209 The fount would be restricted to 90 characters, to fit a 90 button keyboard arrangement and the corresponding 90 channel magazine. The character set was therefore determined by the mechanics of the machine, and the design of the typeface would reflect these constraints. Fitting the Arabic alphabet to the limited character set was achieved mainly by assigning multiple roles to single characters. The approach is well illustrated by the character �‫ ب‬bāʾ. In its manuscript form, �‫ ب‬ bāʾ is given numerous context-dependent shapes (Figure 2.41). In its common typographic form, the different shapes of the letter are significantly reduced; one of Linotype’s existing Arabic typefaces, for example, employed five characters for North America is worth emphasising. Any notion that compromises in the printed appearance of Arabic were the result of Western ignorance does not therefore seem tenable. At the same time, the origin in Lebanon of efforts to simplify Arabic may stem in part from the country’s historical ties to French and US schools and universities, where many in Lebanon’s middle and upper classes were (and continue to be) educated (see also “Nasri Khattar’s Endeavours in Arabic Type Design” on page 86). Western ideas of progress and modernity associated with technological advance encountered in such contexts may have contributed to a readiness to accept the expedience of script simplification for the sake of economic advantage; this may not have been the case in other Arabic countries. 209 See also the discussion of linecaster mechanics and their influence on type design and typesetting in “Mergenthaler Linotype’s First Arabic Linecaster Founts” on page 56.

114

Chapter 2 Isolated

Final

Isolated & final

Initial & medial

Medial

Initial

Figure 2.43 Simplification principle of �‫ب‬ bāʾ, adapted from “L’Arabe Simplifié”, ­specimen (detail), 1960/1, actual size. DTGC, box Arabic general.

�‫ ب‬ bāʾ ­(Figure 2.42 on page 113). But Mrowa’s scheme would take this further by employing a mere two characters to render �‫ ب‬ bāʾ in all contexts (Figure 2.43). The same principle was applied to all the letters of the script, but to different degrees and in Tracy’s words only “where this can be done without undue distortion of traditional shapes”.210 The result was a fount streamlined from 102 to 69 alphabetic characters.211 (The remaining 21 characters would consist of figures, punctuation and spaces.) Although this basic approach to a simplified Arabic typeface was in place in 1954, progress on the project at Linotype was subsequently sluggish. Here, it is worth remembering that other developments besides Arabic and Latin founts also occupied the company. As discussed above, Mergenthaler worked on an Urdu Linotype in the mid-1950s, and as demonstrated by Ross, Indian script developments were pursued at the same time.212 Eventually, in April 1957, an impatient Mrowa accused Linotype of neglecting the project.213 To push it towards completion, Mrowa sent his staff lettering artist Nabih Jaroudi to the Linotype office in London.214 During some two weeks of apparently concentrated effort, Jaroudi revised and corrected Linotype’s work on the typeface. 210

Walter Tracy to Dawood, Letter, 30 January 1964, WT correspondence, folder 18b, DTGC.

211

Some of these characters were half forms, which only produced a meaningful letterform in combination with other characters.

212

See “The Printing Trade and the Middle East after the Second World War” on page 104, and Ross, The Printed Bengali Character and Its Evolution, 155–159.

213

During this time Tracy collected material and pursued his research about Arabic, laying the foundations for his continued and successful involvement in Arabic typography over the next twenty years. However, he “did not, as some people think, have to learn the language”. Tracy, “Composing Room Days (and After),” 13.

214

Nabih Ǧarūdī. Walter Tracy to Jackson Burke, Letter, “Simplified Arabic,” 12 April 1957, Box P3640, NMAH.

Mechanical Composition of Arabic

115

Figure 2.44 “Simplified Arabic”, left: (a) proof of the 12D size under development, April 1957, reduced to 50% linear, below: (b) detail in actual size. DTGC, box Simplified Arabic.

­  proof of 69 characters in regular and bold weights shows the simplified deA sign in development (Figure 2.44). The new fount was based on the 14 pt size of Linotype’s ‘Series 2 with 3’.215 Jaroudi’s remarks on the work, recorded in an internal memorandum, include instructions for the design of characters, their approval or rejection, and advice about the character set.216 The remarks in 215

216

Until 1967 Linotype’s Arabic founts were advertised without proper names but had only series numbers. Arabic 2 with 3 described a duplexed typeface, in which regular and bold weights were struck into the same matrices. Individual founts were labelled with codes denoting point size, depth of strike and series. The fount used – 14 ◊ E24 – was thus 14 Anglo-American point size (14), English depth of strike (◊ ), series E24. Founts with a triangle instead of the diamond shape were struck to the US-depth and a capital D after the point size referred to the Didot point system. For a summary of the differences between the two measurement systems see “Beginnings of a New Technological Era” on page 205. The letter ‫ ��س‬sīn was composed from two characters, yet one of the two was missing from this proof. Walter Tracy, “25th April 1957 et Seq: Discussion with Nabih Jaroudi on Simplified Arabic, Commentary on Characters as at April 1957,” 25 April 1957, WT corre-

116

Chapter 2

particular address design flaws Jaroudi observed in specific characters, such as the inconsistently-sized diacritical dots on characters of �‫ ب‬bāʾ and the unsatisfactory shaping of the head of ḥāʾ in its isolated and final positions. ‫ح‬ Linotype revised the design accordingly, and by late 1957 had produced a trial set of matrices for use in al-Hayat. On 13 December, the newspaper’s front page featured a column of text set in the new typeface (Figure 2.45). This earliest surviving instance of its use is telling, as it shows the design next to the paper’s existing typeface manufactured by Linotype’s main competitor, Intertype.217 The Linotype design appears distinctly linear, an effect achieved by minimising curves along its connecting stroke. The result gives a sense of characters connecting and aligning precisely even when the typeface is printed on to absorbent newsprint paper. By contrast, the Intertype typeface has slightly rounded or sagging connections, an unfavourable effect made worse by frequent gaps within letter-groups that interrupt the intended impression of single, fused shapes. But despite these advantages in appearance, several problems can also be observed in this trial. One is the use of a single character (instead of the conventional two) for the ligature ‫ لا‬lām alif in all contexts. This produces atypical word-shapes, analogous to using a Latin uppercase letter in the middle of a ‫ذ‬ word set in lowercase (see, for example, the first word in the headline � ‫ �ل�لا ����ست���ا‬, li l-ustāḏi (for the professor). Another is this character’s pronounced inclination (also found in ‫ ل‬lām) which combines unpleasantly with the more upright, isolated ‫ ا‬alif – a frequent pairing. A third issue is the spelling errors found in the text as typeset, possibly related to the new fount scheme that required the composition of individual letters using two successive keystrokes.218 Following its trial in al-Hayat, work on the typeface continued at Linotype for another two years. Then, in November 1959, in Linotype Matrix, the new typeface was finally announced as the Mrowa-Linotype Simplified Arabic (Figure 2.46 overleaf). An accompanying article highlighted Kamel Mrowa’s central role in the project: The system is the outcome of close co-operation between Linotype and Mr Kamel Mrowa, proprietor of ‘Dar Al-Hayat’ […], one of the principal newspapers in Lebanon. His appreciation of the aesthetics of the Arabic spondence, folder 18b, DTGC. 217 218

The prospect of taking over a customer from the Intertype Corporation would have furthered Linotype’s interest in Mrowa’s proposal to develop Simplified Arabic. � ‫ ا‬aḥsan (better / best) which lacks a See for example the first word of the column ‫ح����س�ن‬ tooth between medial and final character.

Mechanical Composition of Arabic

Figure 2.45

117

Detail of front page of al-Hayat, 13 December 1957, actual size. This is the first known use of a trial version of Simplified Arabic. The type in the decorated frame is the conventional Intertype fount used at this time by al-Hayat. DTGC, box Simplified Arabic.

118

Chapter 2

Figure 2.46

Article announcing the release of Mrowa-Linotype Simplified Arabic, as printed in Linotype’s trade magazine Linotype Matrix, 32, November 1959, 5 ­(detail), reduced to 50% linear.

script and his grasp of the problems of mechanical composition were invaluable to Linotype in the evolution of the new system.219 In announcing their new typeface, Linotype emphasised the practical advantages of Simplified Arabic. Among these were a greater ease both in learning its keyboard layout and in the actual keying of copy, which contributed to faster composition.220 These advantages were important selling points in 219

Linotype & Machinery Ltd, “Linotype Are First in the World to Produce a System of Simplified Arabic for Mechanical Composition from One Magazine,” Linotype Matrix II, no. 32 (November 1959): 5.

220 In tests conducted at al-Hayat in 1958, a 30 % improvement in composing speeds was reported by the newspaper’s Linotype operators, and it was speculated that this might

Mechanical Composition of Arabic

119

the newspaper industry where production speed and efficiency were crucial. Additionally, the Simplified Arabic scheme allowed the operation of Arabic linecasters to be controlled by coded paper-tape for the first time, resulting in the potential for another three-fold increase in composing speed.221 Together with these productivity gains, Simplified Arabic also cost less to purchase since it required significantly fewer matrices than other Arabic typefaces.222 Apart from the practical and economic advantages Linotype claimed for Simplified Arabic, the Linotype Matrix also asserted that “a reader of a newspaper or journal set in Mrowa-Linotype Simplified Arabic finds it no more different from the normal than italic is different from roman”. Linotype backed up the assertion by pointing out that some characters “that would require considerable alteration if simplified, have been retained in their familiar variety of shapes”.223 Customers were assured that the new type would be “entirely satisfactory to readers”, as it “strikes a happy balance between the need for speed and economy on the one hand and the preservation of traditional typographical forms on the other. It is therefore evolutionary rather than revolutionary.” These statements suggest a perception at Linotype that their new typeface might require some explanation, indeed that the simplified principle was so far-reaching as to be potentially unsettling.224 Tracy, probably the author of the article, may not at this point have felt sufficiently well-informed about Arabic typography to justify the radical approach of the new system with full confidence. Instead, responsibility for it is effectively delegated to Mrowa, with Linotype merely providing the infrastructure:

increase to as much as 50 % once they became better acquainted with the new keyboard layout. Walter Tracy to Jackson Burke, Letter, “Simplified Arabic,” 13 January 1958, Box P3640, NMAH. 221

For a description of the technology see “The Beginning of ­Computer-Aided Arabic Com�position” on page 183.

222

Wear and tear on matrices inside the Linotype linecaster meant that they had to be replaced regularly, resulting in on-going costs for users. Income from the supply and replacement of matrices was an important part of the business models developed by typesetting machine manufacturers.

223

Linotype & Machinery Ltd, “Linotype Are First in the World to Produce a System of Simplified Arabic for Mechanical Composition from One Magazine,” 5. Two letters to which this applied were ḥāʾ and ʿayn. Later, ḥāʾ was simplified; see discussion below. ‫ح‬ ‫ع‬ ‫ح‬ No similar statements are found in earlier articles about Arabic Linotype composition; cf. Linotype & Machinery Ltd, “Millions of People Benefit from the Advantages of Linotype Composition in Arabic,” Linotype Matrix II, no. 23 (October 1955): 2.

224

120

Chapter 2

The system is a tribute to the judgment and enterprise of Mr Mrowa and to the typographic resources of Linotype; and it is of the greatest potential benefit to the whole Arab world.225 It may be conjectured that the conspicuous credit given to Mrowa was prompted by a degree of caution on the part of Linotype to embrace the new typeface. But if the company was exercising caution, it was wholly unnecessary in light of the typeface’s subsequent success. By the next issue of Linotype ­Matrix, in May 1960, Simplified Arabic featured prominently on the cover as one of the “New faces of the 50s” and exemplary of Linotype’s innovation in type design (Figure 2.47).226 The company’s confidence in the new design is also demonstrated by the rapid addition of sizes: by December 1960, 10D and 18D founts were available for purchase. The additions were publicized in Linotype Matrix, 34 (Figure 2.48), which noted the “remarkable success” that Simplified Arabic had achieved “in the short period since it was announced”.227 The release of Mrowa-Linotype Simplified Arabic, Series 2 with 3, marked the start of a new position of influence that Linotype would assume in Arabic typography.228 Intertype Abridged Arabic Soon after Simplified Arabic’s entry into the market and its evident early success, Linotype’s competitor, the Intertype Corporation, began to develop a simplified Arabic typeface of its own. Like Linotype, Intertype manufactured ­machines for hot metal type composition. The mechanics of the Intertype were notably based on those of the Linotype and its introduction in 1913 soon after the expiry of Mergenthaler’s patents was not coincidental. But the Intertype also incorporated improvements on its Linotype rival, as its construction was informed by “research into the principal defects and most troublesome features of extant

225

Linotype & Machinery Ltd, “Linotype Are First in the World to Produce a System of Simplified Arabic for Mechanical Composition from One Magazine,” 5.

226

Both lines contain a composition error of the company’s name which could be interpreted as testimony to Linotype’s inexperience with Arabic typography.

227

Linotype & Machinery Ltd, “Now Available: 3 Sizes of Mrowa-Linotype Simplified Arabic,” Linotype Matrix II, no. 34 (December 1960): 2. Development and manufacture of matrices were often slowed-down by the limited production capacities of the works and 12 months for two weights can be considered relatively fast.

228

The full name of this first version is in reference to the earlier typeface it was based on.

Mechanical Composition of Arabic

121

Figure 2.47 Left: Cover page of Linotype Matrix, 33, May 1960, reduced to 20% linear. Above: detail thereof, actual size.

Figure 2.48

Specimen of three additional sizes of Mrowa-Linotype Simplified Arabic, as reproduced in Linotype Matrix, 34, December 1960, 2, reduced to 70% linear. DTGC, box Simplified Arabic.

122

Chapter 2

linecasters”.229 After a difficult start, the Intertype eventually became a formidable competitor to the Linotype and by 1957 some 27,000 had been sold.230 While Intertype’s activities in Arabic typography are not well documented, it appears that the company had established itself in the Middle East by the 1930s and counted among its clients the prestigious Egyptian daily al-Ahram; 231 by the 1950s, Mrowa’s al-Hayat was also using Intertype machines. After the successful launch of Simplified Arabic in 1959, and possibly influenced by alHayat’s consequent shift to Linotype machines, competition between the two companies intensified. Thus in January 1960, only a few months after Simplified Arabic was made public, a proof showing of a trial Intertype design on a similar, simplified basis was obtained by Walter Tracy (Figure 2.49). Like Linotype, it appears that Intertype derived its design from an existing typeface in its inventory. But the design’s overall appearance in proof is crude and is marred by composition errors probably related to the fount scheme adopted. Remnants of a strong horizontal connecting stroke on the right side of numerous characters, retained for composition in medial positions, impede their use ‫ف‬ at the beginning of words (see, for example, �� fāʾ ) and increase the type’s visual unevenness. For its character set, Intertype’s design shows a number of differences from Linotype’s Simplified Arabic: ‫ �ه‬hāʾ, for example, is represented by two rather than three distinct characters, omitting a dedicated character for medial positions (an omission that was later reinstated); ‫ لا‬lām alif, by contrast, retains two distinct characters for isolated and final positions, where Linotype employs only one. Issues such as these show that despite copying a concept already established by Linotype, Intertype’s design process was not without difficulties, and did not at first produce a convincing result. By June 1960, Intertype was able to provide its customers in the Middle East with a more advanced specimen of a single size of their simplified typeface (16 pt), and announced a further three sizes in development (9, 12 and 18 pt)232 229

The Intertype Corporation was founded in 1911 as the International Typesetting Machine Company. Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 10.

230 Seybold estimates that “Intertype’s sales from 1950 on seem to have equalled or exceeded those of Mergenthaler. In 1956, for example, Intertype shipped 1,150 units”. Seybold, The World of Digital Typesetting, 41. Wallis states that Intertype was only able to secure a firm position in the market in 1918, helped by an order for 31 linecasters from The New York Times. Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 12. 231

Baḥarī, Ṣaḥāfä al-Ġadu, 138.

232

The circumstance under which Linotype obtained this specimen sheds light on the secrecy assigned to new typographical developments. Although the Intertype Corporation had an office in Slough, Buckinghamshire, Linotype had to acquire a specimen through

Mechanical Composition of Arabic

123

Figure 2.49

Left: Intertype Abridged Arabic, early proof of the fount under development, Harris-Intertype Ltd, 1960, actual size. DTGC, box Simplified Arabic.

Figure 2.50

Right: “Namūḏaǧ min al-ʿarabiyy al-muḫtaṣar ǧism 16 bunṭ,” detail of advance specimen, Harris-Intertype Ltd, June 1960, actual size. DTGC, box Simplified Arabic.

(Figure 2.50). The type shows improved alignment between characters and less noticeable right-hand joining strokes. A pronounced horizontality in the lines of text and the shapes of some characters such as ‫ د‬dāl suggest a move towards the Mrowa-Linotype design. But in the typeface finally released by Intertype, additional styling and character set features were incorporated that are distinctive and that indeed improve on Linotype’s simplification scheme. its agent Michael Nahas in Beirut, who explains in the accompanying letter: “This only specimen copy was obtained from our friend mechanic, which should be returned back, as soon as possible, to the customer from whom he borrowed it”. Michael A. Nahas to Walter Tracy, Letter, 7 June 1960, Simplified Arabic box, DTGC.

124

Chapter 2

Notably, Intertype reduced the characters required for ḥāʾ to just one for ‫ح‬ initial and medial position; the letterform conceived for this purpose was better integrated stylistically. The reduction of the ḥāʾ characters also freed up ‫ح‬ positions in the magazine and on the keyboard for other characters. Among these were two for �‫ �ي‬yāʾ, in final and isolated positions, the inclusion of which helped bring words closer to their conventional shapes. In the latter half of 1960, or sometime in 1961, the new design was released as Intertype Abridged Arabic, in four sizes;233 by the beginning of 1962, al-Ahram was using the typeface for large quantities of text (Figure 2.51). In retrospect, the development of Abridged Arabic can be seen as consistent with Intertype’s policy of adapting competitors’ typefaces, a policy that today might be considered unethical, and indeed competitors at the time found it objectionable.234 But if Intertype Abridged Arabic was to a large extent based on the MrowaLinotype Simplified Arabic, it also brought new and different qualities to the simplification of Arabic that were recognised by Linotype as improvements. A New Design: Series 8 with 9 Despite the substantial investment in research and development by Linotype and Mrowa, the patents reportedly registered for its simplification scheme, and Intertype’s putative infringement, no legal action was taken by Linotype.235 While Linotype’s resentment towards Intertype was apparently shared by others in the industry, the differences in Intertype’s simplification scheme and in the design of the typeface itself were apparently sufficient to make a lawsuit for infringement too uncertain.236 233

A lack of documentary evidence has made it difficult to establish a more precise date of release.

234

In a letter to C. A. Ainsworth, breaking the news of the Intertype Abridged face, Tracy characterises it unambiguously: “It is obvious that the Intertype Abridged Arabic is substantially the same scheme as Linotype Simplified Arabic. In view of the fact that we have taken the trouble to apply for patents for our own scheme in a number of countries, it seems necessary to consider whether action should be taken against Intertype.” Walter Tracy to C. A. Ainsworth, Letter, 5 July 1960, Simplified Arabic box, DTGC.

235

While promotional literature and correspondence claim the system to be patented, no such patent could be discovered.

236

In 1962 the Monotype Corporation approached Linotype to establish the reach of the patent and to obtain the permission to adopt the scheme to their machinery. An internal company memorandum quotes the typographical adviser John Dreyfus as follows:

Mechanical Composition of Arabic

Figure 2.51

125

Detail of al-Ahram, 5 February 1962, p. 1, actual size. This cutting fom the newspaper shows the bold weight of Intertype Abridged Arabic alongside a larger size of Intertype’s normal Arabic fount (opening paragraph), together with hand-­ lettered headlines. DTGC, box Simplified Arabic.

126

Chapter 2

At the time Intertype released its Abridged Arabic, Linotype had further sizes of Series 2 with 3 in development. This apparent commitment to the design would seem to make the prospect of an immediate successor improbable. And yet Linotype embarked on precisely this course when it commissioned a new version of the Mrowa-Linotype Simplified Arabic from Nabih Jaroudi.237 Now, however, Linotype would pursue an entirely new design rather than adapt an existing typeface.238 Work on this new design progressed much faster than had been the case with its predecessor and by 1962 a first 12D size was ready. A specimen of the “new design in Mrowa-Linotype Simplified Arabic” credited the typeface to Jaroudi and announced that additional sizes were underway (Figure 2.52). Linotype’s new design was linked to a new keyboard layout. A diagram dated September 1962 shows how the new layout varied from its predecessor, and where characters had been added or repositioned (Figure 2.53). Ironically, the changes reflect exactly those improvements Intertype had introduced to their Abridged Arabic: for example in lieu of four forms of ḥāʾ, only three ‫ح‬ characters are employed, allowing additional characters for �‫ �ي‬yāʾ and ‫ لا‬lām alif to be introduced. But overall, the new design benefitted considerably from being wholly conceived within the now well understood limitations that simplification placed on character set and composition scheme. The design of the characters also changed: rounded and curved elements were reduced in favour of sharper and more angular ones, producing a more linear and even effect; text composition looked streamlined and efficient – appropriately so for a newspaper. The new typeface also benefitted from an apparently improved

“To Tracy’s regret Linotype decided not to go to law against Intertype over this piracy […] As a member of A.Type.I. [sic], I naturally cannot agree that it would be right for the Corporation to steal this patented system from Linotype. The fact that the system has been patented is proof of the fact that it is intrinsically worthy of protection. It would also be unseemly for the Corporation (as a member of A.Typ.I. [sic]) to infringe the Linotype Company’s rights in this design, despite the fact that Intertype have done so”. Anonymous to D. G. Fletcher Rogers, Letter, “Simplified Arabic (Mrowa-Linotype),” 7 June 1962, 1, correspondence folder Arabic (Egyptian 2), MT. 237

This new version would become known as Series 8 with 9. While there is no confirming evidence, the timing of the initiative suggests that it was in part prompted by the introduction of Intertype’s Abridged Arabic.

238

In Walter Tracy’s recollection, the existence of the competitor is entirely omitted, suggesting a purely quality-driven decision: “There was a mixed reception [of the first version], and we were urged to hire a professional scribe to design a new face according to the simplified principle”. Walter Tracy, “Composing Room Days (and After),” 13.

Mechanical Composition of Arabic

127

Figure 2.52 “A new design in Mrowa-Linotype Simplified Arabic”, specimen of 12D size, Linotype, 1962, reduced to 40% linear. DTGC, box Simplified Arabic.

Figure 2.53

Keyboard layouts for Mrowa-Linotype Simplified Arabic, original (above) and revised (below), Linotype, 1962. Keys marked (×) in the revised layout (enlarged in boxes) indicate character set changes copied from Intertype’s scheme. DTGC, box Simplified Arabic.

128

Chapter 2

manu­facture, resulting in precise alignments that minimized gaps between characters and gave the desired impression of fused letter groups. In the succeeding years, additional sizes of the new Mrowa-Linotype Simplified Arabic became available. A specimen from 1963 showed founts in 9D and 10D; an 18D size followed in 1965, and in 1966 a 7D fount. The continuous expansion of sizes indicates that the typeface was indeed popular.239 But if so, the road to popularity was not entirely straight. The profound differences in appearance between Simplified Arabic and other Arabic typefaces that more closely resembled manuscript letterforms ensured that Arabic readers would not embrace the concept of simplification universally. In a letter of December 1969, some ten years after the introduction of Simplified Arabic, Linotype’s Middle East representative Ralph Goodman listed those places where the system had been accepted (Dubai, Lebanon, Kuwait, Syria, Tunisia, Bahrain, Libya, Sudan, Aden, Egypt), where Simplified Arabic founts had been sold but were not wholly accepted (Algeria, Iraq, Saudi Arabia), and where Simplified Arabic had “not yet broken in” (Jordan and Morocco).240 The factors Goodman identified as crucial to the acceptance of Simplified Arabic were pragmatic ones: Of course, where we have broken in, as with Tunisia a few years back, it is mainly for newspapers, and there is still some resistance to using Simplified for bookwork or for Government work, and certainly for the Koran. But the overwhelming pressure of price, speed and ability to move from hot-metal manual on to tape and eventually photocomposition systems, all work in favour of Simplified. It is considerably cheaper to buy a machine using Simplified, a fount of matrices is cheaper in Simplified – the operator can obtain greater speeds when he no longer has the side magazine Keyboard to think about, he can be trained to touch-type as in Roman, […] the printer can move on to tape, and even computers […] and one day photocomposition.241 Goodman’s remarks indicate that the case for Simplified Arabic mainly revolved around the economics of production, as indeed it had for Kamel Mrowa initially. The implication is that gains in efficiency, speed and cost savings would eventually overcome concerns about aesthetics and stylistic ­appropriateness. 239

Walter Tracy claimed that it was “amongst the most popular of all Arabic types”. Ibid.

240 Ralph Goodman to Barnard, Letter, 10 December 1969, WT correspondence, folder 18a, DTGC. 241

Ibid., 2.

Mechanical Composition of Arabic

129

And indeed most newspapers in Arabic countries did eventually adopt simplified typefaces, whether by Linotype or by other manufacturers, making this form of printed Arabic widely read throughout the Middle East. Simplified Arabic’s growing popularity was supported by on-going refinements, though a change of name may have helped, too. As noted above, prior to the making of Simplified Arabic, Linotype’s Arabic typefaces had been identified only by series numbers that for some were very probably confusing or difficult to remember and which in any case hampered effective marketing.242 To address the situation, in 1967 Linotype’s Egypt representative, Hrant Gabeyan, sought and suggested suitable names for all the company’s existing Arabic typefaces, following the practice long established for Latin script typefaces. Mrowa-Linotype Simplified Arabic Series 8 with 9 became Yakout, after the 13th-century calligrapher Yāqūt al-Mustaʿṣimī (Figure 2.54 overleaf).243 Simplified Arabic was conceived in the specific historical circumstances of the Middle East in the 1950s and exemplifies the competing forces that shaped the evolution of Arabic typography in the following three decades. The typeface’s fundamental achievement at the time lay in the original concept, established by Kamel Mrowa for the practical needs of newspaper production. The simplification principles were the primary reason for its success and its widespread adoption, its advantages over earlier fount arrangements making it the obvious choice wherever speed and efficiency overruled concerns for aesthetics or adherence to established conventions. Conversely, although the concept was developed for a soon-to-be obsolescent composition system, the type also conveyed the sense of modernity and progress that was called for. It provided an answer to the urgent needs of a newspaper world in the making, of national presses fuelled by an era of upheaval and change in the Middle East, and (with some contradiction) it also stood for advance through technology.

242

The founts of the first version of Mrowa-Linotype Simplified Arabic, Series 2 with 3, were called 10◊ E26, 14◊ E24 and 18D ◊ E95 for its 10, 14 and 18 Didot sizes, the successor design Series 8 with 9 7D ◊ G69, 9D ◊ G43, 10D ◊ G44, 11D ◊ G41, 14 ◊ G15 and 18D ◊ G48 for its respective size range.

243

Gabeyan was never credited to have conceived the name of one of the widest selling and most read and copied Arabic typefaces in history. Its current incarnation is still marketed by the name Yakout today; ironically its current trademark holder misspells Jaroudi’s name and provides an erroneous date for the typeface’s design. Monotype GmbH, “Yakout Font Family,” n.d., http://www.linotype.com/1618/yakout-family.html (accessed 16 May 2017).

130

Figure 2.54

Chapter 2

“Yakout”, the typeface formerly known as Mrowa-Linotype Simplified Arabic, specimen, not dated, Arabic language cover, reduced to 70% linear. DTGC, box Simplified Arabic.

Mechanical Composition of Arabic

131

As the success of the typeface grew, the roles played by Mrowa and, to a lesser degree, by Jaroudi gradually fell from view.244 Linotype dropped references to Mrowa from their specimens, and by the time of his death in 1966 nothing in their promotional literature suggested authorship by anyone other than Linotype.245 Monotype Hot-Metal Composition The Monotype composing machine was invented by Tolbert Lanston in Washington, DC. First patented in 1887, it was manufactured in modest numbers from 1889 by the firm of Sellers & Co. in Philadelphia, Pennsylvania.246 Of the first 50 machines, four were shipped to the United Kingdom in 1897 in search of investors for the fledgling company. A coincidental encounter of Lanston’s associates on their voyage across the Atlantic with the Earl of Dunraven led to the establishment of the Monotype Machine Syndicate and the purchase of British and Empire (except Canada) rights, relieving the company of its financial woes.247 The British Lanston Monotype Corporation was founded the same year, and a manufacturing plant was built in Salfords, Surrey, half way

244 The last published crediting of Mrowa is in a 1964 article by Tracy in which, contrary to most of his writings, contributors are acknowledged by name. Tracy, “The Flourishing Reed,” 145. 245

Earlier specimens referred to the typeface as Mrowa-Linotype Simplified Arabic. In the latest additions to the size range, 18D pt in 1965 and 7D pt in 1966, the face was designated only as Simplified Arabic. Kamel Mrowa was assassinated on 16 May 1966. In an account of his own career, Tracy makes the somewhat flippant remark that “not long after [the development of Simplified Arabic], he [Mrowa] was shot dead as he left his office one day; for political reasons, not typographic”. Walter Tracy, “Composing Room Days (and After),” 13. Tracy appears not to have appreciated the actual political weight of the killing. Mrowa’s death occurred at the start of a series of events that ultimately culminated in the Lebanese civil war from 1975 to 1990. Historian Charles Winslow argues that his assassination, although “not usually emphasized in these accounts, may well have begun the ‘hostility-reaction formation’ that brought on the protracted civil chaos that wracked Lebanon for nearly two decades”. Charles Winslow, Lebanon: War and Politics in a Fragmented Society (London & NY: Routledge, 1996), 152.

246 Seybold, Fundamentals of Modern Composition, 44. 247

John Dreyfus, “Foreword,” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 3. Seybold gives a purchase price of one million US dollars. Seybold, Fundamentals of Modern Composition.

132

Figure 2.55

Chapter 2

Monotype DD style keyboard (left) and caster (right). From Legros and Grant, Typographical Printing-Surfaces.

between London and Brighton.248 Its mission, undoubtedly related to the position the British Empire still held at the end of the nineteenth century, was to market the product throughout the world, with Lanston Monotype in charge of the Americas. Born of the same quest to mechanise composition as the Linotype, the delay in introducing the Monotype machine was significant. In 1895, two years before the British Monotype Corporation was established, the US Mergenthaler Linotype Company was already shipping 1,076 machines; the Monotype works at Salfords were only operational in 1900, whereas by 1904, 8,000 Linotypes were already in use in the United States alone.249 Arguably, the role of the Monotype system in the trade remained influenced by its later start, and by its mechanical differences from the Linotype.

248 Its name was changed in 1931 to The Monotype Corporation Ltd to distinguish it from the US Lanston Monotype Corporation. In the current context only the British company will be discussed as all Arabic developments originated in the United Kingdom and the terms Monotype and ‘the Corporation’ are used synonymously for the same enterprise. 249 Frank J. Romano, Machine Writing and Typesetting: The Story of Sholes and Mergenthaler and the Invention of the Typewriter and the Linotype (Saleh, New Jersey: GAMA, 1986), 104.

Mechanical Composition of Arabic

Figure 2.56

133

Monotype D style keyboard, counting and perforating mechanism. From The Monotype System, (Philadelphia: Lanston Monotype Machine Co, 1912). http://circuitousroot.com/

The mechanics of the Monotype have been described by numerous authors in great detail, beyond that which is required for the current study.250 Yet, a brief summary of the most significant characteristics is provided here in order to make their influence on the design of the Arabic typefaces more comprehensible. In contrast to the Linotype, the Monotype system cast individual type sorts. It entailed two distinct devices: a keyboard that produced spools of perforated paper-tape, and a caster that was driven by this coded type (Figure 2.55). The system’s flexibility, in many ways akin to manual composition of foundry type, made it the preferred choice for book typography and the “higher quality setting of particular non-Latin scripts”.251 The Monotype keyboard accommodated 274 keys which initiated the punching of a hole in the

250 For an early and detailed technical account see Legros and Grant, Typographical Printing-Surfaces, for a summary with the advantage of historical distance see Seybold, The World of Digital Typesetting, and for the specific influence of the mechanics on Latin typeface design see Southall, Printer’s Type in the Twentieth Century, and Southall’s most recent and comprehensive publication on the subject “Technical History of the Monotype Composing Machines,” in Andrew Boag and Christopher Burke eds. History of the Monotype Corporation (Woodstock: Vanbrugh Press, 2014), 303–414. 251

Ross, The Printed Bengali Character and Its Evolution, 162.

134

Chapter 2

paper-tape ­(Figure 2.56 on page 133).252 The hole’s position signified a unique code which corresponded to an action of the caster. When the tape was fed into the caster, its reading mechanism advanced it over a series of pipes ejecting compressed air. The flow of the air was only possible where the pipes met with a hole, thus controlling the movement of the caster. The majority of these codes related to the positioning of one of the 225 matrices, held in the matrix case, in front of an adjustable mould.253 Once the matrix case was positioned a sort was cast to the correct width and ejected into a galley. Justification of lines could be specified when keyboarding and so the composed type could be used for immediate page make-up in much the same way as manually composed foundry type. Among the advantages was the ease of correction, as a single sort could be replaced instead of the entire line obtained from the Linotype. Moreover, composed matter could be stored for renewed use in the form of tapes, rather than standing type, which made transportation easier. Justification of lines, according to Ross, “perhaps the most significant aspect of Monotype composition”, relied on a measurement scheme that allocated a precise number of predefined units to each character.254 Because of the separation of keyboarding and casting, a width measurement had to be employed.255 Excepting Benton’s self-spacing type, Monotype therefore introduced the first counting routine to the trade with its relative-unit system.256 It was based on a division of the em-quad into 18 equal width units, and the assignment of multiples of these units to all the characters of a fount; the minimum width was five eighteenths, the maximum eighteen.257 A mechanism on the keyboard calculated the sum of units of every character keyed, and notified the ­operator

252

A number of keyboard layouts catered to different tasks and operators could customise their keyboard arrangement for specific jobs, offering a degree of flexibility that contributed to Monotype’s reputation for high-quality typography in the trade.

253

Commonly the matrices were held in a 15×15 arrangement but the introduction in 1963 of the Unit Shift mechanism, in combination with a 16×17 arrangement allowed for a greater variety of character widths. Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 33.

254

Ross, The Printed Bengali Character and Its Evolution, 163.

255

In Linotype composition the line-end decisions could be made by the operator who was in view of the matrices, allowing him to judge whether a word still fitted into the line.

256

Seybold, Fundamentals of Modern Composition, 46.

257

The Monotype System. A Book for Owners & Operators of Monotypes (Philadelphia: Lanston Monotype Machine Co., 1912), 22. In the development of Monotype’s Arabic founts the narrowest moulds were three and four units wide.

Mechanical Composition of Arabic

Figure 2.57

135

The matrix case in its pre-1963 configuration with 15×15 rows. From Legros and Grant, Typographical Printing-Surfaces.

when the end of the line was reached.258 The counting mechanism then indicated on a scale the necessary space to fill out the line for the operator to employ the corresponding spacing keys. This unit system attained a crucial role because all other type-related elements of the Monotype machine depended on its logic. Its influence is most palpable in the organisation of the matrices in the matrix case, in the so-called Matrix Case Arrangement (henceforth MCA) (Figure 2.57). Specific unit values were assigned to every row and the matrices were placed in the position most appropriate to the character they contained; wide letters occupying the rows for the highest unit numbers, narrow letters those for the lowest.259 Despite its apparent constraints this system also offered some flexibility, as the position of individual matrices (and the corresponding keyboard arrangement) could be modified according to each fount and where necessary by the operator (Figure 2.58 overleaf). In addition to the shuffling of matrices, ‘double matrices’ could also be employed. The size of cellular matrices of Monotype founts at 12 pt or smaller was typically 0.2″ × 0.2″, (Figure 2.59 overleaf) with the 258

A bell alerted the operator when the line was within four ems of being complete; the later D keyboard used signal lights for the same purpose. Ibid., 276.

259

Seybold points out that this was particularly noticeable for italic and bold faces, as they could be assigned units most fitting to their characteristics. This contrasted starkly to the Linotype founts that employed duplexed matrices, requiring the bold or italic version of a character to occupy the same width as its roman counterpart. Seybold, Fundamentals of Modern Composition, 56.

136

Figure 2.58

Chapter 2

Partly filled matrix case, demonstrating how individual matrices could be ­organised according to specific needs. Photograph by the author.

Figure 2.59 Above: Drawing of a 0.2″ × 0.2″ Monotype matrix, ­actual size. From “Monotype Matrix Information”, n.d. http://circuitousroot.com/ Figure 2.60 Left: Comparison of different MCAs; top: arrangement as used for 12 pt and smaller; bottom: arrangement as used for 14 pt, 16 pt and 18 pt. From Legros and Grant, Typographical Printing-Surfaces.

­ ossibility of including 0.2″ × 0.4″ matrices for characters that required more p space.260 Larger type sizes, however, relied on the use of bigger matrices to 260 “Display figure matrices are the same width as 0.2" × 0.2" cellular matrices, but are twice the height. The matrices are high enough for 30-point figures. Each 0.2" × 0.4" figure matrix takes the place of two 0.2" × 0.2" matrices in the matrix case. The body of a display figure is the same size as the type with which it is used, but the face of the figure overhangs this body on a platform which rests upon high quads in the lines above.” “Monotype Matrix Information. For the Use and Convenience of Monotype Owners and Operators.” (The Monotype Corporation, n.d.), 1.

Mechanical Composition of Arabic

137

a­ ccommodate wide characters, resulting in a reduction of characters available in the matrix case (Figure 2.60).261 A bigger type size thus prohibited a large character set, a physical constraint that presented considerable difficulties for the adaptation of non-Latin scripts to the Monotype. Initial scepticism from the trade regarding the quality of mechanically composed type was overcome early in the twentieth century when in 1911 the last technical hurdle for full kerning capabilities of the Monotype caster were addressed.262 Henceforth it was generally agreed that the Monotype system “had enough flexibility to reduce the limitations to insignificance so far as character design was concerned”.263 Yet, as will be shown, for Arabic and other non-Latin scripts this notion is contestable, as often the increased complexity of letter shapes faced more severe constraints from the unit system and other limitations of the machine.264 ∵

261

“Matrices for 14, 16, and 18 point are made in two sizes, 0.2″ × 0.4″ and 0.4″ × 0.4″, which are combined for use in one matrix case”. Ibid.

262

Christopher Burke notes that the effort to achieve historical fidelity of the Monotype Caslon revival “seems to have provided the impetus to finally implement the degree of kerning which Monotype machines could achieve”. Christopher Burke, “The Early Years 1900–1922,” The Monotype Recorder, Centenary Issue, One Hundred Years of Type Making 1897–1997 no. 10, New Series (1997): 10.

263

Tracy, Letters of Credit, 40. This stance can be seen as representative for the trade as a whole but there were also critics of these limitations, most explicitly uttered by Dutch designer Jan van Krimpen: “If I have formerly devised means to improve the attempts of the Corporation to copy existing faces – in the first place my own on which we have been working together – I have now found that this is essentially a hopeless enterprise. What can be done is, with the help of the designer if he is available, make a Monotype version after what originally has been a face produced by a type-foundry which is, at the very least, unlikely to have taken the Monotype limitations – if only the limited number of possible widths imposed by the system of unit arrangements – into account.” Jan van Krimpen, “On Preparing Designs for Monotype Faces,” in Type & Typography. Highlights from the Matrix, the review for printers & bibliophiles (New York: Mark Batty Publisher, 2003), 60.

264 Ross suggests that the proportions of a number of type forms found in the Monotype Bengali Series 470 may have been unfavourably influenced by the relative-unit system. Ross, The Printed Bengali Character and Its Evolution, 168.

138

Chapter 2

Here it is worth bearing in mind that the Monotype Corporation’s core business was the sale of machines, and typefaces mainly served as a means to this end. Innovation in design was not important, but the ability to create types acceptable to printers was; according to John Dreyfus typeface design was for many years considered the “Cinderella of the Monotype Works”.265 It follows that the Corporation’s policy of type design was conservative, as its primary role was to support the establishment of Monotype composition: The Lanston Monotype Corporation viewed its typefaces as commercial fodder for the machine, which, after all, was the initial raison d’être of the company. No pretence was made to originality; in the same way that Gutenberg’s first books proved their value by effectively matching their manuscript predecessors, Monotype setting had to match the look of hand composition.266 This changed from the 1920s, notably with the arrival of Stanley Morison as typographical adviser, when Monotype set out on an ambitious programme of typeface developments.267 Numerous revivals of renowned historical models, as well as contemporary designs, and their skilful promotion contributed to the Corporation’s reputation as a champion of typographical quality. Early in the history of the company typefaces were also produced with the purpose of opening new markets worldwide. In 1903 a Gaelic typeface was developed for Ireland, the first Fraktur and Schwabacher types for the German-speaking market followed in 1904, and in 1910 the first Greek Monotype face was cut; by 1915, 133 Fraktur faces and 42 Cyrillic faces were advertised by the company’s journal The Monotype Recorder.268 After the disruption of the First World War which saw the Salfords factory used for the production of machine-gun parts, Monotype’s international typographical programme expanded gradually beyond Europe and the order of new developments roughly corresponded to the importance and economic standing of territories within the British Empire. Among the earliest non-European types were Hebrew, Devanagari and Tamil, publicised in the Monotype Newsletter in December 1934 and reflecting 265

Dreyfus, “Foreword,” 3.

266 Burke, “The Early Years 1900–1922,” 4–5. 267

See for example Sebastian Carter, ‘The Morison years and beyond: 1923–1965’, The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 14–25; and James Moran, Stanley Morison. His typographic achievement (London: Lund Humphries Publishers Ltd, 1971).

268 Burke, “The Early Years 1900–1922,” 6.

Mechanical Composition of Arabic

139

the British presence in India and Palestine.269 Support for Indian languages was further expanded in the following years with the addition of Bengali in 1937, and according to promotional material Telugu and Gujarati could also be made available.270 Monotype’s First Arabic Typeface: Series 507 Monotype’s Arabic programme was initiated in early 1938 in response to demand from India.271 Through its Bombay office, the Corporation learned from Mohammad Elias Burney, the director of the Translation Bureau of the Osmania University in Hyderabad about the institution’s intent to acquire Arabic composition machines.272 The University was the first in India to teach in the vernacular Urdu, and Burney was going to recommend the purchase of Intertype casters for Arabic script composition.273 By implication this meant that its publications would be printed in the culturally inappropriate Nasḫ, as no Nastaʿlīq type existed on the Intertype. But it appears that Monotype was able to convince Burney of the potential benefits of its equipment, and he agreed to withhold his proposal for the procurement until the Corporation could ­commit to developing an Arabic fount.274 With this prospect initial investigations into the feasibility of Arabic composition on the Monotype began. 269 The Monotype Corporation Ltd, Monotype Newsletter, Holiday Number 7 (Monotype Corporation Limited, 1934). 270

Ross, The Printed Bengali Character and Its Evolution, 168. In 1937 the Monotype Newsletter called for clients “interested in the mechanical composition of Telugu, Gujarati, Tamil, Sanksrit, Devanagari etc.” to apply for specimens. The confusion of the Sanskrit language as a script indicates the inexperience of the Corporation with foreign languages and non-Latin scripts during this period, and renders the accuracy of this source questionable. The Monotype Corporation Ltd, Monotype Newsletter, 33 (October 1937).

271

Robinson shows that lithographic printing flourished on the subcontinent from the late nineteenth century when printing in the Middle East was still marginal. The general delay in the development of a printing trade in the Middle East could explain the origin of Monotype’s first Arabic type in what was then British India. Francis Robinson, “Technology and Religious Change: Islam and the Impact of Print,” 27, no. 1 (February 1993): 229–51.

272

The Monotype Corporation Ltd, “Copy of Letter from Bombay Office to London Office,” 15 February 1938, 2, folder Arabic, MT.

273

Kavita Datla, “A Worldly Vernacular: Urdu at Osmania University,” Modern Asian Studies 43, no. 05 (2009): 1117–48.

274

“Copy of Letter from Bombay Office to London Office,” 2.

140

Chapter 2

The Corporation had already developed a working mechanism for the composition of scripts written from right-to-left in 1910, and employed it successfully with its Hebrew types.275 Having solved this mechanical challenge, the principal questions were thus concerned with the matrix-case arrangement and the corresponding design of the type. In the development of non-Latin typefaces Monotype followed a similar approach to the one it used for most early Latin-script typefaces: established foundry type was adapted for mechanical composition.276 This approach was also taken for Monotype’s first Arabic typeface. In a preliminary step the Corporation sourced samples of popular type from the Egyptian Government Press in Cairo, and the Hyderabad Government Press, India.277 In addition to the identification of a potential model, Monotype’s Bombay office also prepared a draft MCA and a keyboard layout (Figure 2.61 and Figure 2.62). From correspondence it emerges that R. V. Pillai, the director of the Hyderabad Government Press, suggested that the type be cast with a 16 pt face on a 14½ pt body, a wish that could not be implemented because of the matrix size limitations.278 In an attempt to compromise between size and character repertoire, Monotype tried to cut a 14 pt face on 0.2″ × 0.2″ matrices to match the proposed layout whilst maintaining the largest type size possible. By May 1938 a first proof of Series 507 was printed (Figure 2.63 overleaf). This trial shows a small selection of characters, set in a random, unintelligible arrangement. Although the proofs were sent to India to gather opinions, it soon became clear that the setup could not work due to technical reasons. A handwritten note on the proof indicates the failure of the first attempt: “Too large for accommodation of the matrices on the mould. Not proceeded with”.279 275

Pierpont, Frank Hinman. Improvements in or relating to the Casting and Composing of Single Types in Semitic Languages. UK Patent 28,692, filed 6 December 1910, and issued 7 December 1911.

276

For example Porson Greek was cut by Richard Austin and Monotype Bengali Series 470 was based on type made for the Calcutta printer H. L. Mazunder. See Robert Bringhurst, The Elements of Typographic Style (Point Roberts & Vancouver: Hartley & Marks Publishers, 2004), 278; and Ross, The Printed Bengali Character and Its Evolution, 168.

277

The Monotype Corporation, Internal Memorandum, “Arabic,” 4 April 1938, folder Arabic, MT.

278

In order to accommodate the large number of sorts of the Hyderabad fount the preliminary layout was based on 0.2″ × 0.2″ matrices, whereas a 16 pt face required 0.2″ × 0.4″ and 0.4″   × 0.4″ matrices.

279

“Trial No. 1, Arabic Series No. 507-14 Point” (The Monotype Corporation, 24 May 1938), correspondence folder Arabic Urdu 507, MT.

Mechanical Composition of Arabic

141

Figure 2.61

Draft of the Arabic MCA provided by Monotype’s Bombay office, 1938, reduced to 20% linear. MT, correspondence folder Arabic 507.

Figure 2.62

Draft of the Arabic keyboard layout provided by Monotype’s Bombay office, 1938, ­reduced to 20%. linear. MT, correspondence folder Arabic 507.

142

Chapter 2

Figure 2.63 “Trial No. 1 Series No. 507, 14pt,” May 1938, actual size. MT, folder master proofs.

It appears that the large character set could not be reconciled with the proportions of Arabic letterforms and the desired type size, demonstrating that also the Monotype system imposed limitations on the design of Arabic type.280 So the fount was made in the 12D size and by October 1938 a new proof in the smaller size had been printed and approved by the Hyderabad Government Press. Yet its layout continued to cause considerable complications that could not be solved by the Monotype Works alone. Repeated exchanges and consultation with the Bombay office were required to establish a working fount synopsis, which was eventually achieved in March 1939. By the following month a new proof was printed in 12D, cast on a 16 pt body, featuring a complete character set and a fully functional fount (Figure 2.64). The proof shows a text sample alongside the character set, illustrating some of the solutions employed to create a working layout. Notably, Series 507 introduced a novel concept to mechanical Arabic composition. Among the characters were twelve isolated diacritical signs (rows two to four from the top). Void of meaning by themselves, these diacritics could be used in combination with characters that only featured a rasm, thus creating numerous composite letterforms. This division into components allowed re-use of the same characters in multiple contexts to create different meanings without the need for dedicated characters. The principle is illustrated in the sample text in the two words ‫ �أ �م�� ا لم ؤ �م ن�����ن‬amīr al-muʾminīna (prince of the believers), in which the ­diacritical ‫ير � ي‬ 280 Type Drawing Office, “507 Arabic,” card index, 2, minutes box, MT.

Mechanical Composition of Arabic

Figure 2.64

143

“Trial No. 2 Series 507, 12D,” April 1939, reduced to 50%. MT, folder master proofs.

dots below the line are unusually deep. These marks are cast as separate sorts and complete the final two ligature characters, disambiguating them as ‫ �ير‬īr and ‫ ي��ن‬īn. Combined with different diacritics the same rasm characters could also be used to compose other letters without the inclusion of further, dedicated matrices in the matrix case: ‫�برت�رث�ر ن�رئ�ر �ب�ن �ت�ن �ث�ن �ن�ن �ئ�ن‬. By this means more letterforms could be composed than characters accommodated in the matrix case, in effect enlarging its capacity and providing a richer fount synopsis. A minor drawback of this approach is the positioning of the two-dot character too deep below the rasm character – a compromise that was unavoidable due to the physical constraints of metal type. Another feature of the fount synopsis aimed at a reduction of characters is the use of half-forms, a concept also found in Linotype’s early Arabic typefaces.

144

Chapter 2

Proofs of Trial No. 2 were returned again to India where it was received with general approval. Mr Pillai expressed his satisfaction with the cutting but remarked on five characters in the new type that differed from the model and asked for them to be corrected accordingly (underlined in Figure 2.64).281 The Works replied promptly, explaining that all concerned characters diverged from the model because they had to be either widened or condensed to fit the unit widths they had been assigned.282 Whereas the last three were supplementary characters outside the standard MCA and could be modified with relative ease, changes to characters 93 ‫ ر‬rāʾ and 166 ‫كل‬ � a ligature of ‫ ك‬kāf and ‫ل‬ lām – would have required changes to the entire fount layout. The 12D fount was initially approved for manufacture in July 1939 but alignment issues of a number of characters and the persistent need for larger type sizes prolonged the development of Series 507.283 Despite the outbreak of World War II in September 1939, work on the typeface continued and, by April 1940, Monotype’s Matrix Department found a way to increase the size of the face without having to resort to larger matrices. By using a newly-made twobladed 18 pt mould it became possible to cast sorts that had a larger face than usual on the smallest matrices.284 With this technical improvement the 14D size could contain sufficient characters to develop a workable fount synopsis; an innovation not limited to Series 507, as noted by the Works: This scheme has been evolved owing to the special requirements of the Arabic languages, but the general principles can be used in conjunction with any Type [sic] face and with various body sizes.285

281

The Monotype Corporation Ltd, “Arabic - Series 507-12pt.,” 30 May 1939, folder Arabic, MT. The proof is dated 21 April, indicating a relatively fast exchange of just over one month between India and the United Kingdom.

282

The Monotype Corporation Ltd, “Arabic, Series 507-12pt.,” 1 June 1939, correspondence folder Arabic 507, MT.

283

Type Drawing Office, “507 Arabic,” 4–5.

284

“In order to cast a .2 matrix on a 16pt. Didot mould, a two-bladed mould has been made in which the neck is stepped to 12pt. to provide a seating for the .2 matrix. When the two blades work together the opening is covered by a double matrix (.2 × .4) and the type is orthodox in shape. When the top blade is locked forward the opening is covered by a .2 × .2 matrix and while the body of the type is 18pt. Didot the neck is 12pt.” The Monotype Corporation Ltd, “Arabic Composition,” Internal Memorandum, 23 April 1940, correspondence folder Arabic Urdu 507, MT.

285

Ibid.

Mechanical Composition of Arabic

145

Manufacture of the 14 pt size began in due course, but other aspects contributed to lengthen the undertaking: in parallel, Monotype was working on an experimental 18 pt size, the addition of characters for other languages, and continued the revision of character shapes.286 By February 1941 draft texts for promotional material announcing the new size explained its setup: While it is necessary to place the larger characters on double matrices, a considerable proportion of the characters come on standard matrices, thus allowing a normal font of any given Arabic language to be assembled in a matrix case. The three languages so far dealt with are Urdu, Pushtu and Arabic (Egyptian) and in each of these the four complete alphabets, figures, points, accents and the most used ligatures have been included in the Matrix Case Arrangements, while alternative and additional characters are supplied as outside characters.287 Trial No. 4a, dated 19 March shows the type under development (Figure 2.65 overleaf). This fount synopsis features 214 characters in the matrix case, with an additional 53 outside characters that could be substituted for matrices sharing the same width in the default layout, thus leveraging Monotype’s ability to cast individual types for hand insertion into mechanically composed matter. In comparison, contemporary Arabic founts on the Linotype 48SM caster contained a mere 124 matrices. Monotype used this marked difference of capacity for a more comprehensive offer of ligatures and alternative sorts, and additional language support – qualities that became typical of its Arabic founts. The characters outside the matrix case are also noteworthy: featured among them are high-tooth characters for use in contexts where a sequence of teeth may lead to ambiguity and have a detrimental effect on legibility. Although common in Arabic handwriting and quality typography, these additional characters were rarely found in European printing.288

286 Type Drawing Office, “507 Arabic,” 8. 287

The Monotype Corporation Ltd, Internal Memorandum, “Arabic - Composition,” 19 February 1941, 1, correspondence folder Arabic Urdu 507, MT.

288 This and other shortcomings of Arabic typography made by European printers were observed and expressed explicitly by the orientalist Hellmut Ritter in a short article about the rules of Arabic composition. Hellmut Ritter, “Über einige Regeln, die beim Drucken mit arabischen Typen zu beachten sind,” Zeitschrift der Deutschen Morgenländischen Gesellschaft 25, no. 100, Neue Folge (1950): 577–80.

146

Figure 2.65

Chapter 2

“Trial No. 4a Series No. 507, 14D,” March 1941, reduced to 75% linear. MT, folder master proofs.

Mechanical Composition of Arabic

Figure 2.66

“Trial No. 8, Series No. 507, 12D (redesigned),” May 1941, reduced to 75% linear. MT, folder master proofs.

147

148

Chapter 2

The fount’s new arrangement was made possible through the invention of a new mould, developed for the 14 pt size, but which proved satisfactory for other sizes too. In March 1941 therefore Monotype decided to remake the earlier 12 pt fount according to the new scheme.289 The redesigned fount is shown on Trial No. 8 from 9 May and displays text settings of the four languages supported by the default character set (Figure 2.66 on page 147). The Arabic setting illustrates at once the principal advantage the Monotype mechanics had over linecasters: the ability to kern, as seen in the final ‫ و‬wāw, was crucial for the correct spacing of words and lines of text. In contrast to linecaster faces, spaces within words after non-joining characters could be reduced to a minimum, clearly distinguishing them from inter-word spaces, and thus contributing to higher legibility. At the level of character design, the aforementioned compromise in the shape of the frequently occurring ‫ ر‬rāʾ is the most severe issue. The squashed appearance of the character becomes prominent in text setting, and attains a startling effect in the common combination with ‫ و‬wāw. While both letters should effectively have identical tails, the unconventional, excessively narrow shape of ‫ ر‬rāʾ arrests the eye in mid-word. Another deficiency is the inconsistent height of diacritical signs in characters in which the dots should be aligned (see ‫ �ش���ب� ي��ه‬in line two), a shortcoming which is not explicable by technical constraints. Individual character shapes exhibit issues of inconsistency, as illustrated by the final form of ‫ ك‬kāf: its vertical ascending stroke is atypically short and its horizontal bar appears too narrow (see the fourth word in line three). Combined with the unconventional upwards movement, the character appears to fall over backwards and breaks the flow and rhythm of the surrounding text. The most remarkable peculiarity of Series 507 is, however, its use for Urdu, the language for which this typeface was primarily conceived.290 A Design for Urdu Urdu writing culture derives historically from the Persian tradition.291 Notably under the Mughal dynasty starting from the sixteenth century, the Persian 289 Type Drawing Office, “507 Arabic,” 10. 290 Urdu is an Indian language closely related to Hindi. It is the official language of Pakistan and the first language of between 60 and 70 million people, the majority of which live in India. “Ethnologue. Languages of the World.,” n.d., http://www.ethnologue.com/ (accessed 16 May 2017). 291

“The literary development of languages of Central Asia and the Indian Sub-continent us-

Mechanical Composition of Arabic

149

l­ anguage and book arts were introduced and fostered in South Asia, and eventually developed into a distinct style. This cultural heritage and proximity led to the adoption of the Persian Nastaʿlīq hand as the preferred writing style for the Urdu language.292 Similar to the rest of the Islamic world, the Mughal rulers, although champions of the arts in general and the book in particular, did not adopt printing for Urdu:293 We know, for instance, that a Jesuit delegation from Goa visited the court of the Mughal emperor Akhbar in 1580 and presented him with a copy of Christopher Plantin’s finest achievement as a printer, the Polyglot Bible in eight folio volumes (Antwerp, 1568–1573). But no call for the opening of an imperial printing press followed – only apparently complete indifference. Printing could not match the exquisite manuscript productions of the imperial studio, nor supplant the administrative and informationgathering networks sustained by an abundance of scribes.294 Printing in the vernacular began only in the nineteenth century and was soon widely embraced; indeed, more so than in the Middle East during the same period: By the 1820s in the Indian subcontinent Muslim reformist leaders were busily printing tracts. By the 1830s the first Muslim newspapers were being produced. By the 1870s editions of the Quran, and other religious books, were selling in tens of thousands. In the last thirty years of the century, over seven hundred newspapers and magazines in Urdu were ing the Arabic script was slow, as they were overshadowed by Persian, which had a cultural development of many centuries behind it. In consequence there are very few early manuscripts in these languages.” R. F. Hosking and G. M. Meredith-Owens, eds., Handbook of Asian Scripts (London: The British Museum, 1966), 22. 292

Nastaʿlīq is a writing style of the Arabic script developed in fifteenth century Persia, see also chapter 4 “Persian Type and Typography” on page 285. There is a large body of literature on Islamic calligraphy, for a recent standard work see Blair, Islamic Calligraphy.

293

Historian John Richards describes the lack of interest of the Mughal emperors, intellectuals and nobles to printing with foundry type as “perhaps the most puzzling cultural divide” with Europe. He does not mention the difficulty of adopting Nastaʿlīq to the European technology of foundry type, which may serve as an explanation for their reluctance. John F. Richards, The Mughal Empire (Cambridge University Press, 1996), 289.

294 Graham Shaw, “Non-Latin Scripts and Printing Technologies: Triumphs and Tribulations,” in Non-Latin scripts, from metal to digital type (London: St Bride Library, 2012), 18.

150

Chapter 2

started. All who observed the world of printing noted how Muslims understood the power of the press. In Upper India at the beginning of the twentieth century 4000–5000 books were being published in Urdu every decade and there was a newspaper circulation of tens of thousands.295 The printing process that was employed in South Asia was lithography. Lithography, invented by the German Alois Senefelder in 1798, relies on the immiscibility of grease and water.296 It entails the application of an oil-based medium to the flat printing stone to obtain the printing image. The surrounding areas are then etched with a solution that creates a surface which accepts water, but will not accept the greasy printing ink. For the actual printing, the stone is kept wet with water, which repels the greasy ink from the non-printing areas, whereas the image created with the oil-based medium accepts the ink. Printing stone and paper are then pressed together in a printing-press, transferring the ink from the stone’s surface to the paper. In South Asia the process became popular because of its proximity to manual text production. Scribes could write directly on to the lithographic stone, resulting in reproductions with little or no divergence from the customary appearance of manuscript texts. Indeed, the uncompromised reproduction of Nastaʿlīq has been identified as a prerequisite for the acceptance of printed texts by the Urdu reading public. According to Shaw, “Lithography made possible the paradox of the ‘mass-produced manuscript’ which met the criteria of cultural authority which the typeset text could not”.297 In addition to the stylistic preference, the cost of labour was so low that employing scribes was more economical than investing in (inferior) typographical equipment. When Mergenthaler first investigated the possibility of mechanical Urdu typography in the late 1920s, it was reported that a whole page of a newspaper written on a lithographic stone cost 4 Indian Annas, approximately .08 US cents, and the same work on a Linotype would have cost one US dollar.298 In 1930 Mergenthaler’s manager of the Foreign Linotype department expressed the view that it was “absolutely necessary, in the designing of the Urdu characters, that they be produced on the original Urdu slope, right to left, 295

Robinson, “Technology and Religious Change: Islam and the Impact of Print,” 233.

296 Twyman, Printing 1770–1970, 25. 297

Shaw, “Non-Latin Scripts and Printing Technologies,” 24.

298 Deputy Chairman and Managing Director to Chauncey Griffith, Letter, “Urdu,” 19 April 1929, Box P3640, NMAH. From the same document it emerges that most printers were considered too poor to afford a Linotype caster and that even the Government Central Press would only require a single machine for its entire printed output.

Mechanical Composition of Arabic

151

in order to hold to the beauty of the language, which it is desired to protect”.299 It appears that staff at Mergenthaler were acutely aware of the strong cultural preference for the Nastaʿlīq style, and it seems that no conclusive decision regarding an Urdu typeface was made at the time. However, the issue of composing the language on a linecaster remained an actively discussed topic. In 1936 a report about Urdu typography, compiled by the Indian émigré to the United States, Hari Govind Govil was obtained by Mergenthaler.300 It stressed the division between Indian printers about the appropriate writing style, and the fruitless attempts to develop a satisfying Nastaʿlīq foundry type by the Hyderabad government.301 Commenting on Govil’s account, Princeton professor of Indo-Germanic Philology, Harold Bender underlined the difficulties entailed in the prospect of typographical composition of Urdu, and advised Mergenthaler that any investment was probably not profitable as “the demand for printed Urdu, always relatively weak, is less now than it was a decade ago, and is likely to continue to decline”.302 Despite the discouragement from Bender, Linotype pursued the idea of the typographical composition of Urdu. In 1939 the Indian office of the company even patented a fount and related composition scheme, but the typeface seems not to have been implemented to any significant extent.303 299 Manager Foreign Linotype Department, “Urdu on the Linotype,” 17 October 1930, Box P3640, NMAH. 300 Govil had come to the United States in 1920 to study at the Boston Institute of Technology. In 1923 he founded the Orient Society which published the bimonthly magazine Orient on art and culture in New York, and the India Society in 1924. Sarah Fedirka, “Reorienting Modernism: Transnational Exchange in the Modernist Little Magazine Orient,” English Language Notes 49, no. 1 (Spring/Summer 2011): 1. 301

Hari G. Govil, Report on the Urdu Linotype, 6 January 1936, 2, Box P3640, NMAH. Typographical composition of Urdu was attempted in Gilchrist’s dictionary as early as 1787, in what appears to be one of the first movable Nastaʿlīq types. Graham Shaw, Printing in Calcutta to 1800: A Description and Checklist of Printing in Late 18th-Century Calcutta (London: The Bibliographical Society, 1981), 107.

302

Harold H. Bender to Chauncey Griffith, Letter, “Urdu,” 26 January 1936, Box P3640, NMAH.

303

Linotype & Machinery Ltd. An improved manner of and a type font for printing in Urdu or like script. Indian Patent 27030, filed 5 December 1939, and issued 4 July 1940. The experimental Urdu type was based on a scheme developed by Abdul Wadood Quadri. It employed a mere 68 alphabetic characters and could thus be accommodated on a single magazine Linotype. But whilst this radical simplification allowed for mechanical composition of Urdu, the resulting appearance of the script proved unacceptable to the general reading public. Thomas King to Chauncey Griffith, Letter, “Urdu,” 4 April 1939, Box P3640, NMAH. This study foregoes a detailed account of Mergenthaler’s attempts to pro-

152

Chapter 2

Disregarding this documented reluctance of the Urdu readership to accept any written form of the language but Nastaʿlīq, the fact remains that in 1938 Monotype’s client at the Osmania University in Hyderabad commissioned a fount in the Nasḫ style. This choice suggests the strong incentive mechanical composition offered. Pragmatic considerations such as higher production speed and the convenience of a single, closed workflow for all composition needs of the Translation Bureau seem to have outweighed the less tangible concern for cultural appropriateness. Irrespective of the unorthodox Nasḫ style, the fount had to cater for the language’s patterns. Written Urdu is marked by the high frequency of aspirated sounds, written with the ‫ �ھ‬do chashmī he, and the recurrence of ‫ے‬ �� baṛī ye, employed for the vowels /eː/ and /æː/.304 The ‫ے‬ �� baṛī ye is particularly influential on the look of written Urdu, and its typographic rendering is among the most challenging design questions in an Arabic fount. The letter’s final form exhibits a wide horizontal return stroke that projects deeply under preceding letters and its alignment above the baseline contributes greatly to line formation (Figure 2.67). In handwriting any preceding elements such as diacritical dots or vowels are written above or below its horizontal stroke, a flexibility that is absent from cast type. Even if its long tail could be implemented through a technically difficult kern, collisions with the lower dots of preceding characters would be unavoidable.305 In the design of Series 507 a solution to this typographical problem was found that demonstrates the influence of the physical properties of metal type. To allow the crucial direction change of the tail and avoid clashes with preceding characters, the final form was artificially lowered (Figure 2.68). Although this appropriately rendered the important return stroke and allowed its projection below preceding characters, it also caused an unnatural step within the line.306 Whereas this step is unfavourable by itself, the design’s deficiency is most noticeable when the final character is composed alongside the dediduce an acceptable Urdu fount as no tangible influence on the evolution of Arabic type was found. It may merit further, dedicated research. 304

There are a total of eleven distinct aspirated sounds in Urdu. For a detailed statistical analysis of patterns in Urdu see Mahinnaz Mirdehghan, “Persian, Urdu, and Pashto: A Comparative Orthographic Analysis,” Writing Systems Research 2, no. 1 (2010): 9–23.

305

The theoretical possibility of casting dots as separate sorts could circumvent this issue, but would severely complicate the typesetting process.

306 In most current typefaces a different approach is taken: to avoid a vertical step and additional kerning, the top connecting stroke is stretched to the length of the return stroke. This practice is questionable, as it distorts the letterform beyond recognition.

Mechanical Composition of Arabic

153

Figure 2.67

Left: Page of a lithographic copy of a handwritten Bible. Note the frequency and alignment of ‫ے‬ �� Baṛī ye, actual size. The marked difference between handwritten and typographical forms of Urdu contributed to the continued use of manual means of text production. From The Holy Bible in Urdu, revised version, (Lahore: Pakistan Bible Society, n.d.), reduced to 70% linear.

Figure 2.68

Right: “Monotype Urdu Series No. 507,” specimen (detail), February 1951, reduced to 65% linear. MT, folder master proofs.

� baṛī ye with preceding characters. As ligature and final cated ligatures of �‫ے‬ c­ haracter should commonly align on the same level, their differences in shape and position further underline the anomalous appearance of the typeform conceived for the final character (Figure 2.69 overleaf).

154

Chapter 2

Figure 2.69

Composition of characters from Figure 2.68, enlarged to 200% linear.

Although Series 507 was primarily intended for Urdu, the specimen of this language on Trial No. 8 is the least successful (see Figure 2.66 on page 147). Whereas Arabic, Persian and Pashto are rendered in a quality that could typically be expected from mechanical composition in this period, the peculiarities of the type combine with the characteristics of Urdu to create the worst rhythm among the four trial settings. But despite the shortcomings in cultural appropriateness and design, continued additions to the typeface suggest that it was accepted by customers.307 In India, Monotype counted among their clients the Hyderabad Government Press and the West Frontier Province Government Press Peshawar; and in the United Kingdom, Percy Lund Humphries and Cambridge University Press used Series 507. The acceptance of the typeface by clients and its use and continued manufacture can be explained by the general paucity of choice for Urdu typography at the time. Moreover, existing Monotype customers were more likely to adopt Series 507 than invest in a completely new setup from a competing manufacturer. With a Nastaʿlīq fount outside the realm of possibility, Series 507 provided a solution that – while far from satisfactory – enabled Monotype Urdu composition with an intelligible result. In conclusion, Series 507 demonstrates typical and recurring characteristics of Arabic hot-metal founts. Based on an existing typeface it was not an original creation and inherited features from the model it was based on. Its design could take advantage of the relative flexibility that the Monotype had over linecasters, notably through the inclusion of more characters and the ability to kern. The latter technical facility attained a decisive role in the quality of Arabic Monotype faces, setting them apart from contemporary linecasting founts. Yet adaptation to the Monotype system also prescribed design decisions: Arabic type pushed the flexibility of the system to its limits, and its mechanics required some compromises. Notably the number of characters, the type 307

During 1941 demand for a heavier face resulted in the development of Series 542, a design that was never completed. Type Drawing Office, “507 Arabic,” 11. A bold companion to Series 507 was developed from 1960 under the name of Urdu Bold, Series 707. Based closely on the lighter weight, it inherited the same design issues.

Mechanical Composition of Arabic

155

size, proportions and configuration of letterforms were all influenced by the caster’s mechanics. Although the intricate link between unit widths, matrix size and MCA always influenced the design of Monotype founts, the concessions were more acutely felt in complex fount synopses, such as those required for Arabic type. Here it is worth bearing in mind that the trade’s appreciation of the system’s flexibility and the quality of its output were shaped by a Latin-centric perspective. Generally, Latin Monotype founts were less prone to compromises because of their greater simplicity, as demonstrated by a typical character set. A Latin fount synopsis could contain the entire regular face, a companion italic, small capitals, figures, fractions and punctuation. By contrast the entire fount of Series 507 could only accommodate the Arabic alphabetic characters of a single style and a narrow selection of punctuation marks. It follows that fewer characters and the similarity of shapes and proportions in the Latin script allowed for a higher degree of fidelity despite the constraints of the machine. A New Typeface for the Middle East Monotype initially envisaged the use of Series 507 also for Arabic language composition, but as the styling of the typeface was tailored towards the specificities of the Indian market it was not accepted in Egypt.308 Unfavourable feedback from the Middle East established that a different design was needed, leading to a new typeface that corresponded to the region’s typographical preferences. Here, the company’s representative attained a key role. The three principal manufacturers of hot-metal typesetting equipment, Intertype, Linotype, and Monotype, employed representatives in numerous countries early in the twentieth century, and travelling sales people promoted their products around the globe.309 The first issue of Intertype’s journal Interludes boasted that 10,000 of its casters were in use in more than 60 countries, listing its international representatives and their respective territories.310 Monotype maintained a permanent representation in Egypt through the firm of Joseph 308 “The general opinion [is] that the Arabic script you have produced is good, and a step in the right direction, but it does not come up to the standard which is required by them”. Quixley to H. L. Buckle, Letter, 25 January 1944, folder Arabic (early correspondence), MT. 309 For an (anecdotal) account of the life of a Mergenthaler salesperson in the two decades before World War II see Orcutt, Merchant of Alphabets. 310

Intertype Ltd, “Untitled,” Interludes 1, no. 1 (January 1931): 2. Leopoldu Calu of Alexandria was responsible for Egypt, the Sudan and Palestine.

156

Chapter 2

Lindell from 1939, and Linotype’s subsidiary Société Linotype Française had a branch in North Africa from the 1920s.311 Traditionally, such agents were in charge of sales in the region and, as they were often representing multiple companies, only loosely affiliated. Nevertheless, the expertise in local needs and the close contact to clients in the region also lent a consulting role to them, which attained increasing importance with the expansion of the market from the middle of the twentieth century. As an intermediary between local customers and the remote offices in Europe and the United States, representatives provided first-hand information about the local trade, possessed language skills that the manufacturers often lacked, and demonstrated an understanding of the practical needs in daily practice. This position provided a degree of influence on the Arabic programmes of the manufacturers, and representatives frequently took a supportive stance for new developments for the regional market. When the Egyptian publisher Dār al-Maʿārif expressed interest in purchasing Arabic Monotype equipment in 1941, Lindell became an important contact between the client and the company in the United Kingdom.312 By January 1944 the publisher formally requested an estimate for four Monotype composing machines with keyboards and founts in different sizes.313 The prospective client rejected Series 507 for its purposes, and suggested instead that a new design be conceived modelled on a typeface used at the Egyptian Government Press. Lindell thus began to investigate this proposal for its legal implications.314 By 26 May 1944, al-Maʿārif provided the Monotype Corporation with samples of two existing founts: the type used by its own printers and the one of the Government Press (Figure 2.70). Lindell shipped a box of the latter type to the 311

The Monotype Corporation Ltd, “Backmatter,” Monotype Newsletter no. 39 (May 1939). Lindell had offices in Cairo and Alexandria, representing companies selling machinery and equipment for printing and textile industries. Describing his travels to North Africa in 1928, Reginald Orcutt noted: “The Société Linotype Française of Algiers was a branch of a parent company in Paris, that was daughter of the British Company and grandchild of our head office in Brooklyn.” Orcutt, Merchant of Alphabets, 83.

312

Dār al-Maʿārif, (the house of knowledge). The battle for North Africa swung in favour of the Allied forces from late 1942, and was brought to a decisive victory in May 1943 when the last German forces were defeated in Tunisia. Encyclopædia Britannica Online, “World War II,” https://www.britannica.com/event/World-War-II (accessed 16 May 2017).

313

Quixley to Buckle, 25 January 1944.

314

Monotype had already obtained samples of this face in 1939 from the Hyderabad Government Press. With reference to the findings of lawyers in Egypt, precedents of similar cases and research of patent records, Lindell concluded that “the letters used by the Egyptian Government Press are not subject to patent”. Ibid., 5.

Mechanical Composition of Arabic

Figure 2.70

157

Specimen of the Egyptian Government Press fount in 15 pt; note the ­composition errors in lines one and six of the second paragraph: the word ‫ �ب�ن‬bin (son) is ­rotated by 180°. Reduced to 85% linear. MT, correspondence folder Arabic (Egyptian I).

158

Chapter 2

United Kingdom and noted that the client was prepared to accept a Monotype version of this design. Alongside the type, al-Maʿārif also provided the desired fount synopsis, specifying the priorities for inclusion of characters in the matrix case.315 It appears that no immediate development resulted from this enquiry, and when Lindell wrote again to the Corporation in April 1945, he made a strong case to proceed with the development. Urging the need for a new Arabic fount, he emphasised that the Egyptian Government Press “could not legally object” against Monotype’s use of their typeface as model and added that “all available faces in the M.E. [Middle East] are copied from others’ designs”.316 Moreover, Lindell considered the necessary adaptation to the unit system sufficient to differentiate the Monotype version from its model: […] even if you decide to cut the G.P. face, this will never be as true as their actual face, which is for hand composition. In order to bring the G.P. face to unit value some differences will be unavoidable.317 It appears that Lindell’s suggestion was followed by the Corporation as records show that experiments with both founts were conducted from May 1945. In preliminary tests characters were reduced in the projection room to 18D and used to develop photographic trials. Eventually, the Egyptian Government Press face was chosen as the model for the new Monotype fount and a few trial characters were manufactured.318 A first proof of this new face was printed in January 1946 under the production name Arabic Standard, Series 409 ­(Figure 2.71). Its Trial No. 1 features a selection of 29 alphabetic characters in 16 pt, set 315

Joseph Lindell to Percy Goodall, Letter, 26 May 1944, correspondence folder Arabic (Egyptian I), MT.

316

Joseph Lindell to The Monotype Corp. Ltd, Letter, “Re: Arabic - Your Letter X/ES.160 Dated February 1, 1945,” 17 April 1945, correspondence folder Arabic (Egyptian I), MT.

317

Ibid. This echoes Jan van Krimpen, who argues that Monotype adaptations of foundry type can only be seen as distinct designs, bearing insufficient resemblance to the original to be called by the same name. van Krimpen, “On Preparing Designs for Monotype Faces.” Lindell further mentioned that his business had been limited to English language composition, for the only Arabic equipment sales went to Europeans producing bilingual publications. This commercial inconvenience – a sizable part of the business of representatives derived from commissions – may indicate that Series 507 was used only by customers with a less developed sensibility for Arabic typography.

318

The Monotype Corporation Ltd, “Arabic,” 31 May 1945, correspondence folder Arabic (Egyptian I), MT. The rationale for this choice is not documented.

Mechanical Composition of Arabic

Figure 2.71

159

“Trial No. 1, Series 409, 16-pt,” January 1946, actual size. MT, folder Series 409.

i­ ndividually as well as in sample words. The fount appears largely satisfactory, but exhibits also a few unsuccessful characters whose proportions may have been overly compromised by the unit system (see characters 93, 177, 312). On submission to the client for assessment in March of the same year, Series 409 was rejected by Al-Maʿārif, “mainly on account of its size”, according to an internal memorandum.319 In reaction to this proof the client provided additional 319

The Monotype Corporation Ltd, “New Arabic Composition Matrix,” 12 March 1946, correspondence folder Arabic (Egyptian I), MT. Additional information about this rejection is not documented and it remains unclear whether the fount was considered too big, too

160

Chapter 2

samples of the two sizes that were required, leading Monotype to cancel the development of the fount. By April 1946, Series 409 was superseded by Series 549.  onotype Series 549, the Unwitting Mechanisation of M Qurʾān Typography Monotype’s staff and its representative Lindell were aware of the source of the model for their new Arabic typeface Series 549, but no evidence suggests that anyone involved with the type’s adaptation knew about its previous historic role. Commonly called the Government Press face, or the Būlāq type, it has a prominent, though not exhaustively documented history.320 Developed in the beginning of the twentieth century at the Būlāq printing house in Cairo, the typeface was used for the press’s own publications.321 Following closely the Nasḫ hand, the typeface was one of the first that successfully reproduced this established manuscript style in type. It employed approximately 470 sorts, including many ligatures and alternate forms to achieve a degree of elegance and authenticity not found in most earlier Arabic foundry type (see Figure 2.70 on page 157).322 The typeface attained its historic status when in 1924 it was used to print one of the first typographically composed Qurʾāns that was endorsed by a Muslim authority and widely distributed (Figure 2.72).323 Hitherto the theologically approved norm for the copying of the Qurʾān was to handwrite the text and then reproduce it by lithography; the departure from this orthodoxy must be small or otherwise disproportional. 320

See “Arabic Publishing at the Turn of the Twentieth Century” on page 26.

321

According to Hammam, the type was ”chosen by a High Committee composed of distinguished men of letters and technical experts on Arabic calligraphy” and the matrices were engraved in 1906. Hammam, “History of Printing in Egypt,” 158.

322

A notable exception is Ohannis Mühendisoğlu’s type in the Yeni Hurufat. See Milo, “Arabic Script and Typography,” 122.

323

Earlier typographical Qurʾāns made by European scholars and printers existed, but lacked authority among Muslims. According to Roper a small number of Qurʾāns began to be printed in Egypt in the last decades of the nineteenth century, but remained of marginal impact. See Geoffrey Roper, “The History of the Book in the Muslim World,” in The Oxford Companion to the Book, ed. by Michael F. Suarez and H. R. Woudhuysen, vol. 1, Oxford Reference Online (Oxford University Press, 2010), http://www.oxfordreference.com/view/10.1093/acref/9780198606536.001.0001/acref-9780198606536-e3986?rskey=PCys83&result=7&q=qur%27an (accessed 16 May 2017).

Mechanical Composition of Arabic

Figure 2.72

161

Page from the King Fūād Qurʾān; in the box to the right a detail from line six illustrates how elongations were employed to accomodate the floating marks, enlarged to 400% linear. Al-Qurʾān al-Karīm (Cairo: al-Azhar Madrasa, 1924), 14, actual size. Courtesy of Thomas Milo.

counted among the most profound events of Arabic typographical history. The Cairo edition, also named after its patron King Fuʾād, was ground breaking on numerous counts. In addition to the use of foundry type for the text, it entailed

162

Chapter 2

a complete recension for the needs of traditional Islamic Qurʾān scholarship.324 The recension was conducted by religious scholars of the highest instance of Sunni Islam, the Azhar Madrasa, therefore making the King Fuʾād Qurʾān the authoritative version of the Muslim holy book, a position it retains to the current day; indeed, the importance of the concomitant recension and the use of typography to give shape to the Qurʾān cannot be overemphasised. In addition to being one of the first instances in which foreign technology was accepted for this most guarded of texts, the Cairo edition normalised spelling variations of earlier Qurʾāns and provided recitation aids combined with an orthography that was based on the conventions established under Caliph ʿUṯmān.325 This new recension resulted in unprecedented spellings, as shown in a recent study by Milo.326 Drawing from comparisons with multiple editions, Milo shows how the unparalleled orthographic precision of the Cairo spellings led to new – and in writing unknown – renderings of words. Accomplished with foundry type, the appearance of these spellings was shaped by the inherent characteristics of the medium. In manuscript practice, vowels are positioned according to the height and depth of each letter, but here the physical properties of metal type did not allow such flexibility. Instead, vowels aligned horizontally above and below the characters in lines of separately cast sorts (Figure 2.73). Also, elongations between letters, commonly governed by strict rules, were introduced in previously unknown positions in order to carry the large number of marks

Figure 2.73

Diagram of the nominal configuration of sorts for vocalised text setting, detail from line three, enlarged to 200% linear.

324

Hartmut Bobzin, “From Venice to Cairo: On the History of Arabic Editions of the Koran,” in Middle Eastern languages and the print revolution, (Westhofen: WVA-Verlag Skulima, 2002), 170.

325

“It was printed orthographically in such a way that the shape of the original 7th-century consonantal text is clearly presented, but at the same time it included all the vocalization and other indications which are required for understanding the meaning and for correct recitation”. Roper, “The History of the Book in the Muslim World.”

326

Thomas Milo, “Arabic Amphibious Characters. Phonetics, Phonology, Orthography, Calligraphy and Typography,” in Vom Koran zum Islam, ed. by Markus Groß and Ohlig (Berlin & Tübingen: Verlag Hans Schiler, 2009).

Mechanical Composition of Arabic

163

that could not otherwise be accommodated. In conclusion, it follows that the simultaneous introduction of a new, authoritative edition of the text, and its composition with foundry type resulted in the establishment of a previously unknown image of the Qurʾān; incidentally, this image was in effect typographical, and thus set a new, technology-influenced standard for an acceptable rendering of the Arabic script.327 ∵ Based on a typeface of such standing, Series 549 held great prospects for success for Monotype. In its first printed trial, dated 8 April 1946, 11 characters of the new face are featured in 16D and 18D sizes (Figure 2.74 overleaf).328 Although only a small sample, the new proof shows considerable improvements over the rejected Series 409. The design of the previously unsuccessful characters ‫ ك‬kāf, ‫ و‬wāw and ‫ ر‬rāʾ adopts more conventional proportions and follows more closely the style of the model. Overall, the new fount exhibits features that are found in the foundry type source: the shapes of individual characters are dynamic and combine well with one another, contributing to a pleasant movement in reading direction. Its style is inspired by manuscript Nasḫ and echoes its elegance and rhythm in a normalised typographic form. These qualities are maintained and further amplified in the second trial of the fount of 26 June (Figure 2.75 overleaf). Limited to one size only (16D), the proof contains a larger selection of characters, set in two lines of words and word-fragments. Besides the consistently successful addition of characters, this new proof exhibits the full impact that Monotype’s kerning ability had on the typographic ‫أ‬ rendering of Arabic. In the sequence ‫� ر‬, the wide tail of ‫ ر‬rāʾ projects deeply ‫ع‬ into the open counter of ʿayn, and thus contributes to the ideal spacing of ‫ع‬ the word-shape. Early in the development of Series 549 the inclusion of vowels was discussed as a useful addition to the fount conspectus, but its implementation posed additional problems for the engineers at Monotype:

327

Numerous later editions are lithographic copies of the original movable-type edition. See for example ‬)Cambridge: Ǧamʿiyyä an-Nuṣūṣ al-Islāmiyyä, 1983( ‬Al-Qurʾān al-Karīm‫‫‫‬.

328

The two sizes appear to be developed from a single set of drawings and patterns, a common practice in the manufacture of matrix founts for hot-metal composition.

164

Chapter 2

Figure 2.74

“Trial No. 1, Series No. 549, 16D and 18D,” April 1946, actual size. MT, folder master proofs.

The question of composing Arabic, with all accents, in one operation, is difficult, if not an impossible operation. It is, therefore, suggested that a compromise can be reached by using a limited number of accents in the same die-case [matrix case], eliminating certain characters.329

329

Joseph Lindell to The Monotype Corp. Ltd, Letter, “Re: Accents for Arabic Series 549.,” 17 January 1947, correspondence folder Arabic (Egyptian I), MT.

Mechanical Composition of Arabic

Figure 2.75

165

“Trial No. 2, ­Series No. 549, 16D,” June 1946, actual size. MT, folder master proofs.

A selection of the most frequently used accents was thus chosen and designed with the goal of making them usable in combination with both the 16D and the 18D founts. The newly conceived characters were to be cast on a 6D body that should allow their use with both sizes, each cast on smaller bodies (10D and 12D respectively). Yet, their inclusion within the standard die-case proved too complicated on grounds of the varying sizes and the required overhangs of vowels. Eventually it was decided that Series 549 would provide accents only as outside characters.330 Without any formal announcement about the new typeface, it appears that by the summer of 1947 the 16D size was manufactured and delivered to clients in Egypt.331 In a letter to the Typographical Committee it was reported that Series 549 was “a great success” and that “experts in Cairo said it was the best 330

Type Drawing Office, “Series 549,” card index, 2, minutes box, MT.

331

Later Monotype promotional material gives the 1948 as official publication date.

166

Chapter 2

Arabic face in existence”.332 Despite this excellent feedback, a single size limited its usability and Lindell called for the manufacture of additional founts (10D and 12D) for magazine and railway timetable work. In the meantime work on the 18D and accents continued for another year, and by August 1948 the expansion of the Series was given an urgent incentive. The government of Egypt was planning its 16th Agricultural and Industrial Exhibition in February 1949. Monotype’s representative Lindell realised the potential for the Corporation to showcase its new Arabic typeface in this venue, and reiterated the need for additional sizes.333 In preparation for the fair, work on three sizes (12D, 16D, 18D) was put in hand, yet only the largest could be completed in time. In a leaflet designed for the event, Series 549 was first publicly announced at the Cairo fair (Figure 2.76). The English text emphasised the novelty of the typeface in achieving “the reproduction of the Arabic Script in all its beauty, grace and harmony in a degree never before attained by mechanical means”.334 It further acknowledged the long development time and the contributions of organizations from the region, without whom the development of the typeface would not have been possible. On a separate sheet, the division of keyboard and caster is underlined as a feature which enables global cooperation: The Arabic text of this leaflet was keyboarded in Cairo. The spool of perforated paper was sent to us by Air Mail and the type cast on a ‘Monotype’ Caster in England.335 Although a promotional text, the claims in the specimen did not exaggerate the quality of Series 549 (Figure 2.77). Owing to the possibility of kerning, word shapes are compact and support the readability of the face, an unassailable advantage over the Linotype that was recognised and advertised: “because there are no mechanical limitations preventing the kerning of characters, the beauty of the ‘flowing script’ can be preserved in the mechanical setting of ‘Monotype’ Arabic”.336 Indeed, the decisive contribution of Monotype’s en332

Quixley to Elfried Silcock, Letter, “Arabic 549-12 D and 10 D.,” 13 August 1947, correspondence folder Arabic (Egyptian I), MT. No printed samples of this earliest complete fount of the Series could be found.

333

Percy Goodall to Paulson, Letter, “Cairo Exhibition,” 16 August 1948, correspondence folder Arabic (Egyptian I), MT.

334

“A Specimen of ‘Monotype’ Arabic” (The Monotype Corporation, 1949), 1, folder Arabic, MT.

335

“A Specimen of ‘Monotype’ Arabic.”

336

The Monotype Corporation Ltd, “Arabic Composed on ‘Monotype’ Machines,” n.d., 1,

Mechanical Composition of Arabic

167

gi-

Figure 2.76

“A Specimen of ‘Monotype’ Arabic”, February 1949, reduced to 45% linear. (a) cover page (b) loose inlay. MT, folder Arabic.

neering ­excellence to the type’s success cannot be overstated, and its role was implicitly acknowledged in a paragraph with unusual technical detail: Apart from preserving the natural appearance of the Arabic Script the possibility of casting characters overhanging the body width, allows full use to be made of a minimum number of matrices as various accents can be used when needed next to an overhanging character. The overhanging character can rest on the body of the accent for support, thus allowing an accent to be used either above or below the character. When an overhanging character is to be used without an accent a high space (similar in width to the body of the accent) can be used to support the overhang.337 folder Arabic (early correspondence), MT. See also “Annex 1” on page 475. According to a typescript that may have been a draft for the official specimen text, 90 of the 308 characters in the fount were kerned. The Monotype Corporation Ltd, “Untitled Typescript,” n.d., folder Arabic (early correspondence), MT. See also “Annex 2” on page 478. 337

The Monotype Corporation Ltd, “Arabic Composed on ‘Monotype’ Machines,” 1. The text

168

Chapter 2

Here one should bear in mind that besides the technical flexibility of the Monotype system and its successful use for the Arabic script, the design of the fount was shaped in equal measure by the quality of its model – the Egyptian Government Press face. Monotype’s Series 549 faithfully reproduced the character of the foundry type, yet even improved on it in terms of consistency and manufacture. Its character shapes are well proportioned and their design ‫ف‬ avoids all ambiguities between similar letters (see medial �� fāʾ,  ʿayn and ‫ع‬ ‫م‬ mīm), a characteristic which is further amplified through the inclusion of a large number of commonly used ligatures. A minor shortcoming that was cop‫ف‬ ied from the source is found in the small counters of letters such as ‫ و‬wāw, �� fāʾ and mīm, which tend to visually close-up and create unnecessarily dark ‫م‬ patches on the page. The manufacturing quality of the fount is significant for the excellent overall impression as the alignment of characters is faultless and no gaps are discernible within letter groups when printed. Minor remaining issues in Series 549 can be found in a number of marks that ‫ ف‬are poorly ‫ن‬ positioned, and in certain combinations (see for example �‫ ��خورو‬in line six) the kerning of the isolated ‫ و‬wāw appears exaggerated. Overall, however, the typeface marks a milestone in the development of Arabic typography. On both macro and micro levels, the typeface excelled above all previous Arabic founts for mechanical composition and the majority of earlier foundry type founts. Series 549 successfully combined the readability and aesthetics of a typeface deemed suitable for printing the holy Qurʾān with the advantages and convenience of mechanical composition. According to Monotype’s promotional material, composition of between 110 and 140 characters per minute was possible with the new fount – six times faster than type composed by hand.338 At the occasion of the Cairo exhibition, Monotype’s newly appointed managing director Elfrid Silcock travelled to Egypt and oversaw the release of the new 18D size of the typeface. In the following report he noted the positive reception of the new design and the further need for additional sizes – a ­demand that was met with a range of founts developed over the coming years.339 In a hint at future developments, Silcock’s report also raised the need for fully vocalised Arabic for use in educational literature.340 This hitherto impossible does not specify whether diacritical dots or vowel marks are meant by the term “accent”. 338

Ibid., 2.

339

In its hot-metal form the typeface was eventually available in 10D, 12D, 14D, 16D and 18D.

340 Elfried Silcock, Report on Visit to Egypt of Mr Silcock from February 12th to 17th 1949 (Salfords, Surrey, 7 March 1949), 3, folder Arabic (early correspondence), MT, Salfords, Surrey.

Mechanical Composition of Arabic

Figure 2.77

169

Arabic text shown in the specimen, p. 3, actual size. MT, folder Arabic.

feature became the driving force for the logical development of Series 549: a fount for mechanical composition with full vocalisation capabilities.

170

Figure 2.78

Chapter 2

An iǧāzä, the Islamic certificate of competence in calligraphy. Penned in Ṯuluṯ and Nasḫ styles this iǧāzä illustrates the flexibility of vowel position and size in the script’s manuscript form. ʿAli Ra‌ʾif Efendi, 1791, 28 × 21 cm. 1-88-154.129, Selections of Arabic, Persian, and Ottoman Calligraphy, African and Middle Eastern Division, Library of Congress, Washington, D.C.

Monotype Series 589, Full Vocalisation The use of floating marks in Arabic type caused considerable difficulty before type lost its physical properties with the introduction of photocomposition. In Arabic writing the short vowels and other supplementary marks are placed in a second pass once the rasm and the diacritical dots have been written, thus allowing for a highly contextual configuration. The position and, in the case of some vowels, also the size of the mark may react to the surrounding elements (Figure 2.78). The characteristics of foundry type made the reproduction of this flexibility practically difficult. Vowels and marks could either be cast on the same sort as the letter they were intended to modify – a highly inefficient approach as it entailed an exponential increase of sorts that had to be included in a fount – or the casting of separate vowel characters for combination with other sorts.

Mechanical Composition of Arabic

171

While the latter was more economical for fount-manufacture, the handling of the minuscule vowel sorts dramatically increased the complexity of composition, making it error-prone and time consuming. In addition, vowels cast on individual sorts lacked solidity because of their small size and were thus vulnerable to breaking or bending under the pressure of the printing press. Most foundry type employed a combination of both strategies, casting the most frequent‫ أ‬and semantically pertinent combinations on single sorts (for example the � alif hamzä), whilst providing individual characters for discretionary marks that could be added where required.341 Although functional in principle, this system imposed its own limitations on the rendering of Arabic in print. Because the height of the body is generally consistent for one size of foundry type – irrespective of the character a sort carries – a mark could not always be placed in a position that reflected the size of the letterform (see Figure 1.6 on page 24). Kerning of marks, and in the most exacting composition the use of varying body heights, could alleviate the problem, but not fully resolve it (Figure 2.79). The first founts designed for mechanical composition on linecasters were yet more restricted by the constraints of the machines. Matrices on linecasters occupied a large vertical space that prevented the composition of separate vowels above or below characters (Figure 2.80 overleaf). On the Linotype it was not possible to cast separate lines of marks, for the operator would have to remember the position of vowels relative to the character in the preceding line, manually adjusting for variations of width without visual guides. The only remedy that was presented in specimens of Linotype founts was detrimental to the quality of the composed matter. Every vowel was provided as two separate matrices, one cast with a connecting stroke at the baseline height, one without; (Figure 2.81 overleaf) the former was to be used with join-

Figure 2.79

341

Detail of a page from The M.E.C.A.S. Grammar of Modern Literary Arabic, (Beirut: Khayats, 1965), enlarged to 140% linear. This publication is exemplary for quality Arabic composition, yet the properties of foundry type remain visible in the excessively high position of the vowels in the first words of each line.

Arabic is generally comprehensible without full vocalisation, but in cases of ambiguity vowels may be employed for clarification in otherwise un-vocalised text.

172

Chapter 2

Figure 2.80

Justified row of Arabic Linotype matrices. From Al-munaḍḍadä al-ʿarabiyyä, Mergenthaler Linotype, 7.

Figure 2.81

Specimen of vowel matrices set of the Linotype 24 pt Arabic 26◊ 544 fount, ­November 1955, reduced to 90% linear. DTGC, box Arabic general.

Figure 2.82

Detail of specimen of vocalised Linotype composition, February 1959, reduced to 75% linear. DTGC, box Arabic general.

ing ­characters, the latter was to be composed with initial, final and isolated characters.342 In showings of vocalised text by the manufacturer, vowels thus trailed or preceded the characters they were meant to modify (Figure 2.82). Although decipherable, this scheme severely impedes readability as vowels float in the line without any clear visual relation to the characters which are meant 342

The same principle was employed in the simplification system of Ahmed Lakhdar ­Ghazal, see page 90 and following.

Mechanical Composition of Arabic

173

to ­carry them, requiring the eye to travel repeatedly back and forth to connect the disparate elements. The inappropriateness of the system can be demonstrated by the workaround that was employed by a Linotype client in Bahrain. In the production of The Accented Arabic elementary school books the printer had to resort to a cumbersome process (Figure 2.83 overleaf). He first typeset the text without vowels in a standard Linotype fount (18D Traditional Arabic 2 with 3), and printed it with letterpress. After this step, the editor of the series manually wrote the vowels in the correct positions, a lithographic film was made and the final books were eventually printed in offset.343 Considering the effort in time and the cost the printer incurred in this process, it becomes clear that the results obtained from the linecaster were unacceptable for such work. Consequently Linotype had limited, if any, access to markets that required vocalised typesetting. In view of these difficulties, which effectively prevented the composition of vocalised text on linecasters, Monotype had the opportunity to reply to a demand that the competition was unable to satisfy. As mentioned above, an integrated solution for the composition of fully vocalised Arabic had been considered during the development of Series 549, but was side-lined due to the added complexity. However, the matter received a new stimulus in 1952 when the director of Al-Maʿārif, Chafik Mitry, informed Monotype that a competitor had offered to manufacture accented matrices according to their requirements.344 In a follow-up letter explaining the motivation of the client, Lindell noted that al-Maʿārif had reduced its hand compositors by almost 75% since the introduction of Monotype equipment, but that increasing demand for educational books made the mechanical composition of vocalised text a pressing concern.345 Monotype, eager to prevent a competitor from entering the field, thus renewed its commitment to the development of fully vocalised Arabic and began to investigate a proposal articulated by Mitry:

343

Michael A. Nahas to Walter Tracy, Letter, 10 February 1968, WT correspondence, folder 18b Dib-Aoun, DTGC.

344

Chafik Mitry to The Monotype Corp. Ltd, Letter, 21 July 1952, folder Arabic (early corres­ pondence), MT. Incidentally, two days after this letter was sent, the coup d’état of the Free Officer Movement toppled the monarchy. The new socialist-inspired regime brought far-reaching changes of politics that came to define Egypt’s history for the next 50 years. Encyclopædia Britannica Online, “Egypt,” n.d., https://www.britannica.com/ place/Egypt (accessed 16 May 2017).

345

Joseph Lindell to The Monotype Corp. Ltd, Letter, “Re: Al-Maaref Printing & Publishing House, Cairo,” 29 July 1952, folder Arabic (early correspondence), MT.

174

Figure 2.83

Chapter 2

Page of The Accented Arabic elementary school books, Bahrain, 1968, p. 69, reduced to 70% linear. DTGC, box Arabic general.

Mechanical Composition of Arabic

175

[Mitry] claims that he has been partially successful in his trials, by casting some of the characters three or four units less than the unit width for which these characters have been designed, thus allowing a space by the side of the character for the accent to be placed.346 In a preliminary report, the Works assessed the suggestion positively, with some reservations about its technical feasibility. Whilst the making of new matrices and a new MCA were acceptable, the report considered it doubtful whether some of the characters would be possible at all, “owing to the difficulty of casting very kerned letters”, and noted that some matrices might be impossible to produce because of “the reduction of bevels necessary to accommodate the accents”.347 In addition to the technical difficulty, the political turmoil in Egypt following the revolution in July was palpable in the publishing trade. Writing to Monotype’s UK office in March 1953, Lindell reported about the economic hardship as exports stalled and caused a substantial trade deficit; yet, his assessment of the printing trade suggested a positive influence of the new regime, for “with the compulsory education system more and more books are being printed” and with the country’s “newspaper output […] greater than ever […] prospects for new composing machines are definitely good”.348 Here Lindell cautioned that the competition from Intertype and Linotype was fierce, and that Monotype’s range of typefaces was insufficient to maintain its position in the market. Whilst urging the development of additional weights and sizes, Monotype’s representative unambiguously stated what he considered of highest strategic importance: “If we can offer the accents, we will have something which the Slug Machine [linecaster] can never offer and it will open a new field for us”.349 With his suggestions of development priorities, including missing fount sizes and bold faces for existing types, Lindell conveyed Mitry’s suggestion to send a qualified technician of al-Maʿārif’s staff to the United Kingdom. There 346

Anonymous, “PG/MKB/1239,” 4 September 1952, correspondence folder Arabic (Egyptian I), MT.

347

Works to Head Office, Letter, “Arabic ‘Naskh’. Series 549,” 29 September 1952, folder Arabic (early correspondence), MT.

348

Joseph Lindell to Elfried Silcock, Letter, “Re: Arabic Series,” 11 March 1953, 1, correspondence folder Arabic (Egyptian I), MT. Lindell reported a trade deficit of £ 70 million due to poor cotton prices; he also noted that “it was really hard for the new regime to tackle with [sic] the problems resulting from the situation which they had inherited from the bad management of the old regime”.

349

Ibid., 3.

176

Chapter 2

he could assist in the planning and manufacture of the fount for vocalised composition. Monotype’s managing director, Silcock, replied to the suggestion with general agreement, yet cautioned that prolonged development times were to be expected for the recommended programme because of previous “commitments in the supply of matrices for other countries which do not have the facilities of mechanical composition in their own language at present.”350 Numerous exchanges ensued between Lindell, who urged a fast expansion of the typeface range, and Monotype, which reiterated its manufacturing capacity limitations. Eventually it was agreed for the technical adviser of al-Maʿārif, Mr Karouth, to visit the Monotype Works and help with the fount development. A joint investigation by Karouth and the Works found that the method suggested by the client was feasible, provided that a new mould with a satisfactory operating mechanism was designed, and new founts, MCAs and keyboard equipment were made. The proposed rationale for the new system illustrates the important role of the technical aspects in the type’s conception: The position required for many accents necessitates a larger body size being used, but to obtain as many matrices as possible in the M.C.A. they must be kept to .2 × .2 where possible. The three bladed mould allows for these requirements. It is proposed to cast the equivalent of Series 54918D on 24D body with 4D total overhang top and bottom and 549-16D on 20D body with approximately 2½ total overhang top and bottom.351 Whereas a conventional mould opened only in two dimensions according to the point size and the set size, the three-bladed mould provided a way to vary the height of the body on which the face was cast, thereby allowing for the intricate system of overhangs and supporting spaces that was required for vocalisation. Further it was proposed to cast the type as an 18D face on a 24D body to provide sufficient vertical space for the accents above and below. Together with the technical improvements of the apparatus, design changes of characters completed the proposed system: In order to preserve the continual joining of characters when they overhang on to the accents, necessitates the design being widened [sic]. Other characters have to be re-made to obtain overhangs at specific sides. By 350

Elfried Silcock to Joseph Lindell, Letter, 14 May 1953, 1, correspondence folder Arabic (Egyptian I), MT.

351

Works (T.D.O.) to The Secretary, Typographical Committee, Letter, 29 March 1954, 1–2, folder Arabic (early correspondence), MT.

Mechanical Composition of Arabic

177

altering some characters we can obtain the use of the same keyboard equipment for both 549- 16D and 18D. Thus it is necessary to re-make approximately 90% of existing matrices and it would seem preferable to make the new characters as new founts for accented Arabic, leaving the existing series as the non-accented version.352 Based on this assessment, Monotype agreed to the new fount’s implementation and the collaboration between manufacturer, representative, and customer remained close throughout the following stages. Elements that were completed at Monotype were sent to Egypt for review and correction, and subsequently returned with suggestions for improvements. From July the fount was no longer referred to by the earlier denomination, but given a distinct Series number: 589.353 By October 1954 a new MCA was drafted and a first proof in 24D, Trial No. 1, was printed with the purpose to “show in line one the position of the accents relative to the characters and at the same time to prove the casting possibilities of these accents”. (Figure 2.84 overleaf)354 In line with Mitry’s wish for an aesthetically pleasing result, the accents were cast on matrices that were only three units wide, a practice that was not recommended by the Works because “the practicability of casting 3-units under normal conditions is very doubtful and not to be advocated”.355 But despite such reservations the opinion of the client was regularly sought by the manufacturer; indeed, it was decisive for the continuation: “We await the comments and criticisms from Mr. Karouth before being able to proceed further.”356 The importance of the collaboration between the various parties cannot be overemphasised, for the success of this development depended on their combined technical, aesthetic and linguistic skills. This fruitful partnership can be demonstrated by the informed design choices that are evident from the revisions made at al-Maʿārif. In January 1955 Lindell returned drawings to Monotype after Karouth had corrected and, in part, completely redesigned the characters they featured. The accompanying letter notes that all the revised drawings

352

Ibid.

353

Type Drawing Office, “589 Arabic Naskh Accented,” card index, 1, minutes box, MT.

354

Works (T.D.O.) to The Secretary, Typographical Committee, Letter, “Arabic Naskh Accented. Order No.E.328,” 8 October 1954, 1, correspondence folder Arabic (Egyptian I), MT.

355

Ibid., 2. Commonly the smallest width in Monotype matrix founts was 5 units.

356

Ibid.

178

Chapter 2

Figure 2.84

“Trial No. 1, Arabic Naskh Accented, Series No. 589, 24D,” October 1954, actual size. MT, folder master proofs.

contained observations about problematic issues, and illustrates the decisionmaking by one example: In the light of several years of use and experience in our 549, Mr Karouth has found it necessary to alter some of the characters in order to obtain the accents with perfect characters, for example: Letter ‫و‬ The opening has been enlarged because when printing on poor quality paper (absorbent) the opening of this letter is entirely covered by ink. Mr Karouth has opened the eye of this and some other letters to avoid this inconvenience.357 By October 1955, a new proof, Trial No. 2, was composed with the new threebladed mould, showing seven fully vocalised words and a number of individual characters (Figure 2.85). The quality of this specimen far exceeded that of vocalised typesetting on linecasters, and came close to the precision of the foundry type model. It employed a system that combined a selection of characters cast with accents on a single body, and a large number of marks that 357

Joseph Lindell to The Monotype Corp. Ltd, Letter, “Re: Accented Arabic,” 1 April 1955, 1, folder Arabic (early correspondence), MT.

Mechanical Composition of Arabic

Figure 2.85

179

“Trial No. 2, Arabic Naskh Accented, Series No. 589, 24D,” October 1955, actual size. MT, folder master proofs.

were cast separately. These individual matrices were provided in more than one vertical position and were made to work satisfactorily through the heavy use of kerns. Most marks were represented with two characters, one with a kern projecting to the left of the body, the other with an overhang on the right. They combined with the characters that carried marks, equally featuring numerous and often wide kerns, in an intricate system of interlocked sorts (Figure 2.86 overleaf). Because of the many overhangs the operator had to have a profound understanding of the interrelationship between matrices, for neighbouring characters had to support their respective kerns on their shoulders. For example some characters with a 3-unit overhang could not be placed adjacent to any other character, but required an accent or a high-space as support, effectively prescribing full, rather than partial vocalisation. Some shortcomings in position are due to the minimal ّ ُ width of the cast sorts, as can be َّ ‫ ف‬seen above narrow initial characters in ��‫ ت��ؤَدِ ب�ُ ن‬tuʾaddibunī and in particular in �‫ �ي‬fiyya, ‫ِي‬ ِ where the body of the diacritical dot prevents the vowel from being positioned further to the left (Figure 2.87). This proof established the feasibility of the concept, and hence the prospect of providing the previously impossible high-quality mechanical composition of fully vocalised Arabic. The relevance for the market and the economic inter-

180

Chapter 2

Figure 2.86 Illustration of the interlocking sorts of Monotype’s system for the composition of floating marks and vocalisation. From Gerhard Lieser, “Production of High Quality Arabic Texts on a CRT Filmsetting Machine”.

Figure 2.87 Detail of Figure 2.88, enlarged to 200% linear.

est of this proposition can be gathered from the reaction of Monotype’s representative C. Day: In my letter of 15th May from Bahrain, I stressed the point that the introduction of this face would greatly enhance our selling prospects in the Arab world, and I now go further and say that I have every reason to believe that orders could be secured, in the Beyrouth area, for three complete composing units before the end of this year, provided we can state that Accented Arabic will be ready for the market early in the new year.358 Despite the positive feedback, the typeface could not be completed in 1955 and it took a further five months to arrive at a complete and fully satisfactory fount.359 In March 1956 Trial No. 3 was printed as the final proof before 358

C. T. Day to Percy Goodall, Letter, 29 October 1955, 1, folder Arabic (early correspondence), MT.

359

Among the remaining work was the design of specialised keybutton banks that reflected the fount synopsis and facilitated the complex composition task: In a draft layout from November 1955, accents that kerned left were placed on the left bank, and those kerning to the right were located on the right bank. Percy Goodall to Joseph Lindell, Letter, “Accented Arabic - Series 589-24D,” 25 November 1955, 1, folder Arabic (early

Mechanical Composition of Arabic

Figure 2.88

181

“Trial No. 3, Arabic Naskh Accented, Series No. 589, 24D”, March 1956, actual size. MT, folder master proofs.

182

Chapter 2

public release (Figure 2.88 on page 181). In addition to the completion of the character set, improvements in the positioning of accents (see for example the accents above initial ‫ ل‬lām) indicate the continued revision and evolution of the system. Moreover, further tests were carried out on the machine to ascertain its satisfactory performance, which was finally deemed acceptable by the technical superintendent in April 1956.360 Based on Trial No. 3, Series 589 was approved for manufacture.361 At its inception in 1956 this typeface stood apart as the only option for mechanical composition of fully vocalised Arabic. But in addition to this unique selling point, the quality of the composition it enabled was unmatched at the time. The constraints that were imposed by the Monotype system – notably fewer characters than featured in Series 549 and the elongation of character shapes for the accommodation of marks – remained well within the scope of acceptable compromises for the typographic rendering of Arabic. Through the collaboration of an informed client with the engineers and draughtspeople at Monotype the quality of the fount was brought to an unprecedented level. Arguably Series 589 set a standard that remains competitive to the current day, and may be considered superior to many later typefaces that could have benefitted from improvements of technology. But whilst it constituted a major achievement in the evolution of Arabic typography and excelled at its specific task – fully vocalised text – the typeface remained a niche-product. Initially only available in 24D, too large for most text setting and without complementary weights to create typographical hierarchies, Series 589 had a narrow range of applications.362 Moreover, the technical complexity of its manufacture and use were defined by the characteristics of hot-metal composition, and as such the typeface’s system became obsolete with the next technological advance, due to unfold within only a few years.

­correspondence), MT. The final keyboard arrangement has not been found. 360 Works to Stanley Morison, Letter, “P/H.694,” 30 April 1956, correspondence folder Arabic (Egyptian 2), MT. 361

Type Drawing Office, “589 Arabic Naskh Accented,” 2.

362

A 20D fount was added in 1960 and a bold version of Series 549, Series 649 was issued in April 1959. E. A. Firmage to Joseph Lindell, Letter, 24 April 1959, folder Arabic (early correspondence), MT.

Mechanical Composition of Arabic

183

The Beginning of ­Computer-Aided Arabic Composition Throughout the twentieth century, typographical technology for Arabic lagged behind its Latin equivalent. Advances of equipment commonly found their way into practice years after the technology had been soundly established for the Latin script, often even towards the end of its life-cycle. For example photocomposition, widely used by the 1960s, only attained its full benefits for Arabic composition a decade later and, as Ross pointed out in 2002, even at the start of this century “digital composition has so far failed to redress the paucity of high quality non-Latin fonts”.363 Similarly, the 1960s saw the introduction of a technology to Arabic typography that had been pioneered much earlier. Teletypesetting was developed by Walter Morey in Newark, New Jersey, in the 1920s.364 Its principle was similar to Monotype’s system, in which a separation of keyboarding and composition allowed for the keying of text in a different location than the composing room. As the name suggests, teletypesetting (TTS) combined this principle with telecommunications technology. In TTS so-called ‘perforators’ with keyboards of great similarity to conventional typewriters produced coded paper-tapes that could then be translated into signals and transmitted to remote locations by means of telephone or telegraph lines (Figure 2.89 overleaf).365 Its origin and primary purpose was a rapid and direct link from news-gathering agencies such as Associated Press to the composing rooms of newspapers, as explained in an account by Walter Tracy: In the news agency the news ‘copy’ is typed out on a special keyboard, the output of which is a perforated tape, in which each character is represented by a unique code. When the tape is fed into a transmitting device, the codes are translated into impulses which can be received many miles away by a receiving device in the office of the subscribing newspaper. This receiver translates the impulses into perforations on a tape, like the original, and this tape can be fed into an electronic device which automatically activates the keys of a Linotype machine.366 (Figure 2.90) 363

Fiona Ross, “Non-Latin Typesetting in the Digital Age,” in Computers and Typography 2 (Bristol: Intellect Books, 2002), 52.

364

Seybold, Fundamentals of Modern Composition, 60.

365

The simplicity of the keyboard and its similarity to the typewriter allowed newspapers to employ staff with lesser qualifications than fully trained linecaster operators who required a formal four to six year apprenticeship. According to Seybold, this practice was only followed in case of acute labour shortages. Ibid., 64.

366 Tracy, “Advances in Arabic Printing,” 88–89.

184

Chapter 2

Figure 2.89

Left: Friden Flexowriter 2300 series 8-level punched tape typewriter. From ­Phillips, Computer Peripherals and Typesetting.

Figure 2.90

Right: Fairchild Teletypesetter operating unit for tape control of linecasters. From Phillips, Computer Peripherals and Typesetting.

This system had a number of advantages. Primarily, newspapers could economise time as the re-keying from hard-copy could be eliminated. Moreover, a single tape could be used to control multiple linecasters, increasing output and reducing the number of operators required, thereby saving the newspaper money.367 And yet, TTS was adopted late by news agencies and newspapers. Only in 1951 Associated Press and United Press International began the transmission of remotely typed stories.368 Among the obstacles for the widespread use of TTS was a lack of typographical uniformity on the receiving side. ­Column widths, typefaces and justification varied between newspapers and rendered the use of a single master tape impractical. Hitherto, linecaster founts were not based on a unit counting system, for the wedge-shaped space bands allowed for variable justification of lines. Hence, and contrary to the Monotype system, the matrices of linecaster founts had arbitrary widths. The 367

The possibility of controlling casters with tape during night-shifts furthered the savings on wages. However, in the United Kingdom union agreements limited this practice to “one operator to supervise two tape-controlled Linotypes” (Linotype & Machinery Ltd, “TTS for Book Composition Offers More Than Double the Speed Without Any Loss of Quality,” Linotype Matrix II, no. 29 (June 1958): 3.

368

Seybold, Fundamentals of Modern Composition, 60. The same year the London Times newspaper established a TTS system within the houses of parliament, enabling it to print an “up-to-the-minute report of Parliamentary proceedings in successive editions” (Linotype & Machinery Ltd, “The House of Commons, The Times and Linotype,” Linotype Matrix II, no. 10 (Summer 1951): 1.

Mechanical Composition of Arabic

185

matrices were assembled in view of the operator who could, subject to his skill and ­experience, make line-ending decisions based on this visual information. But with a separation of keyboarding and casting, the remote keyboard operator had to have some indication of width to define the line ends. Therefore a counting mechanism was conceived, and TTS linecaster founts were made based on a set number of units, introducing a constraint long absent from their design. Linotype adopted an 18-unit system for its newspaper founts, allocating between 6 and 18 units per character.369 But in contrast to Monotype’s flexible system that assigned units depending on typeface and its design requirements, the TTS unitised founts assigned the exact same unit value to the same characters in every fount.370 Thus, a letter h of the Corona typeface had to fit within the same number of units as the letter h of Excelsior, necessitating substantial homogenisation of proportions.371 Once in use, the unitised founts allowed for the transmission of justified tape as by “selecting suitable unit founts it is possible to compensate for changes in line measure”.372 For the Arabic script TTS composition first became a possibility with the introduction of Simplified Arabic and its substantially reduced character set. The limited number of key-buttons of the perforator keyboard had hitherto prohibited its use for Arabic founts. A promotional publication of the first version of the typeface, the Series 2 with 3, emphasised the novelty of the ­technology: La commande par ruban TTS, avec tous les avantages de vitesse et d’économie, qu’elle comporte, entre enfin dans le domaine de la pratique pour les imprimeurs d’arabe.373 369

Arthur H. Phillips, Computer Peripherals and Typesetting: A Study of the Man-Machine Interface Incorporating a Survey of Computer Peripherals and Typographic Composing Equipment (London: Her Majesty’s Stationery Office, 1968), 387.

370

Walter Tracy’s account of the Monotype system testifies to its wide appreciation. He notes that the implementation of the unit system “had enough flexibility to reduce the limitations to insignificance so far as character design was concerned, as the high quality of the types in the English Monotype list demonstrates.” Tracy, Letters of Credit, 40.

371

In a discussion of the severe limitations of the Olympian design for the TTS unit system, Matthew Carter likens the difficulty of incorporating any degree of originality in the drawings to “getting the proverbial oyster through a keyhole” (“Olympian, A New Type for Newspapers,” Interpresgrafik no. 7 (1971): 46.

372

Phillips, Computer Peripherals and Typesetting, 387.

373

Control through TTS tape, with all its advantages in speed and efficiency, finally enters the domain of practice for Arabic printers, approximate translation by the author. “L’arabe simplifié, Mrowa-Linotype,” n.d., DTGC.

186

Chapter 2

There is no extant evidence of a unitised version of Simplified Arabic, its TTS composition might therefore have relied on a later development, the Multiface Perforator. This device contained a counting mechanism based on a 32-unit to the em system. It measured the brass widths of non-unit matrices and rounded them to the closest 1/32 of an em, thereby allowing the operator to make decisions about the position of the line-ending code. Rounding errors largely cancelled each other out, but could occasionally lead to lines being too tight or too loose to cast.374 Yet, this system remained unsatisfactory for the composition of Arabic text. The increase in speed was moderate because a typical TTS perforator keyboard featured only 63 keys and therefore relied on the frequent use of the shift and rail controls, slowing the output speed (Figure 2.91). Moreover, the operator had to recognise and choose between the various contextually dependent forms of a letter, further slowing down the composition. Likewise, the system was not equipped to adequately address justification of Arabic. Hyphenation is not possible in the Arabic script, and justification is not obtained through spacing-out of the line or letter-spacing, but through a transformation of lettershapes. In correctly justified text, word-spaces and inter-letter spaces always remain visually constant and filling of a line is achieved by means of the elongation of connecting strokes or final swashes. The employment of these elongation principles are governed by style-dependent rules.375 In most typographic renderings this aspect of the script is emulated with a dedicated typeform, commonly called Kashida, which is added between two joining characters.376 Yet, the fidelity of this justification technique to the elongation principles employed in manuscript production is at best approximate. Kashida characters were designed as straight lines to allow for sequential use. The resulting typographic rendering therefore created exaggerated horizontal lines, bearing little resemblance to the integral transformation of letter shapes known from the manuscript form (Figure 2.92 overleaf). Moreover, the application of this device in typographical composition imposed certain patterns. As compositors (or operators) could not always predict how many words a line of type would contain, justification by means of the Kashida tended to be employed towards the end of a line, resulting in an imbalanced overall impression. In addition to these practical constraints, we know that type compositors and machine operators often lacked training and knowledge of the customary 374

Seybold, The World of Digital Typesetting, 63.

375

These rules are poorly documented outside the calligraphic domain and are often ignored in writing and current typographical practice.

376

In calligraphic practice the Arabic word employed to describe the principle is madd, denoting extension or lengthening; occasionally, taṭwīl is used synonymously.

Mechanical Composition of Arabic

Figure 2.91

187

“T.T.S perforator keyboard Mrowa Linotype Simplified Arabic”, 1959, reduced to 35% linear. DTGC, box Simplified Arabic.

d­ istribution of elongation – a situation that is no different today. The theoretical option of TTS control of Arabic linecasters with Simplified Arabic thus offered limited benefits compared to regular operation. The Development of Character Selection and Kashida Insertion Routines Under the rule of Gamal Abdel Nasser, al-Ahram had become Egypt’s semiofficial newspaper, closely aligned with the government’s policies.377 In 1957 Nasser appointed his personal friend, Mohamed Hassanein Heikal, as the paper’s editor-in-chief, bestowing unprecedented political significance and influence on the newspaper.378 But far from turning al-Ahram into a propaganda organ, Heikal’s tenure made it an internationally respected newspaper with greatly improved accuracy and objectivity of reporting. During this period it was called “The New York Times of the Arab world” and thus attained a rolemodel status for the press in the Middle East.379 In the mid-1960s a grand new building was erected for al-Ahram. It was located in the poor Būlāq district, and while it was widely criticised for its indulgence, it also represented the

377 Ǧamāl ʿAbd An-Naṣr. 378 Muḥammad Ḥassanayn Haykal. 379

Encyclopædia Britannica Online, “Muhammad Hassanein Heikal,” https://www.britannica.com/biography/Muhammad-Hassanein-Heikal (accessed 16 May 2017).

188

Chapter 2

Figure 2.92

Detail of al-Ahram, 18 March 1966, p. 3, actual size. The right column, set in Intertype Abridged Arabic bold, is a typical example of justification with Kashida extension strokes. Note how the hand-lettered headlines copy the horizontal elongation of typographical composition, an apparent influence of technology on lettering practice. Courtesy of the IZM Aachen.

aspirations of the newspaper.380 Heikal was known for his insistence on high production quality and, apparently in search of the best machinery for the new premises, al-Ahram approached Linotype in 1965 and indicated its interest in 380

Edward R. F. Sheehan, “The Most Powerful Journalist in the World, His Newspaper and His Book,” in The Cairo documents: the inside story of Nasser and his relationship with world leaders, rebels, and statesmen (New York: Doubleday, 1973), xxiv.

Mechanical Composition of Arabic

189

Figure 2.93 Linotype Elektron linecaster. The TTS tape input is seen at the very right. From Phillips, Computer Peripherals and Typesetting.

acquiring its most advanced Elektron linecasters (Figure 2.93).381 The press had been using Intertype machinery for decades, and the potential changeover became a matter of great importance for Linotype. Because of the change of supplier, few, if any of the exigencies and dependencies that resulted from upgrades of existing machinery applied to this request. It thus provided an opportunity for the re-evaluation of the newspaper’s setup, and consequently more freedom in conceiving new approaches to al-Ahram’s composition needs. These circumstances opened the door to advances in typesetting technology that had not yet found their way into Arabic typesetting. Here, another contributor has to be introduced. The Compugraphic Corporation, a recent newcomer to the printing trade, played a decisive role in the technological upheaval that the industry experienced in the 1960s. It was founded in 1960 by William Garth Jr., a former president of the Photon Corporation, and Ellis Hanson, a computer engineer, in Brookline, Massachusetts.382 381

Walter Tracy to D. S. Turner, Letter, “Simplified Arabic,” 29 October 1965, WT correspondence, folder 18a, DTGC. The Linotype Elektron was released in 1962 and publicised as the “fastest linecaster in the world [with] a speed of 15 newspaper lines per minute”. Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 32. It also marked the technological endpoint of linecasters as the last model that Linotype ever developed.

382

See William W. Garth IV, Entrepreneur. A Biography of William W. Garth, Jr., And the Early History of Photocomposition (Self-published, 2002).

190

Chapter 2

The company aimed to develop, build and promote typesetting computers to the graphic arts industry, before eventually producing its own phototypesetting devices.383 In 1963 Compugraphic released the Linasec, a device it had commissioned from the Massachusetts-based Wang Laboratories. The Linasec was a special purpose justifying computer, which meant that it had a relatively simple hard-wired justification logic: It required that unjustified text tapes be fed to the machine which accumulated the character widths and added end-of-line codes to an output TTS tape for the operation of linecasters and TTS-compatible phototypesetters. No hyphenation logic was provided, instead a falloff word was displayed electronically for the human operator to pick an apt break point.384 Thus, the idea that a computer could be used to facilitate tasks hitherto performed by the operator of a linecaster gained ground, and the prospect of automated Arabic justification became tangible.385 Linotype first approached Compugraphic in 1964 with an enquiry about its potential capabilities for Arabic composition and Ellis Hanson articulated the idea of a character selection computer, as well as “a logic system whereby the Kashida could be automatically inserted between characters as necessary to fill out a line”.386 Nothing came of this first idea, but here one has to bear in mind that the use of computers for typographical composition remained embryonic until the second half of the decade: It was not until 1965 that computer composition really developed a momentum and continuum with the appearance of the general-purpose and software programmable minicomputer …387 383

See Wallis, Typomania, 83.

384

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 33.

385

“[…] it is not surprising that during the 1960s it was realized that a computer could be made to perform this function, the computer being programmed to assess the total widths of the characters required to fill a line and where necessary to insert hyphens (or in the case of Arabic, kashida extension strokes) to achieve correct justification”. Tracy, “Advances in Arabic Printing,” 89.

386

Frank Bryant to W. P. Keegan, Letter, “Arabic,” 28 December 1966, WT correspondence, folder 18b, DTGC. There is no documentation of the reasons why the system was not implemented, but we know that the Linasec’s capabilities were considered insufficient for Arabic composition.

387

Wallis, Typomania, 83.

Mechanical Composition of Arabic

191

For Arabic composition, however, the first steps in computerisation were already taken with hard-wired machines when al-Ahram confirmed its planned change of equipment in September 1966. That same year Compugraphic had released a new special-purpose justifying computer that promised to lend itself to the specificities of Arabic composition. The Justape was significantly cheaper than the earlier Linasec, and it was designed to output justified paper-tape by modifying the word spaces rather than the breaking of words.388 (­ Figure 2.94) Its justification principle was therefore more suited to Arabic where hyphenation is not permitted, and an implementation could be attempted. Against this background, Hrant Gabeyan, Linotype’s recently hired representative in Egypt became a key figure in the invention of computer-aided Arabic composition. Born in Cairo in 1937 to two refugees of the Armenian genocide of 1915, Gabeyan was brought up in modest circumstances in postcolonial Egypt.389 He learned Arabic and French in a Christian school and was taught Armenian by his mother at home. Following secondary school and military service, Gabeyan was employed as a salesperson by the firm of Antoine

Figure 2.94 Left: The Justape, Compugraphic’s second special purpose justifying computer. Above: Control panel of the Justape, demonstrating the basic functionality of the first typesetting computers. From Phillips, Computer Peripherals and Typesetting. 388

The machine’s retail price was $8,200, roughly half the $15,000 to $18,000 one had to pay for the older Linasec. See Wallis, A Concise Chronology of Typesetting Developments 1886– 1986, 37. For the difference between the two devices see Phillips, Computer Peripherals and Typesetting, 413.

389

Hrant Gabeyan, interview by the author, Audio recording, 11 March 2011, DTGC. Substantial detail of Gabeyan’s life story is recorded in the interview, much of it going beyond the scope of this study. The complete record is stored in the DTGC.

192

Chapter 2

Figure 2.95

Hrant Gabeyan at his desk in the offices of Antoine Homsy’s firm, c. 1960. ­Courtesy of Hrant Gabeyan.

Homsy in 1960, receiving training in numerous aspects of the printing trade (Figure 2.95).390 Homsy was an influential Egyptian family business, active in various domains; it acted as authorised reseller for European companies, among them Linotype, and by 1962 Gabeyan became the company’s direct contact person in Egypt.391 With the increasingly far reaching sequestration policies of the Nasser government, the firm of Homsy was put under state control in 1964, severely limiting its scope of action. Gabeyan’s account illustrates the crippling impact of the presence of military personnel that substituted the managers of the firm, ultimately driving him to approach Linotype for employment opportunities. Consequently, in 1965 Gabeyan was hired to represent the company in Egypt and the Sudan.392 Early traces of Gabeyan’s work for Linotype reflect attention to detail, as well as genuine concern for the quality of Arabic type. His initiatives went 390 In a specimen of Mrowa-Linotype Simplified Arabic, the firm of A. & R. Homsy, 8 rue Chawarby Pacha, Cairo, is listed as the official Linotype representative. 391

In a letter to Michael Nahas, Linotype representative’s in Beirut, Tracy noted: “Mr. Proudian left Homsy’s several months ago. L&M business is being dealt with by Mr. Hrant Gabeyan, who also deals with the other graphic arts business equipment in the office.” Walter Tracy, “Arabic 10 with 11,” 13 March 1962, WT correspondence, folder 18b, DTGC.

392

Gabeyan recalls that the salary he was offered was roughly fifteen times his prior income with the firm of Homsy. Gabeyan, interview.

Mechanical Composition of Arabic

193

b­ eyond regular customer relations and sales duties, and comprised reviews of typefaces, quality checks of marketing material, and generally help with language and local customs. For example, he alerted Linotype to orthographical errors in specimens – as a salesperson Gabeyan was acutely aware of the embarrassment composition errors and spelling mistakes caused. But while concern about the presentation could be seen as part of a salesperson’s job, quality control and manufacture revisions, duties that he assumed casually, went beyond the portfolio of a company representative. Because Linotype lacked Arabic script and language expertise, Gabeyan effectively began to perform the dual role of salesperson and Arabic typographical consultant. Although nominally his status remained unchanged, over the following years Gabeyan left numerous marks on Linotype’s Arabic programme. His most significant contribution began when al-Ahram reaffirmed its interest in Linotype machinery in 1966. Whilst general ideas for computer-aided Arabic composition had been discussed between Hanson and Tracy, at this stage it seemed less than certain whether a ground breaking change of approach was conceivable. Indeed, Tracy appears to have been uncertain how Arabic justification was handled in practice and asked Gabeyan to gather the relevant information and brief a small team of Linotype employees:393 In thinking and talking about possible developments in the future, we have been wondering how TTS perforator keyboard operators allow for the inclusion in the line of the kashida extension stroke – which is needed to fill out a line when there are not enough spaces, as in narrow columns. You may already know the answer to this. But if you do not, I wonder if you would call in at al-Ahram and ask the perforator operators there what method they use. You will then be able to inform Dr. Lamberti and Mr. Emery during your forthcoming visit to Head Office.394 The general and vague nature of this enquiry illustrates the uncertainty in this early stage of the process. To Tracy it seemed neither clear how this system of justification was employed, nor whether automation was practicable. In order to establish the feasibility of the concept, a meeting between Edward 393

According to Gabeyan’s personal recollection he commenced research into the subject prior to this enquiry and out of his own initiative. Gabeyan, interview.

394

Walter Tracy to Hrant Gabeyan, Letter, “TTS Composition of Arabic,” 20 September 1966, WT correspondence, folder 18a, DTGC.

194

Chapter 2

Figure 2.96 Hrant Gabeyan in front of Linotype House, 21 John Street, London, January 1966. Courtesy of Hrant Gabeyan.

Emery, manager of the Graphic Systems department, Gabeyan, Sergio Lamberti, a programmer of Linotype S.P.A. Milan, and Tracy was called in October 1966 at the Linotype offices in 21 John Street, London (Figure 2.96).395 During this meeting and the ensuing discussions, the fundamental principles of computer-aided Arabic text input were laid. It was agreed that a computer should be employed for two tasks that had hitherto been performed by the operator: selection of the correct contextual form of each letter depending on the position within a word, and justification by means of Kashida extensions were now to be programmed and computer-controlled. As a result operators would have to depress one single key to access all contextual forms of the letter without any concern for justification of the line. The principle was as revolutionary as it was simple. As Linotype’s Arabic specialist, Gabeyan was tasked to articulate rules for the definition of the contextual substitution logic. Moreover, he 395

The timing of Gabeyan’s visit to the United Kingdom (from 2 to 19 October 1966) for this meeting suggests that his knowledge of the Arabic script was considered essential for the successful development of the system.

Mechanical Composition of Arabic

195

should gather and document the conventions governing the use of the Kashida. At the same time discussions with Compugraphic established that the Justape could be adapted to the specific requirements of the planned character-selection and justification system. In a letter reminding Gabeyan of the urgency of his task, Emery confirmed the prospect of using Compugraphic’s device: If you see Mr. Yassin [the director of al-Ahram] in the near future I think you could safely inform him that our preliminary investigation indicates that the “JusTape” can be modified to provide the Kashida requirements, and that we shall be able to quote for such a machine when we have received the logic specification on which you are presently carrying out research.396 In parallel to the conceptual work, Linotype drafted a patent specification in November 1966 to anticipate any claims by the competition.397 The rush for patent protection indicates the significance assigned to the invention, and the concern for its commercial potential, expressed by the patent manager Frank Bryant in a letter to Compugraphic: “The whole object of the exercise is to keep Intertype and other computer manufacturers out of the way, or to stop them stopping us.”398 By 9 December, Gabeyan had developed two tables containing the fundamental principles of character selection and Kashida insertion for Simplified Arabic (Figure 2.97 overleaf).399 They described in binary form the principles 396

Edward S. Emery to Hrant Gabeyan, Letter, “Re: Al Ahram, Cairo. (Egypt),” 24 October 1966, WT correspondence, folder 18, DTGC.

397

In a memorandum to colleagues involved in the project Bryant requests them to “treat this matter as being strictly confidential and not mention it to anyone until you are sure that a patent application has been filed.” Frank Bryant to Edward S. Emery, Letter, “Arabic Keyboard,” 7 November 1966, WT correspondence, folder 18a, DTGC. The fact that Gabeyan, despite his important role, was omitted from the recipients may indicate a culture that did not treat employees in foreign offices as equals to their British colleagues. Gabeyan only learnt of the existence of the patent from the present author during an interview in 2011, causing profound disappointment.

398

Bryant to Keegan, “Arabic.”

399

Gabeyan recalls: “When I unrolled my very long chart on the longest coffee table in the lounge of the Shepheard’s Hotel in Cairo, in front of Mr Walker, it was so long and colourful that he said: ‘I understand nothing of what is written down there, but I admire the work you have done, and therefore I offer you £50.’ These were the first and last words of praise coming from anyone at Linotype.” Hrant Gabeyan, email message to the author, “The Patent,” 9 September 2011.

196

Figure 2.97

Chapter 2

“‘Kashida’ Logic for Linotype Simplified Arabic,” (detail), 9 December 1966. The actual document measures approximately 185 × 115 cm. DTGC, locker C1.

of ­contextual substitution of Arabic characters and laid-out rules governing the correct placement of the Kashida extension between adjacent letters. The expression of these two principles of the Arabic script in a form that could be translated into software and thereby automate a significant element of the composing task was to form the basis of Arabic typography in the postmechanical era. Yet at this time Linotype was not in a position to write the required software, and Compugraphic had no expertise in Arabic.400 The 400 From a letter Lamberti sent to Compugraphic, it emerges that he could not have conceived the software himself: “I do not know how you will be designing this logic, I thought one could make a diode matrix and asked Hrant to draw the table with this in mind but now that I have seen it I think it will be very “heavy” to drive directly so it may result in being very expensive. You have far more experience in this type of circuitry than I have, so I will not try to suggest anything but I will be very interested to see your solution.” Sergio Lamberti to Ellis Hanson, Letter, 14 December 1966, WT correspondence, folder 18b, DTGC.

Mechanical Composition of Arabic

197

various factors of this system therefore required a collaborative effort, pooling the skills of a group of specialists – a dynamic that came to be typical for subsequent Arabic type developments. Here the data tables compiled by Gabeyan were sent to Compugraphic; and in December 1966 Tracy travelled to the United States to instruct Hanson in principles of Arabic typography.401 Hanson, an experienced programmer, was quick to translate the data sheets into a workable substitution logic, grouping characters into classes “to reduce the character selection logic to reasonable proportions”.402 At the same time that Hanson and Tracy worked on the implementation of the system, on 24 December 1966 Linotype applied for a patent describing the general idea of the computer logic.403 The filing for a patent before the system had been developed fully suggests the importance that was assigned to this project; given its collaborative nature, the unilateral decision to patent the system appears to have been controversial too.404 On 19 January 1967, al-Ahram’s official order was received by Linotype. It was a substantial transaction covering four Elektron casters, ten perforators, the Justape computer and maintenance tools. Indeed, the sale was of such importance that B. Saunders, manager of the machine order department, wrote a personal letter to Gabeyan, expressing his “great pleasure” in the name of the company.405 401

Tracy, “Composing Room Days (and After),” 13.

402

Ellis Hanson to Arthur Henry Walker, Letter, 22 December 1966, Private collection Hrant Gabeyan.

403

Lamberti, Sergio. Means For Controlling Typographic Composing Machines. UK Patent GB1162180, filed 24 December 1966, and issued 20 August 1969. The naming of Lamberti as the inventor seems to have been a choice of convenience by Linotype, rather than genuine crediting of authorship as he was neither involved in the assembly of the required information, nor the implementation in software.

404

Linotype justified their action citing concern for anticipation of a patent by Intertype. In a letter to Bill Garth, the president of Compugraphic, Walker assured him that “there is no intention of us to anticipate your filing in this matter and as soon as you are in a position to file and in fact have filed we will then investigate and find out whether Intertype have filed anything. If they haven’t then we will withdraw our provisional patent immediately.” Arthur Henry Walker to William (Jr.) Garth, Letter, 28 December 1966, WT correspondence, folder 18b, DTGC. The patent was not withdrawn and granted to Linotype in 1969. Compugraphic, in turn, filed its own patent for the system. See Hanson, Ellis. Control System for Typesetting Arabic. US Patent 3,513,968, filed 24 January 1967, and issued 26 May 1970. No conflict between the two parties over their respective patents was found.

405

B. A. Saunders to Hrant Gabeyan, Letter, “Al Ahram, Cairo,” 19 January 1967, WT correspondence, folder 18, DTGC.

Figure 2.98

“Arabic Perforator Keyboard Logic for Linotype Traditional Arabic,” 31 January 1967, 185 × 115 cm. DTGC, locker C1.

198 Chapter 2

Mechanical Composition of Arabic

199

In the meantime Gabeyan continued to gather and assemble data to improve the logic under development. By 31 January 1967 two additional elements of the system were completed. The character selection and Kashida placement rules for Traditional Arabic, as well as a detailed set of guidelines of stylistically preferred Kashida positions, were ready to be translated into software (Figure 2.98 and Figure 2.99). From the cover letter Gabeyan sent with the package of information, it appears that he sought to perfect the system out of his own initiative: As you know, the Kashida logic chart is general, and it shows whether a certain character accepts or not the Kashida. Actually, the chart is correct, but what comes hereunder, is a matter of artistic taste. The places where I am suggesting to insert Kashidas are the most preferable places, which can facilitate reading, rest the reader’s eyes, and give a better design to the article in general, and to the typeface specially. Unfortunately, these days there are few operators, not more than 15%, who know where to place the Kashida, and even if they know, they ignore it. Therefore it is a great success too if the Justape is able to insert the Kashida in the desirable place. As a matter of fact, immediately after I have finished the Kashida and Character Selection Logics for the Simplified Arabic, I have begun studying, what I call “the artistic insertion of the Kashida”, with calligraphers, operators, mechanics and Arabic language teachers, and I have been able to get the following results; and I did this because I felt that it was essential to guide the Justape, to give a character more importance than the other.406

Figure 2.99 406

Detail of Figure 2.98 showing Gabeyan’s signature and date.

Hrant Gabeyan to Walter Tracy, Letter, “The ‘Kashida’,” 31 January, 1967, Private collection Hrant Gabeyan.

200

Chapter 2

Figure 2.100 “‘One-Character’ Keyboard for Simplified Arabic,”, fourth version, March 1968, ­reduced to 40% linear. Note the ‫ ا ل‬al key, no. 25 in the bottom row. DTGC, box Simplified Arabic.

Moreover, Gabeyan was instrumental in the design of the corresponding keyboard layout (Figure 2.100). In collaboration with al-Ahram’s make-up editor Hišām Baḥarī, and as a result of Gabeyan’s work on the character selection routine, arose an idea that had ramifications beyond the commission by al-Ahram.407 َ‫ا‬Baḥarī had undertaken a survey of frequency counts of the definite article ‫ ل‬al in newspaper columns. Because Arabic grammar requires the definite article suffixed to all qualifying adjectives of a definite noun, it is among the most frequent letter combinations; according to Baḥarī’s study a typical slug of 10 Cicero contained from three to six definite articles.408 He therefore suggested devising a single matrix containing the full article. Placing both characters on one individual matrix could reduce up to six keystrokes per line and therefore increase composition speed. Whilst the concept originated from the development of al-Ahram’s keyboard scheme, it coincided with Linotype customers complaining about matrices wearing out because of frequent use. Tracy recognised that the proposal could help solve this technical problem while also improving keyboarding efficiency, and, having confirmed 407

According to Gabeyan, Baḥarī’s frequent initiatives rarely had much practical value: “Every time I met him, he gave me fifteen hundred ideas, I am not exaggerating. And once, once, among these fifteen hundred examples, or ideas he gave, there was one good one. And this is why I listened to him and this is why I cooperated with him.” Gabeyan, interview.

408

~45 millimeters. Gabeyan, interview.

Mechanical Composition of Arabic

201

it with Michael Adib Nahas from the MELO, the idea was implemented for all Arabic Linotype founts.409 Based on Gabeyan’s suggestions, and in collaboration with Compugraphic, the first Arabic one-button keyboard was thus designed. Its input method revolutionised Arabic composition, as now operators no longer had to key every character, but could rely on the computer for the selection of the appropriate letterform. This system increased composition speed, and became the basis for subsequent Arabic input schemes. ∵ The completion of the system and installation of the machines was to take another year, but by 4 April 1968, Tracy was able to report to Michael Parker, typographical adviser of Mergenthaler, about the successful completion of the project: Although the special Justape developed by Ellis Hanson is not yet installed in Cairo, it has been thoroughly tested and found very satisfactory. The chief engineer from al-Ahram has been here and he is delighted with the result. As you know, the computer added to the Justape does two things: it makes a choice between the several forms of a letter, and it adds kashida extension strokes as part of the justification process. It seems that it places the kashidas more sensitively than a live operator does (he tends to put the things near the end of a line without too much regard for typographic preference).

409

By September 1968 new keyboard layouts and corresponding matrices were manufactured and replacement kits advertised to customers in the Middle East. As in other cases, no public credit was given to the originators of the scheme, but internally Tracy underlined Gabeyan’s contribution: “I must emphasise that the idea is not mine. It comes from Hrant Gabeyan, who thought of it recently in his study of the ‘one character’ nonjustifying perforator keyboard, which is an essential and valuable element in the scheme we are devising for al-Ahram.” Walter Tracy, “Simplified Arabic (also Traditional Arabic),” 21 March 1967, WT correspondence, folder 18a, DTGC.

202

Chapter 2

Figure 2.101 Meeting at Linotype’s Middle East Liaison Office in Beirut at the occasion of the launch of the Elektron. From left to right: Ralph Goodman (MELO), Haïdar Hashem (MELO mechanic), Mr Pophaïdes (agent in Cyprus), unknown agent in Tunis, Hrant Gabeyan, Michael Adib Nahas (MELO), Arthur Henry Walker (L&M Sales Director), E. J. Naldrett (Linotype Pacer web-offset press Sales Director), Raymond Nassif (agent in Lebanon), Selim Tohmé (MELO mechanic), Tom Marsden (Chief Technician of Elektron linecasters), Kasim Khayat (agent in Iraq), Messrs Kabbabé Sr and Jr (agents in Sudan). Courtesy of Hrant Gabeyan.

It is very satisfying that all the hard work put into the development of the special equipment by Hanson and Hrant Gabeyan – with a little schoolmastering from me – has proved worth while.410 410

The delay of the project was probably caused by the Six-Day War in June 1967 which resulted in a devastating defeat of Egypt. Walter Tracy to Michael Parker, Letter, “Computer Arabic,” 4 April 1968, Box 3614, NMAH. Despite this appraisal of Hanson’s and Gabeyan’s work, neither of them was ever credited publicly for their achievement. The closest to a recognition of Gabeyan’s crucial role is found in an article by Tracy in which he notes: “a member of the Linotype Middle East staff spent many weeks in compiling tables of information to enable the programs to be perfected”, (“Arabic Without Tears,” in The Penrose Graphic Arts International Annual, vol. 68 [London: Northwood, 1975], 124.). Only much later Gabeyan was given the opportunity of a public forum at the Symposia on the History of Printing and Publishing in the Languages and Countries of the Middle East in Mainz (2002) and Paris (2005). His brief article in the first conference proceed-

Mechanical Composition of Arabic

203

Figure 2.102 Tom Marsden explaining the features of the Elektron at the MELO, Beirut. ­Courtesy of Hrant Gabeyan.

The installation at al-Ahram marked one of the last developments of Arabic type-making in the hot-metal era, and heralded the beginning of computeraided typography. Its technology was ground-breaking, and undoubtedly contributed to the excellent reputation al-Ahram held well beyond the Middle East: The new Ahram building, finished a year and a half after the defeat of 1967, is one of the half dozen most advanced newspaper complexes in the world. When A. M. Rosenthal, managing editor of the New York Times, visited the building in 1970 he was astonished. ‘There isn’t anything in America to compare with it,’ he stated.”411 Whilst the introduction of computers to Arabic composition was pioneered with hot-metal linecasters (Figure 2.101 and Figure 2.102), the breakthrough that revolutionised Arabic typesetting came in combination with an entirely ings is the only published trace of his work (“Modern Developments in Arabic Typesetting”); it is rarely cited and attracted scepticism in the field as it was not substantiated by any other mention of his name. 411

Sheehan, “The Most Powerful Journalist in the World, His Newspaper and His Book,” xiv.

204

Chapter 2

different technology. As Tracy noted, “the whole concept was a notable piece of creative work, and the experience gained in developing it proved immensely valuable in the next great advance in Arabic typesetting” – photocomposition.412

412

“Advances in Arabic Printing,” 89.

Chapter 3

Photocomposition: Towards Immaterial Type Beginnings of a New Technological Era Soon after the Second World War two events marked the beginning of a technological revolution that fundamentally altered the way type was made and composed. In 1946 the first experimental model of the Intertype Fotosetter was demonstrated at the US Government Printing Office, and the two French engineers René A. Higonnet and Louis Moyroud came to an agreement with the Lithomat Company that eventually led to the development of the LumitypePhoton typesetter.1 Whereas the Intertype Fotosetter was a direct adaptation of hot-metal principles to a photographic process (a so-called first-generation machine) the revolutionary approach to composition embodied in the Lumitype-Photon was ahead of its time. It boasted features that were only adopted by the competition years later and thus pointed the way ahead for typesetting technology. Such parallel developments of machines of different evolutionary stages remained a characteristic of the photocomposition era, when often radically new inventions confronted the established practices of the trade.2 From the late nineteenth century, attempts were made to develop a machine for photographic composition of text, but, with the exception of Japan where photocomposition gained traction before the Second World War, no device found commercial acceptance.3 By the 1950s, the potential of the new composition technique was apparent as a large proportion of printing 1 Seybold, The World of Digital Typesetting, 72. Seybold gives a good overview of different machine generations. The principal source for the development of the Lumitype-Photon is Alan Marshall, Du plomb à la lumière : La Lumitype-Photon et la naissance des industries graphiques modernes (Paris: Éditions de la Maison des sciences de l’homme, 2003). 2 For a discussion of the Lumitype-Photon see “Arabic Type by the Photon Company” on page 261. 3 For a number of early machines based on photographic principles see Wallis, Typomania, 21. On Uhertype see Christopher Burke, Active Literature: Jan Tschichold and New Typography, (London: Hyphen Press, 2007), 222–46; and Roger Muench, “The origins of modern filmsetting: the Uhertype: a research report”, Journal of the Printing Historical Society 3, New Series (2001): 21–40. For an overview of typographical history in East Asia see Martin J. Heijdra, “The Development of Modern Typography in East Asia, 1850-2000”, The East Asian Library Journal 11, no. 2 (2004): 100–168, accessed 16 May 2017, https://library.princeton.edu/eastasian/EALJ/heijdra_martin_j.EALJ.v11.n02.p100.pdf.

206

Chapter 3

was already done by offset lithography, instead of letterpress. Cast type, with its raised printing surface, was not directly usable for the planographic offset printing process, while, conversely, photocomposition could also be employed in letterpress printing through the introduction of flexible raised plates created photographically.4 In addition to its adaptability to newer printing processes, photocomposition eliminated some of the complications inherent in letterpress printing: storage of type and composed matter, a considerable problem with any form of metal type, became effortless with film.5 Page make-up was fundamentally altered as pieces of film, adhesive tape and cutters replaced the formes, galleys and furniture previously employed. Also the expensive and slow process of reproducing images by engraving, a necessity in letterpress printing, was done away with in the composition of pages for offset printing, another factor contributing to its increasing popularity. For Arabic typography, the flexibility of film offered additional benefits, as directionality played a smaller role in a medium that could easily be mirrored, a feature that rendered the special equipment for right-to-left composition obsolete.6 Indeed, the change from a rigid, three-dimensional object to a two-dimensional image on a flexible medium was at the centre of this new technology. Not only was the material light-weight, pliable, and redefined the process of page make-up, it also rendered questionable long-standing typographical terms. Measurement of type size had been a controversial question in the trade since the eighteenth century, and numerous attempts to devise a consistent system that would be acceptable across borders failed.7 By the middle of the twentieth century there were two principal systems in use: the Anglo-American point in the Englishspeaking world, and the Didot point in continental Europe.8 But irrespective of the system, measurement always related to the type body – the height of 4 James Moran, “Filmsetting, Bibliographical Implications,” The Library XV, no. 4, Fifth Series (1960): 232. 5 “A drawer 2 feet by 4 feet will hold film equivalent to 8,000 lb. of metal”. Ibid., 236. 6 This advantage was recognised and advertised in The Monotype Recorder: “Absent too, and cheerfully forgettable, is any special equipment for setting languages such as Hebrew or Arabic which require reverse delivery to the galley”. The Monotype Corporation Ltd, “Metal to Film: What Is Involved in the Transition?,” The Monotype Recorder 42, no. 2 (Spring 1961): 1. 7 For a concise summary of the historical developments see Tracy, Letters of Credit, 21, and Andrew Boag, “Typographic measurement: a chronology,” Typography Papers no.1 (1996): 105–22. 8 Use of the two systems extended into the countries that were formerly in the zones of influence of the respective powers, but most regions using the Arabic script employed the continental Didot point.

Photocomposition: Towards Immaterial Type

207

Drawings

24D

20D

16D

14D

12D

10D

Figure 3.1 Comparison between the drawings by Jan van Krimpen and the types made from them. Adapted from Southall, Printer’s Type in the Twentieth Century.

the piece of metal that carried the typeface.9 In the absence of the body, the notion of type size became more abstract and turned into “a concept which relates to the amount of white space necessary to accommodate the distance from the lowest descender to the highest ascender, plus clearance above and below the letter”.10 But the notion of type size changed its meaning not only in relation to its measurement, but also in relation to its appearance. In the manual cutting of punches, practised prior to the mechanisation of typesetting, every individual fount was an original and thus allowed for fine adjustments of character shapes for each size of a typeface (Figure 3.1). With the introduction of the pantograph for the making of matrices as used in hot-metal composition, a single original drawing could be used to create multiple founts. The minute adjustments of character shapes, required for an ideal reproduction in different sizes, was hence compromised, with effects detrimental to the quality of the printing type. Conscious of the effect, the principal machine manufacturers retained the practice of employing more than one master drawing to produce a range of founts – a distinct set of matrices being necessary for each size.11 9

This principle accounts for the different apparent sizes of founts of nominally equal dimensions. A face which is ‘small on the body’ therefore appears smaller than the nominal point size would suggest.

10

Seybold, The World of Digital Typesetting, 35.

11

According to Ross, normally three masters were used for all sizes of a typeface. The Print-

208

Chapter 3

Figure 3.2 “Intertype presents New Horizons for the modern printer,” promotional publication for the Intertype Fotosetter, n.d., reduced to 50% linear, DTGC, locker C1.

Figure 3.3 Fotosetter single- and doubleletter matrices. From Phillips, Computer Peripherals.

However, with the advent of photocomposition a single set of matrices could be employed to create the entire range of type sizes that the machine was able to produce. Where this was practised, the results were inferior to those obtained from hot-metal composition, as proportions and features of the character images could no longer be optimised for their intended use. Moreover, the employment of a single matrix for the setting of different sizes introduced issues inherent to the photographic process. An increase or decrease of the projected image results in a change of intensity of the light that reaches the photographic surface, causing irregular exposure. To compensate for this problem, different means were employed: filters could be mounted to alter the light intensity, the aperture of the lens could be adjusted, and some devices were programmed to double or triple expose the image for large sizes “where the degree of magnification was so great that the character might otherwise ­appear

ed Bengali Character and Its Evolution, 185. Southall remarked that Frederick Goudy’s practice to cut “a 9:1 range of sizes from a single set of patterns” was unusual. Printer’s Type in the Twentieth Century: Manufacturing and Design Methods (London & New Castle, DE: The British Library & Oak Knoll Press, 2005), 24.

Photocomposition: Towards Immaterial Type

209

to be ‘washed out’”.12 In reference to the use of a single set of matrices for all sizes, Seybold stated that “very small point sizes – such as six point – seem too light, and some of the serifs tend to disappear”.13 Monotype reproduced its hotmetal approach of size-dependent master drawings for the first generation of photocomposition devices it manufactured, offering A, B and C matrices specifically conceived for size ranges; according to Boag, however, “most customers just set all sizes from the B matrices”.14 Developed as an evolutionary step from existing hot-metal casters, the first-generation filmsetters were largely based on their predecessors. The mechanics resembled those of the hot-metal machines as much as possible, as exemplified by the Intertype Fotosetter, the first commercially successful device (Figure 3.2). It employed the same technique for the release, assembly and redistribution of matrices into the magazine, only its justification mechanism was changed and its metal pot was replaced by a light source. Its brass matrices contained a photographic negative as character image, but featured the same teeth and notches that governed their redistribution (Figure 3.3). The light source exposed individual, stationary matrices, projecting the character image through a lens on to the film carriage. The matrix defined the width of the character image, and controlled the film advance, but contrary to the hotmetal machine no spacebands were employed for justification. Empty blanks of a uniform width were inserted when a space was keyed, and once the line was completed, the camera mechanism measured the difference between the width of the assembled matrices and the desired line length. The resulting discrepancy was then automatically distributed throughout the entire line, in effect letterspacing all characters.15 Among the advantages of the Fotosetter was the departure from the linecasting principle. Every matrix was individually exposed, and thus allowed for kerning, a feature that was emphasised by the manufacturer as the possibility of “close fitting composition […] for all letter combinations” (Figure 3.4 overleaf).16 The machine’s output was advertised 12

Seybold, The World of Digital Typesetting, 95.

13

Ibid.

14

Andrew Boag, “Monotype and Phototypesetting,” Journal of the Printing Historical Society no. 2, New Series (Winter 2000): 60.

15

Seybold, The World of Digital Typesetting, 74. Later models resolved this drawback and provided means to add additional space only between words. This initial shortcoming is indicative of the non-linear nature of technological progress, where a newer product is not necessarily better than its predecessor.

16

C. L. Boileau, “The Intertype ‘Fotosetter’ Machine,” in The Penrose Graphic Arts International ­Annual, vol. 47 (London: Lund Humphries Publishers Ltd, 1953), 91.

210

Chapter 3

Figure 3.4 Optical system of the Fotosetter. Matrices are exposed individually, the film carriage moves according to the set width of the matrix. From Phillips, Computer Peripherals.

Figure 3.5

Close-up of individual Monophoto film matrices against a lit background. ­Photograph by the author.

at a speed of eight characters per second, the maximum that the camera unit was able to expose, noting that this was “in excess of manual operation of a keyboard”.17

17

Ibid., 92. It follows that the speed was determined by the skill of the operator, and implies that no direct advantage in composition efficiency could be obtained from the Filmsetter.

Photocomposition: Towards Immaterial Type

211

Monotype began its research into a photocomposition machine in 1944. In volumes of the Penrose Annual for 1949 and 1950 the new machine was announced publicly, and by 1952, the Corporation exhibited first experimental versions of the Monophoto to “selected confidants”.18 Its mechanics resembled those of the hot-metal caster in many ways, and the keyboard producing the paper-tape was almost identical.19 Whereas the Monophoto Mark 1 employed a single glass plate containing 15 rows of 17 matrices, all later models used a metal carrier that held individual film matrices in much the same manner as the hot-metal matrix case (Figure 3.5).20 This resurrected flexibility, introduced in the Mark 2 of 1957, contributed to its acceptance in the marketplace.21 Largely similar principles were maintained, incorporating only gradual improvements, up to the Mark 5, the final version of the machine introduced in 1969: .

All five releases […] slavishly adopted the basic mechanical principles of the metal keyboard and caster: Monotype seemed uneasy with electronics, and the continued dependence on 31-channel tape shows that they were also uneasy about the application of computers to typesetting.22 Also the size-range available remained identical from 6 to 24 point, and the composition speed of under three characters per second (increased by one in the Mark 5) was about the same as the hot-metal caster.23 This apparent 18

Elfried Silcock, “‘Monotype’ Photo Typographical Composing Machine,” in Penrose Annual, vol. 44 (London: Lund Humphries & Co. Ltd, 1950), 102. Wallis underlines that this machine was not based on the Rotofoto machine, developed in the late 1930s by George Westover as an endeavour independent of the Corporation; “Monotype Time Check,” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 50. Boag, however, purports a link between Monotype and Westover’s invention. “Monotype and Phototypesetting,” 58. This supposition is not supported by the earliest patent for the Rotofoto: Westover, George. Improvements relating to Methods and Apparatus for Photographically Composing Text-matter. UK Patent 499,859, filed 27 April 1937, and issued 27 January 1939.

19

Boag, “Monotype and Phototypesetting,” 59.

20

From 1963 the case contained 272 matrices in a 16 × 17 arrangement.

21

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 29. The first Monophoto Filmsetter was installed in the United Kingdom in the same year.

22

Boag, “Monotype and Phototypesetting,” 62–63.

23

Phillips, Computer Peripherals and Typesetting, 465. According to Phillips the maximum speed of Mark 1 and 2 was at 9,600 characters per hour, which translates to 2.6 per second.

212

Chapter 3

c­ onservatism may, however, have been informed by very pragmatic considerations. Similarity of concepts brought about similarity of use and thus facilitated the shift in technology. Operators of hot-metal machines required little, if any, retraining on the early Intertype or Monotype filmsetters, therefore allowing for printers’ transition “without investment in equipment beyond both their comprehension and their means, and without much disturbance of labour”.24 A further, crucial, factor in favour of Monotype over its competitors was the company’s reputation for the versatility and reliability of its equipment, and the quality of its typeface catalogue. The latter was adapted fairly rapidly for photocomposition, and new designs were commissioned for the new technology.25 The high speed at which designs were converted was facilitated by the relative ease of the process. In contrast with the multistaged and laborious processes of matrix-making for metal type, matrix-making for direct-photography photocomposing machines is conceptually very simple. Large character images – character masters – are photographed to produce small character images on the matrix.26 But as David Saunders, a long-time employee of Monotype pointed out, during the adaptation of the company’s hot-metal type library to photocomposition, “some aspects of the change in technology were not fully acknowledged in working practice.”27 At first, existing copper patterns, previously used for pantographic punchcutting, were painted – the character black, the background white, photographed and used for matrix making. In this process hairlines often broke down, and manual corrections of the new photographic masters were needed.28 Thus, the apparent simplicity of this process was untenable, and the direct adaptation of existing hot-metal character configurations (either as drawings or patterns) by no means guaranteed a successful rendering of the face in the new composition and printing processes. Divergence from the expected appearances of established typefaces was often inevitable, and the speed at 24

Moran, “Filmsetting, Bibliographical Implications,” 233.

25

Boag, “Monotype and Phototypesetting,” 75.

26

Southall, Printer’s Type in the Twentieth Century, 86.

27

David Saunders, “Two Decades of Change 1965–1986,” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 27.

28

Ibid.

Photocomposition: Towards Immaterial Type

213

which typefaces were adapted for the new technology did not always allow for the required revisions.29 Arabic on the Monophoto In contrast to the late development of Arabic for the Monotype hot-metal caster, begun towards the end of the life cycle of the technology, there was no delay in the adoption of Arabic for photocomposition. In April 1957, the year that the Monophoto “clinched initial commercial acceptance” and saw its first installations in South Africa, the United Kingdom, and the United States, Series 549 was scheduled as the 13th typeface that was to be made for the new machine.30 In August, a pending order from the Board of Secondary Education in Lahore, Pakistan, added a commercial perspective to Arabic Monophoto founts.31 The client enquired about Series 507 and 549, specifying a range of sizes, together with the Latin typefaces Bembo and Times.32 In an initial report, written about one month later, the Works summarised the state of the potential order. At this point, and contrary to the first enquiry, Series 507 was no longer required; and instead of Series 549, Series 589 was discussed because its accents offered a wider range of use. Because there was no prior experience with the making of Arabic photographic matrices, it was suggested that a few experimental characters be made to ascertain that characteristics of the design would be maintained when reduced for the glass negative. But in the report’s last paragraph, the Works articulated one of the most pertinent issues in the adaptation of a design from one technology to the other:

29

Long-standing Monotype employee Robin Nicholas recalled that “there was a scramble to get the typeface library into new formats” and that “in those early days of phototypesetting a lot of the typefaces didn’t get the care and attention that they should have done. It was a case of getting them available as quickly as possible”. Simon Esterson and John L. Walters, “Robin Nicholas,” Eye Magazine 21, no. 84 (Autumn 2012): 72.

30

Wallis, “Monotype Time Check,” 51. The Monotype Corporation Ltd, “‘Monophoto’ Typography - Progress Schedule,” 11 April 1957, correspondence folder Monophoto, MT. Arabic is preceded by multiple sizes of Times, (items 1–4), Grotesque, Gill, Times Semi-Bold, Baskerville, Walbaum, Devanagari, Bodoni and Plantin.

31

E. A. Firmage to C. Fellows, Letter, 14 August 1957, correspondence folder Monophoto, MT.

32

The order of distinct sizes is indicative for the novelty of the technology, as it appears that the capability to produce multiple sizes from a single set of matrices was not yet appreciated.

214

Chapter 3

In Exotic founts the existing designs have usually been planned purely on the basis of casting in type and can, therefore, have features which are unnecessary in photographic setting. It has therefore been suggested that to a varying extent more efficient and acceptable designs could be made for ‘Monophoto’, if these founts were re-planned and redrawn for ‘Monophoto’ only. It is realised that the work and time factor involved would of course, create quite a delay in supplying the first customer’s requirements.33 This instance is a rare example in which the technological restrictions inherent in a type’s design where recognised and explicitly articulated. It demonstrates the Works’ critical position in the manufacturing chain, and indicates its concern for best practice. It also highlights the permanent, often incompatible, tension between pragmatic necessity in the form of customer requirements, time frames and budgets, and the wish to produce the best possible product. Here, the potential to liberate the design from its precursor was recognised by the management, as evident from its reply: We have considered the points raised in your memorandum regarding Arabic Urdu and it is felt that the filmsetting of this face should not necessarily have the restrictions imposed by the hot metal method.34 How the new composition technology was going to be used to overcome the said restrictions remained unspecified, and further evidence about the making of this fount has not been found. The urgency of the need for the adaptation of Monotype’s existing type catalogue to the new technology may account for the decrease in detailed records. Whereas the design and manufacturing process is well documented up to the 1950s, the following decades lack a comparable depth of sources. By 23 April 1958, a letter to Chafik Mitry, the Egyptian publisher who had been instrumental in the development of Series 589 for hot-metal, informed the client that the Corporation had recently produced a version of the typeface for the Monophoto.35 It noted that the enclosed film showed the fount in 24D, and 33

Works to Head Office, Letter, “‘Monophoto’ Typography,” 11 September 1957, 2, correspondence folder Monophoto, MT.

34

Head Office to Works, Letter, “‘Monophoto’ Typography Arabic Series 589,” 23 September 1957, correspondence folder Monophoto, MT.

35

E.  A. Firmage to Chafik Mitry, Letter, 23 April 1958, folder Arabic (early correspondence), MT.

Photocomposition: Towards Immaterial Type

Figure 3.6

215

“Arabic in All Its Glory”, announcement of Monophoto Arabic Series 589 (detail). From The Monotype Recorder 42, no. 2 (Spring 1961), 24, actual size.

that three further sizes had been composed from the same matrices with pleasing results. Whereas no printed proof of this date could be found, The Monotype Recorder featured the typeface in 1961 (Figure 3.6).36 Shown in two sizes (12D, 24D), the small sample does not, however, appear to reflect a design revision that overcame the restrictions imposed by hot-metal composition. Rather, the same features and shortcomings seen in the earlier version seem to be copied: characters whose joining strokes had been elongated to ­accommodate

Figure 3.7 Comparison of hot-metal (left) and photocomposition (right) versions of Series 589, enlarged to 200%.

36

The Monotype Corporation Ltd, “Arabic in All Its Glory,” The Monotype Recorder 42, no. 2 (Spring 1961): 24.

216

Chapter 3

vowels on their bodies retained their proportions despite the obsolescence of the dimensional constraints of metal type; the accents did not benefit from new vertical positioning in relation to the character height, but largely kept their compromised configurations; and the spacing of the new photographically composed fount shows no improvements, with certain sequences appearing distinctly better in the hot-metal version (Figure 3.7 on page 215). How the typeface was eventually adopted to the new technology is not apparent. In lack of documentation of the making of Monophoto Series 589, the way in which Monotype revised other Arabic series may shed light on the approach that was taken in adapting hot-metal designs to photocomposition. In January 1960 the Corporation began investigating the making of Series 507 for the Monophoto. The new version was to be based on the 12D hot-metal fount but designated as 16 pt, whilst allowing for a range from 12 to 24 point. The principal considerations were related to the size of the face that could be accommodated on the matrix, the overhangs and their relation to the line feed, the alignment of characters and the integrity of hairlines in photographic reduction.37 According to the Works’ second conversion proposal, approximately half of the 550 characters had to be remade because hairlines in the drawings and copper patterns had to be thickened.38 It appears, though, that although a degree of strictly mechanical adaptation was required, no further re-evaluation of the design was pursued. Whereas previous Arabic typeface developments necessitated repeated consultation with clients and numerous trials that often led to changes and revisions, the records for the making of the Monophoto version of Series 507 are brief. Following the Works’ initial investigation in January, manufacture proceeded according to the customer’s requirements.39 By May trials of the new fount were approved, and the development was completed no later than November 1960.40 37

H. R. Knight, Arabic 507 Series, Handwritten report, 16 February 1960, 4, correspondence folder Arabic 507, MT.

38

H. R. Knight, 507 Arabic (Scheme 2), Handwritten report, 24 February 1960, 1, correspondence folder Arabic 507, MT. Characters which featured hairlines thinner than .0015 inch at 8 pt were remade.

39

Type Drawing Office, “507 Arabic,” card index, 16, minutes box, MT. The identity of the customer could not be established, but successive documented use of the typeface by publishers in Iran and Pakistan suggests either of them as the clients of the first installations.

40

The Secretary, Typographical Committee to C. A. Poore, Letter, “‘Monophoto’ Trials Arabic, Series 507-12, 16, 24 Point Dated 20. 4. 60,” 9 May 1960, correspondence folder Arabic Naskh Bold 649/659, MT.

Photocomposition: Towards Immaterial Type

Figure 3.8

217

“Page from Telephone Directory, ‘Monophoto’ Filmsetter work, printed by Bank Melli Press Iran, Tehran,” sample of Arabic Series 507 (detail). From The Monotype Recorder 42, no. 2 (Spring 1961), no page number (detail), actual size.

By the following issue of The Monotype Recorder in spring 1961, three Arabic Monophoto faces were reportedly in use with clients in Pakistan and Iran, and a short article featuring samples of printed matter from Iran underlines the potential of the technological shift for Arabic (Figure 3.8): Of all the major scripts of the world those within the ‘Arabic’ group have most to gain from the coming of the Filmsetter. Not only can an essentially ‘calligraphic’ script be set forth with all the advantages of linked and overlapping characters to its obvious aesthetic advantage: what is just as striking to the printer is the economic gain made possible by a method which involves no ink-spread thickening of the sometimes very small ‘counters’ […] A size of Arabic which would be difficult to read in a print from metal type remains brilliantly clear and legible in filmsetting.41 Although in principle correct, the advances Monotype’s promotional text assigns to the new technology are modest. The sharper printed image obtained from photocomposition and lithography was not unique to Arabic, and incidentally, Monotype’s hot-metal composition had proved that letterpress 41

The Monotype Corporation Ltd, “Arabic in All Its Glory,” 24.

218

Chapter 3

Figure 3.9

“12 point Urdu No. 2”, March 1952, enlarged to 200% linear. NMAH, box P3640.

­ rinting could in effect eliminate the gaps between joined letter groups. A secp ond article in this issue of The Monotype Recorder discusses the introduction of the Monophoto by the Pakistani newspaper The Daily Jang, noting that “what ‘film’ supersedes, in this historic instance, is not type of any kind, good or bad, foundry or machine: it is calligraphy.”42 It underlines the difficulty of typographical composition of Nastaʿlīq and the resulting continuity of handwritten daily newspapers in Pakistan; in apparent reference to Mergenthaler’s earlier efforts at the development of an Urdu fount (Figure 3.9), the article notes: One attempt, about ten years ago, to go over to type had ended disastrously; the method of composition used sorely handicapped the design of the face, and the reaction of the readers was such that the production of the paper was suspended for good after about a week.43 According to The Monotype Recorder, The Daily Jang was going to be composed in the newly made Monophoto Urdu 549 and its related bold weight, Series 649, in what was described as a “momentous change – from the scribe’s pen direct to filmsetting, by-passing Gutenberg’s invention of metal type”.44 42

The Monotype Corporation Ltd, “‘By-Passing Gutenberg’ in Karachi,” The Monotype Recorder 42, no. 2 (Spring 1961): 13.

43

Ibid., 14. Although clearly intended as a sneer at the competitor, the fact remains that Mergenthaler’s fount was completely unacceptable and that it contributed to the closing of the Dawn newspaper. According to the indologist Norman Brown, acting as a consultant for Mergenthaler, “several reasons had been responsible for its death, among which one that was commonly cited was Dawn’s insistence upon printing in our face […] This seems to confirm the feeling you and I had that our old face should never have been allowed to go to Pakistan”. W. Norman Brown to Jackson Burke, Letter, “Urdu,” 30 June 1951, Box P3640, NMAH.

44

The Monotype Corporation Ltd, “‘By-Passing Gutenberg’ in Karachi,” 14.

Photocomposition: Towards Immaterial Type

Figure 3.10

219

“‘Monophoto’ Urdu Series 549,” specimen (detail). From The Monotype Recorder 42, no. 2 (Spring 1961), insert, actual size.

Yet, the samples of the face shown would not suggest its wide acceptance in Pakistan (Figure 3.10). The design is hardly adapted to the letters and patterns found in Urdu, instead strongly resembling the Egyptian fount Monotype’s Series 549 was based on. The newly introduced characters lack the quality of the original foundry type, and their combination makes for an unfavourable hybrid, marred by inconsistencies, poorly executed typeforms and a lack of authenticity. Rather than a new approach to the typographical composition of Urdu, which photocomposition could have allowed, Monophoto Series 549 Urdu must be considered an ad-hoc solution of little merit; a result that is not surprising given the speed of new fount manufacture. Expansion of Monotype’s Arabic Type Catalogue In parallel to the adaptation of the existing series to the new technology, corresponding bold weights were made for hot-metal composition. The provision of a bold weight matching Series 549 had been urged by Monotype’s Egyptian agent Lindell from 1953, but manufacture began only in late 1957 after repeated demands by the Corporation’s representative in Beirut, C. T. Day.45 Day 45

In April 1957 Day explained the need for additional weights thus: “Usually when I am discussing Arabic with clients and prospects, I am confronted with Linotype and Intertype Brochures in Arabic which feature quite a good range of light and heavy Arabic faces and

220

Chapter 3

Figure 3.11 “Trial No. 1 ‘Monotype’ Arabic Naskh Bold, Series No. 649, 16D,” June 1958, reduced to 60% linear. MT, folder master proofs.

s­ uggested basing the bold version on Series 549, but added the unconventional request for every character to share the same set width in the heavier fount.46 According to Day, this scheme had already been employed for Devanagari and could facilitate the automatic composition of regular and bold weights in one line; he did not, it appears, appreciate the compromises of shape and proportion that the drawing of bold characters within the dimensions of the regular entailed.47 Without apparent consultation with either Lindell or a client, Monotype proceeded with Day’s suggestion and produced a first proof of Series 649 Arabic Naskh Bold in 16D by June 1958, noting that the fount was planned “on the same set, units and matrix sizes as Series 549-16D in order that the two Series are interchangeable” (Figure 3.11).48 By September, both people cannot understand why it is that we have produced such a good Light Arabic Face and do not appear to be able to do anything about a heavy face to work with it.” C. T. Day to Percy Goodall, Letter, 28 April 1957, folder Arabic (early correspondence), MT. 46

Quoted from E. A. Firmage to The Secretary, Typographical Committee, Letter, “Bold Arabic,” 5 November 1957, folder Arabic (early correspondence), MT.

47

In a curious contradiction this suggestion brought the design constraint of Linotype’s duplexed founts to the Monotype system.

48

The Secretary, Typographical Committee to E. A. Firmage, Letter, 3 July 1958, folder Arabic (early correspondence), MT.

Photocomposition: Towards Immaterial Type

Figure 3.12

“Trial No. 2 ‘Monotype’ Arabic Naskh Bold, Series No. 649, 16D,” April 1959, reduced to 60% linear. MT, folder master proofs.

221

222

Chapter 3

Day and the Karachi office of the Corporation conveyed the opinion that the fount was satisfactory, and its completion was approved.49 A second trial from April 1959 shows the new weight in an advanced state and, according to a letter to Lindell, it was ready for commercial distribution (Figure 3.12 on page 221).50 Designed on 13¼ set, the face shared the same character widths as the regular, yet had to accommodate the thicker strokes typical of a bold weight, a constraint with noticeable – and at times detrimental – effect on the proportions of letterforms (Figure 3.13). These compromises of design, it appears, were thus accepted to economise in terms of development and equipment. Because the bold fount had an identical synopsis as the regular, it facilitated practical tasks. The two weights could be easily substituted, saving clients the purchase of dedicated equipment. A fount change thus only required the substitution of the matrix case, rather than keybanks, keybars, stopbars and the justifying scale. Copyfitting tables were equally applicable to either fount. Whilst Series 649 was a late development for hot-metal casters, the approach taken in the design of this bold weight was maintained in successive photocomposition adaptations. From September 1959, Series 659 was developed as a complementary bold weight for Series 589 on the Monophoto. The specification for the new fount noted that the drawings should be made to be suitable for photographic as well as hot-metal matrices and for the design to be “as 649 is to 549”.51 The document gives further insights into the changing attitude towards new typeface developments that appear to be increasingly influenced by time and efficiency considerations. It stated that the original drawings of Series 589 and 649 should be used for the new fount, and that one

Figure 3.13

Comparison of set widths of characters in Monotype Series 549 and Series 649. Note how the proportions of characters in the bold weight are compromised as they are crammed into the widths of the regular weight. Illustration by the author.

49

C. G. Turner to The Secretary, Typographical Committee, Letter, “Arabic Naskh Bold, Series 649-16D,” 23 September 1958, folder Arabic (early correspondence), MT.

50

E. A. Firmage to Joseph Lindell, Letter, 24 April 1959, folder Arabic (early correspondence), MT.

51

John Goulding, “Specification for New Fount,” 30 October 1959, correspondence folder Arabic Naskh Bold 649/659, MT.

Photocomposition: Towards Immaterial Type

223

single set of drawings had to serve for both Monophoto B-range photographic matrices and 24D hot-metal matrices; the priority was classified as “urgent”, an indication that may explain the note that no trials were required.52 This sense of urgency is equally found in the making of Series 707, a bold fount based on Series 507; initially commissioned as a hot-metal fount in 1960, by 1961 a photographic version was commenced in parallel, as noted in the specification: “This Series has been made fairly recently in type – in fact it has not yet been finally proofed and approved. Whilst waiting however we must start making for film”; as in Series 659, no trials were required and its priority was defined as “soon as possible”.53 As has been pointed out above, the pace of adaptations was not limited to Arabic founts, but extended to the entire typeface catalogue of the Corporation. In a change to earlier practice, and in contrast to Monotype’s reputation for typographic quality, the technological shift imposed a new and different logic to type-making. It appears that acceptable, rather than exceptional quality, had to suffice for the Corporation in the early years of the photocomposition era, when a speedy offer of founts for the new machines took priority over the thorough revision and adaptation of designs: The breaking down of fine lines was immediately evident, but the general loss of weight that would arise from the lack of inkspread and the split ink film in the offset lithographic process was overlooked. The more radical treatment necessary to compensate for the differences between, on the one hand hot-metal composition and letterpress, and on the other hand, filmsetting and lithographic printing, was not applied.54 The urge to extend Monotype’s Arabic programme during the first half of the 1960s was not confined to the adaptation of existing designs to the new machines and the making of derivative weights, but also entailed original typeface developments according to customer demands (Figure 3.14 overleaf).55 52

Ibid.

53

John Goulding, “Specification for New Fount,” 20 October 1961, correspondence folder Arabic Naskh Bold 649/659, MT.

54

Saunders, “Two Decades of Change 1965–1986,” 27.

55

Among these developments was an experimental Urdu Nastaʿlīq design for photocomposition. Designated as Series 679, trials of the type were made (Figure 3.15 overleaf), but the resistance to a compromised form of the script proved too significant to ascertain its acceptance in Pakistan. In the assessment of the manager of the Eastern Area (the Indian subcontinent), L. A. Collier, the type was “perfectly legible and infinitely bet-

224

Chapter 3

Figure 3.14

Photographic reproduction of a Nastaʿlīq design provided by a certain Mr Minai. Although it was attempted to create a typeface based on this model, no finished implementation resulted from it, actual size. MT, correspondence folder Urdu Nastaliq.

Figure 3.15

(a) “Urdu Nastaliq, Series 679, 10-pt”, trial proof, n.d., actual size, (b) box to the right shows enlargement of the text. MT, correspondence folder Urdu Nastaliq.

Photocomposition: Towards Immaterial Type

225

Among them was a new type development that a customer required for his recently ordered Monophoto filmsetter. Simplified Arabic for Monophoto In early 1962, Monotype’s representative in Beirut, D. Stevens, conveyed to the Home Office that a potential purchase of a Monophoto filmsetter was subject to the provision of a simplified Arabic typeface. The customer was Friar Cortbawi, “a priest in charge of a School in Beyrouth”, who wished to have two weights in a single matrix case because much of his work required the mixing of regular and bold within a single line.56 According to Stevens, the client insisted that a simplified design was necessary in order to “get around the limitations of the matrix case capacity” – a reasonable observation as Monotype’s Arabic founts were hitherto limited to a single weight per MCA. Stevens’ assessment of this request, however, suggests that considerations other than pragmatic ones may have informed the client’s desire: The Simplified Arabic face in Light and Bold appears to be growing in popularity, in keeping with the Lebanese character that one must be upto-date in all things, whether good or bad, the main thing is to be with the latest. It is suspected that Fr. Cortbawi is under some pressure by his customers and also as a leading modern printer feels he should conform with the way of thinking.57 Despite Stevens’ suggestion that the appeal of Simplified Arabic – introduced only three years earlier – was driven by a Lebanese fashion, the fact remains that Monotype felt pressure to provide equipment that could compete with the efficiency of linecasters. From June 1962 the Corporation began investigations into the adoption of simplification principles for its founts and

ter than, and closer to Nastaliq than, the Linotype Nastaliq face, but we cannot pretend that it resembles calligraphed Nastaliq very closely.” L. A. Collier to E. A. Vesey, Letter, 26 August 1965, 1, correspondence folder Urdu Nastaliq, MT. 56

E. A. Firmage to L. A. Collier, Letter, 3 May 1963, 2, correspondence folder Urdu, MT. Although not attested by evidence, this probably refers to the Imprimerie Catholique de Beyrouth, one of the principal publishers in Lebanon during this period.

57

Anonymous, “Simplified Arabic, Extracts from Two Letters from Mr D. Stevens,” n.d., correspondence folder Arabic (Egyptian 2), MT.

226

Chapter 3

machinery and started with a review of existing approaches.58 In a letter to the patents manager requesting legal advice, John Dreyfus, Monotype’s typographical adviser, gave an incisive account of the events surrounding the release of Mrowa-Linotype Simplified Arabic: Mrowa is a newspaper proprietor in Beyrout [sic]. He patented his system in Lebanon. Linotype has patented the system in eight other Arabicspeaking countries. Intertype has stolen the system and have introduced one improvement into their version. To Tracy’s regret, Linotype decided not to go to law against Intertype over this piracy.59 From the same letter it emerges that Dreyfus strongly discouraged copying the design without Linotype’s permission, and, crucially, that Monotype’s agents in the region estimated that “the major part of printing done to-day is in this system” (see also page 120 and following).60 Simplified Arabic thus had not only found wide acceptance in a relatively short time-span, but it started to dominate the appearance of printed Arabic. Incidentally, and decisively for the Corporation, it became an expected style for typographical composition: Cortbawi, the prospective client from Beirut, made the fount’s manufacture a condition for his purchase of the company’s filmsetter. Monotype thus decided to follow Linotype’s design, and approached the competitor to obtain permission to copy its simplification principles.61 The new fount was to be based on Series 549 and reduced to the 121 character synopsis of the linecaster layout (90 magazine and 31 side case characters); as this new fount merely meant a simplification of the existing design, it was anticipated that three-fifths of the characters could be taken from the existing Series and only 45 to 50 needed a revision in accordance with the new layout.62 However, against Monotype’s expectation, Linotype did not give permission for use of its simplification

58

John Goulding, “Simplified Arabic (Mrowa-Linotype),” 4 June 1962, correspondence folder Arabic (Egyptian 2), MT.

59

Anonymous to D. G. Fletcher Rogers, Letter, “Simplified Arabic (Mrowa-Linotype),” 7 June 1962, 1, correspondence folder Arabic (Egyptian 2), MT.

60

Ibid.

61

An internal memorandum notes that the Corporation has “started negotiations with Linotype and [does] not anticipate any difficulty regarding the use of the system”. Anonymous to C. A. Poore, Letter, “Simplified Arabic,” 22 August 1962, correspondence folder Arabic (Egyptian 2), MT.

62

Ibid.

Photocomposition: Towards Immaterial Type

227

­ rinciples on the Monophoto.63 Whilst the development of a hot-metal verp sion was granted, it was considered uncommercial to allow the typeface’s adaptation to the new technology as long as Linotype had not established its own photocomposition programme and it was agreed to reconsider the position in early 1964. However, by June 1963 an internal Monotype memorandum sheds light on the renewed attempts that were made to acquire the rights for the manufacture of a simplified Arabic fount on the Monophoto: I emphasised to Tracy that we had received an order from the Lebanon for a Filmsetter, and whilst we had no intention of doing anything in an unethical manner, we were, nevertheless, very concerned if Linotype insisted in not agreeing to make an arrangement with us. I told him that I was pretty certain that the news would very soon get to the ears of the energetic Intertype representative in the area, and that Intertype would certainly have no scruples in copying the face in film, as they had already done in hot metal.64 Although no further documentation of the exchange between the two companies has been found, it appears that this commercial pressure had its desired effect. Whether Linotype eventually agreed to the arrangement, or whether Monotype pursued the development regardless is not known, but by January 1965 the schedule for new fount developments for the Monophoto lists Arabic Modern and bold, Series 749 and 759 as item three, in progress at the Type Drawing Office.65 No detailed evidence of the consequent development is extant, but according to references in the Monophoto schedule the fount seems to have been proofed in November 1965 and approved for manufacture by January 1966.66 An undated proof displays a sample of the fount in five sizes (Figure 3.16 overleaf). Judging from this showing, Series 749 was Monotype’s least successful Arabic typeface to date, failing on multiple levels. Although warned by Tracy about Linotype’s experience in making the first Simplified Arabic version, suggesting that “it was not a satisfactory solution to try and introduce characters 63

Vice President and Secretary Mergenthaler Linotype Company to Sir F. G. Wrisberg, Letter, 8 March 1963, Box 3614, NMAH.

64

The Monotype Corporation Ltd, untitled typescript, 13 June 1963, correspondence folder Arabic (Egyptian 2), MT.

65

The Monotype Corporation Ltd, “Typography Schedule ‘Monophoto’ Filmsetter,” 1 January 1965, correspondence folder Monophoto, MT.

66

The Monotype Corporation Ltd, “Typography Schedule ‘Monophoto’ Filmsetter,” 3 January 1966, correspondence folder Monophoto, MT.

228

Chapter 3

Figure 3.16

“Arabic 749,” proof, n.d., reduced to 70% linear. MT, correspondence folder Arabic Naskh.

from an existing fount”, Monotype pursued this route to a predictably poor effect:67 joining is uneven, and does not give the required impression of continuity in joined letter groups; no formal consistency is achieved as individual characters appear disjointed, probably a consequence of the adaptation of an existing typeface without sufficient re-evaluation. In particular the newly created characters lack the quality of the hot-metal precursor; it appears that the Corporation developed this typeface without consultation or advice by either clients or company representatives in the Middle East as had previously been 67

The Monotype Corporation Ltd, untitled typescript.

Photocomposition: Towards Immaterial Type

229

the case, and the defects of the design suggests an apparent lack of expertise in the Arabic script. Whereas the Drawing Office successfully adapted existing foundry type to the Monotype system, Series 749 betrays the Corporation’s limits in creating new Arabic founts without guidance. The inadequacy of the fount may perhaps also explain the scarcity of information pertaining to this design. The expansion of Monotype’s Arabic programme in the 1960s appears driven by immediate demands, and lacking a coherent, informed strategy. The upheaval caused by the change of technology may have contributed to rushed, short-sighted adaptations of founts, but cannot account alone for the consistent application of poor judgements during this period. Here, it may be suggested that the Corporation’s lack of a typographical adviser who was trained and experienced in Arabic typography became a liability as soon as the market expanded. Without the discriminating clients who were behind Monotype’s first hot-metal Arabic types, the Corporation appeared at a loss in its decisions, committing to unsuccessful designs and producing mediocre founts. In contrast to Linotype, Monotype’s hot-metal début in Arabic typography was generally superior to its subsequent type developments for photocomposition. Arabic on Linotype Photocomposition Machines In contrast to Intertype and Monotype, Linotype’s début in photocomposition entailed a complete break with its hot-metal technology. The company’s first device developed for photomechanical composition was announced in 1955 as the Linofilm, and it had little in common with a linecaster (Figure 3.17 overleaf).68 It was a typical ‘second-generation’ machine, implying that it was no longer based on a hot-metal precursor, but conceived wholly as a phototypesetter. Devices of this kind still used photographic character images which were directly projected on to photo-sensitive film or paper. But other design characteristics of second-generation typesetters provided higher speed, larger character sets, and increased flexibility over their precursors.69 Most secondgeneration machines had two units: a dedicated keyboard with additional functional controls, and a photographic unit. The keyboard unit produced a coded paper-tape to drive the photo unit. 68

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 28.

69

The notion of machine generations is discussed in detail in Seybold, The World of Digital Typesetting, 84–112. Southall is critical of Seybold’s definition as “it lumps together in his third generation of machines the fundamentally different technologies of scanned matrix and digital photocomposition.” “A Survey of Type Design Techniques before 1978,” 49.

230

Chapter 3

Figure 3.17

Linofilm system, consisting of the photo-unit (left), and the keyboard unit (right). Note the tape reader on the right of the photo unit. From Phillips, Computer Peripherals and Typesetting.

Figure 3.18 A Linofilm fount grid holding 88 photographic matrices. From Phillips, Computer Peripherals and Typesetting.

From the Linofilm keyboard 18 fount grids, each holding 88 characters, were accessible, thus providing a total of 1,584 characters (Figure 3.18).70 The range of sizes from 6 pt to 54 pt was obtained from the device’s optical system with a built-in zoom-lens.71 Further to the increased character repertoire and the 70

Phillips, Computer Peripherals and Typesetting, 497.

71

However, Southall points out that “the limited magnification range of the Linofilm’s zoom-lens optical system meant that a total of five grids with different character image

Photocomposition: Towards Immaterial Type

231

relative ease of changing sizes, the system featured much improved means for justification. In contrast to linecasters, the Linofilm was able to letter-space and kern, providing a crucial improvement to linecaster composition.72 Linotype did not adopt Arabic for its first photocomposition devices; indeed, the Linofilm’s slow development does not suggest high demand: it went into commercial production in 1959, a full four years after its first public appearance, and the first book set on the machine in the United Kingdom was made no earlier than 1962.73 During the 1950s and 1960s, phototypesetters remained prohibitively expensive for many printers, and the big installations of newspapers shifted only slowly to new technologies.74 Moreover, early photocomposition devices often suffered from teething troubles, exacerbating the scepticism of a largely traditionalist trade in face of the new technology. As Wallis pointed out, the Linofilm, and the later Linofilm Quick (Figure 3.19 overleaf) were “idiosyncratic” machines, and “it was not until the Linotron 505 crystallised in 1967 and the Linofilm VIP in 1970 that Linotype became a potent force in phototypesetting”.75 It was probably not by mere coincidence that explorations of the new technology for Arabic were only initiated around the same time: as noted before, technological change tended to be implemented later for Arabic than for Latin script typography. The reasons for this lag were varied, but appear to have been influenced by the manufacturers’ marketing policies. A letter from Linotype’s Middle East representative Ralph Goodman, enquiring with the UK office about future plans for Arabic photocomposition, reveals some of the dynamics influencing adaptation to the new technology:

sizes were needed to cover the whole of the machine’s size range […] for a single typeface.” Printer’s Type in the Twentieth Century, 89. 72

“Character width calculation is based on an 18-unit em. The minimum justifiable space is 4 units; units can be added for letter spacing or subtracted for kerning.” Phillips, Computer Peripherals and Typesetting, 498. Linotype had first introduced a unit system for composition with TTS equipment, and built on these principles in the Linofilm; see “The Beginning of ­Computer-Aided Arabic Composition” on page 183.

73

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 31–32.

74

As late as 1978, a reason cited against a shift to photocomposition was that in addition to the substantial initial investment, also running costs were higher than in hot-metal as photographic material was expensive and could not be reused. See Leslie G. Heath and Ian Faux, A Manual for Phototypesetting (London & Manchester: Lithographic Training Services Limited, 1978), 3.

75

Lawrence W. Wallis, “From Frame to Desktop: a Century of Typesetting,” in Typomania (Upton-upon-Severn: Severnside Printers Limited, 1993), 74.

232

Chapter 3

Figure 3.19

(Left) Linofilm Quick filmsetter. From Phillips, Computer Peripherals and Typesetting.

Figure 3.20

(Right) Comparison of boundaries of hot-metal characters to photocomposition characters, sketch by Walter Tracy, reduced to 40% linear. DTGC, From Walter Tracy to Michael Parker, Letter, “Arabic on Quick.”

We realise perfectly that, for the Middle East, we are not to push the Linofilm-Quick at this stage, preferring to recommend hot-metal, and that we are only offering the Linofilm Quick when it is necessary to try and hold sales which might be going to competitive equipment.76 76

Ralph Goodman to Edward Emery, Letter, “Linofilm Quick - Arabic,” 11 October 1967, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC. According to Gabeyan, the ‘pushing’ of hot-metal and Simplified Arabic was unscrupulous: “The product spread, because we were asked to sell more main-magazine machines, equipped with Simplified. As expected, newspapers were the first to adopt it, some printers followed and the tough ones gave in little by little, as there were less and less side-magazine machines manufactured, and if produced, the delivery time was not less than ten months; very discouraging. Towards the end I used to sell second-hand side-magazine machines, totally done-up at Altricham with new engraved serial numbers, quoted at the price of brand new ones. I was asked to keep it strictly confidential. The displeased, tolerant and to certain extent naïve readers got used to the Simplified day after day, the same way they are getting used to all the typefaces produced these days by charlatans.”

Photocomposition: Towards Immaterial Type

233

According to Goodman the high demand for fount grids pushed the company’s manufacturing capacity to its limits, by implication suggesting that no demand for Arabic grids could be fulfilled. Goodman’s assessment of the market emphasised the success of Monophoto machines, and underlined that enquiries for Linotype equipment were becoming increasingly serious, making the implementation of Arabic photocomposition an important development.77 It was not the first time the prospect of Arabic photocomposition had been raised, but it reinitiated the necessary investigations. Already in July 1967 Tracy had asked the Middle East Liaison Office (MELO) whether the new possibility of kerning should be discretionary for the operator or be the default behaviour in founts made for phototypesetting.78 Similarly, in apparent reaction to Goodman’s letter, Tracy also liaised with Mergenthaler who were in charge of the manufacture of the Linofilm to establish other features the new technology could offer for Arabic type.79 For example, Tracy realised that the absence of physical type allowed for much improved joining: characters created photographically could be made to overlap in print. This was particularly relevant for Simplified Arabic, as characters could now be made without the right-hand join, improving their appearance when used in initial position (Figure 3.20). The Linofilm Quick also had the ability to compose ‘floating accents’, therefore potentially facilitating the typesetting of vocalised texts. The adaptation and improvement of this feature to make it accessible for Arabic was seen as a crucial improvement, and Tracy requested Gabeyan’s help in assembling a comprehensive list of the required characters. Numerous exchanges between the UK office, the MELO and Gabeyan followed, yet little tangible progress was made. In January 1968 Goodman reiterated the urgency of the task, and expressed his frustration at falling behind the competition: As you know, long letters are flying around between Altrincham, London, Beirut and Cairo about what the Quick will be able to do for Arabic, which perforators can cope, how many accents we can offer, in what position. Hrant Gabeyan, email message to the author, “Miscellaneous”, 8 February 2016. 77

Goodman recognised some of the potential photocomposition could offer for Arabic typesetting, and suggested exploring Dr Edward Plooij’s ideas for a radically new approach to Arabic typography that would break with long held conventions. See also “En�deavours in Arabic Type by Dr Edward Plooij” on page 274.

78

Walter Tracy to Michael A. Nahas, Letter, “Arabic,” 19 July 1967, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC.

79

Walter Tracy to Michael Parker, Letter, “Arabic on Quick,” 27 October 1967, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC.

234

Chapter 3

While Monophoto are schooling and selling. They are now at least three years ahead in this particular field, and had resolved their new layout for Arabic Film Composition […] by 1964/5.80 The spread of the Linotype group across the globe, and the particular expertise of individuals who worked far removed from each other made this task particularly complicated. Broadly speaking, exchanges about every detail had to be maintained between four different locations: in Egypt, Gabeyan was the key figure for providing specific information about Arabic; the MELO in Lebanon delivered information about customers in the entire region; Mergenthaler in the United States manufactured the filmsetter and the film grids, and was most familiar with the technicalities of the machine; and in the United Kingdom Linotype made the character drawings for fount manufacture.81 But while the management of information across countries was difficult and time consuming, it may also have been beneficial. According to a study of the strategies typesetting manufacturers employed in the face of radical technological change, the geographical distribution of the Linotype group helped it to acquire and develop new competence, spurred innovation through competition, and allowed the integration of varied, location-specific know how.82 For Arabic typography, the MELO and the company representatives in the region were a crucial element in this skill-mix. Yet, the implementation of Arabic on the Linofilm Quick was further delayed by bottlenecks in the capacity of fount manufacture. In January 1968 Tracy tasked the Type Drawing Office (TDO) to produce “character image material […] as quickly as possible” and to supply it to Mergenthaler for the making of the Linofilm grids. He specified how the drawings for the photo­composition 80

Ralph Goodman to Arthur Henry Walker, Letter, “Linofilm Quick and Arabic,” 5 January 1968, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC.

81

From 1968 the German foundry D. Stempel AG, who had been manufacturing Linotype matrices from 1900, began the production of photocomposition founts for the Linotype group. Hans Reichardt, Chronik der Schriftgiesserei Stempel (Frankfurt am Main, 2002), 21, http://www.schriftenservice-d-stempel.de/ChronikStempel.pdf (accessed 16 May 2017).

82

“For instance, the strength of the printing press industry in Germany provided Mergenthaler [or rather the Linotype group] with strong ties, and helped the firm to understand how technological changes in printing presses might affect typesetters. At the same time, the strong history of high-quality typography in the United Kingdom provided Mergenthaler with insight into typefaces, and the US organization could tap into the strength of the software sector.” Mary Tripsas, “Surviving Radical Technology Change through Dynamic Capability: Evidence from the Typesetter Industry,” Industrial and Corporate Change 6, no. 2 (1997): 373.

Photocomposition: Towards Immaterial Type

235

fount were to differ from the hot-metal version, noting that “there are two requirements to be observed: the images to be supplied must conform with the Quick limitations, and the characters should be modified where necessary to get the benefit of the Quick advantages”.83 The machine’s limitations were primarily related to the introduction of a unit system as character drawings had to be made to fit within multiples of the base unit; its principal advantages were the ability to kern and the possibility to superimpose joining characters for a fused appearance.84 Yet, by March 1968 the work had not started, and it was not to start for another two years. Only by February 1970 Tracy reported to Gabeyan that “various necessary matters (one of them being the development of Ettela’at Simplified, see page 290 and following) had delayed the implementation of Arabic on the Linofilm Quick, but that progress was now being made.85 It was another nine months before Linotype had finished the drawings, sent them to the United States, and Mergenthaler had made the film grids to produce trial settings.86 Finally a showing of Arabic composed on a filmsetter was at hand, and in the words of Tracy it was “very unlikely that any other manufacturer can at present show specimens of film output as impressive as these.”87 In the meantime, in 1967 Linotype & Machinery merged with K.S. Paul Ltd to become Linotype-Paul Ltd.88 K.S. Paul had supported the Englishmen Peter 83

Walter Tracy to E. Beattie, Letter, “Simplified Arabic on Quick,” 23 January 1968, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC.

84

It was also investigated whether the ‘Accent-centering machine’ could provide satisfactory vowel placement, but no evidence of its use could be found. Michael Parker to Walter Tracy, Letter, “Linofilm Quick Centering Accents,” 4 June 1968, 1, Box N9940, NMAH.

85

Walter Tracy to Hrant Gabeyan, Letter, “Simplified Arabic on Linofilm Quick,” 5 February 1970, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC.

86

In a letter sent to Gabeyan with the proofs, Tracy explained: “unfortunately, when we sent the photographic copies of the characters to MLCo in February, it turned out that they contained a fundamental error and all of them had to be returned and re-worked.” Walter Tracy to Hrant Gabeyan, Letter, “Al-Ahram: Film Setting,” 11 December 1970, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC.

87

Ibid. Unfortunately the mentioned samples, or any other visual evidence was not found.

88

1967 is commonly cited for the merger but only from 1 October 1969 the company’s name changed. An internal communication to this effect stated: “As you know, our working relationship with Linotype and Machinery is becoming steadily closer, and cooperation between various departments of the two companies is building up very well. It is now intended to cement this relationship by changing our company name to ‘Linotype-Paul Limited’. This will take effect from October 1 of this year.” J. Parsons to Walter Tracy, Letter, 15 October 1969, WT correspondence, folder 16, DTGC.

236

Chapter 3

Figure 3.21

Linotron 505, showing from left to right the control unit with the paper tape bin, the photo unit, the tape merger and the keyboard unit. From Phillips, Computer Peripherals and Typesetting.

Purdy and Ronald McIntosh in developing the PM Filmsetter 1001, the device that was to be sold as the Linotron 505 and became “the first popular CRT machine” (Figure 3.21).89 The acquisition of K.S. Paul was strategically important for the Linotype group as it provided expertise in the upcoming CRT technology, and computerisation in general. According to Gabeyan this was manifest in the makeup of their staff where a considerable generation gap reflected the shift in the trade at large. Whereas K.S. Paul had young, active and energetic employees, Gabeyan describes Linotype staff as “archaic, often in the slow lane for most of their thought and reasoning process and decision-making, working in old, not to say ancestral, premises and workshops”.90 Over the next two decades Linotype-Paul Ltd became a key to the Linotype group’s technological advancement and general rejuvenation, an evolution of the company that was missed by some competitors.91 89

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 38.

90

Gabeyan, email message to the author, “Miscellaneous”.

91

According to Tripsas, Intertype and Monotype had also considered acquiring the company, but eventually decided against it. In an interview McIntosh recalled the reaction of Monotype’s president J. Matson to their device. When the two inventors approached him, he replied: “We see nothing that will equal the excellence of our [hot metal] product in the next 25 years”. Tripsas, “Surviving Radical Technology Change through Dynamic Capability: Evidence from the Typesetter Industry,” 367.

Photocomposition: Towards Immaterial Type

237

Towards an Arabic Photocomposition Programme Based on the successful adaptation of Simplified Arabic to the Linofilm Super Quick, Linotype planned its efforts towards the photocomposition of Arabic for the coming years around this device.92 However, continued developments of machinery soon led to a review of the company’s programme. In 1970 two new machines had been released by Mergenthaler: the Linofilm V-I-P (Figure 3.22) and the Linotron 505C. Both featured a programmable minicomputer, thus marking a decisive turning point in the capabilities of phototypesetters.93 Besides this shared feature, the two devices were built on fundamentally different principles. Whereas the V-I-P was an advanced second-generation phototypesetter, the 505C was a third-generation machine. It was based on the Linotron 505, the C denoting the built-in computer. Although the 505C still stored character images as photographic matrices, in contrast to secondgeneration machines it did not project them directly on to the photosensitive material. Instead, the scanned image was “[painted] on the face of a cathode ray tube, stroke by stroke”.94 The light of the strokes then passed through a

Figure 3.22 Linofilm V-I-P, left the photo unit, right the keyboard unit. From Phillips, Computer Peripherals and Typesetting. 92

Whilst most correspondence made no clear distinction between the Linofilm Quick and its successor, Simplified Arabic was implemented on the Linofilm Super Quick, released in 1968.

93

According to Wallis, the V-I-P (Variable Input Phototypesetter) “sounded the death knell for hard-wired front-ends in a phototypesetter”, as software now assumed control of the photo unit, as well as typographical composition functions beyond justification and hyphenation. “Dates with Phototypesetting,” in Typomania (Upton-upon-Severn: Severnside Printers Limited, 1993), 52; A Concise Chronology of Typesetting Developments 1886– 1986, 42.

94

Seybold, The World of Digital Typesetting, 112.

238

Chapter 3

Figure 3.23

(left) Linotron 505C character-generating system showing the 16 fount grids (3), and (right) writeout system showing the writeout CRT (1). From Phillips, Handbook of Computer Aided Composition.

lens and on to the photographic material, “exposing not a character at a time, but a stroke or slice of the character at a time”.95 (Figure 3.23) Devices such as the 505C achieved high output speeds, lending themselves particularly well to newspaper composition.96 Initially it appeared that right-to-left composition, a critical feature for Arabic, was not available on either device, but by August 1971, it emerged that the V‑I‑P would be capable of Arabic composition.97 At the same time a representative of al-Ahram visited Linotype-Paul and “made it quite clear that in his opinion Al Ahram would be far better [off] to invest in a 505C system than a Super-Quick system”.98 According to Walker, the client’s typesetting volume was 70,000 lines per day, besides newspaper composition, and thus justified an installation of the faster Linotron 505C. Development of Arabic for the Linofilm Quick was thus halted, and efforts concentrated on the two new devices. 95

Ibid.

96

In the 505 output speed depended on resolution. At its highest resolution of 1300 lines per inch, the 505 yielded 90 characters per second, which could be doubled at half the resolution. Phillips, Computer Peripherals and Typesetting, 549–50.

97

Tracy and Emery enquired with Mergenthaler whether the V-I-P would be an option for the Arabic photocomposition programme. The reply appears to have been negative, and only conversations between Linotype and Mergenthaler staff at the IPEX printing fair revealed that “there was no great problem in producing a right to left reading version of the V.I.P.” Edward Emery to John Heidenreich, Letter, “V.I.P. for Right to Left Reading Composition,” 10 August 1971, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC.

98

Arthur Henry Walker to Hrant Gabeyan, Letter, “Al-Ahram Phototypesetting,” 10 August 1971, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC.

Photocomposition: Towards Immaterial Type

Figure 3.24

239

Linofilm V-I-P A-range fount strip of Helvetica Condensed, reduced to 90% linear. From Phillips, Handbook of Computer Aided Composition.

The V-I-P employed founts in the form of film strips, each carrying 96 characters (Figure 3.24); although multiple fount strips could be employed to accommodate a single typeface, this had to be kept to a minimum as fount changes greatly slowed the machine. Sizing was effected through a combination of the size of the character image and the position of the zoom lens: “an ‘A’ range font could produce ten different sizes within the range of six to 24 points, and by mounting a ‘B’ range font the six point lens setting produced an 18-point image with the larger font master, and the 24-point setting produced a 72-point image”.99 Its character selection and escapement system involved rotating drums, a zoom lens, a rotating mirror and a curved output plane, and was among the more complex conceived for second generation machines (Figure 3.25 overleaf).100 The character limitations of the machine became a subject of repeated exchanges at Linotype. Because the setup of Arabic on the V-I-P was determined by the choice of fount layout, the general design approach to Arabic typefaces became a central issue. Broadly speaking, Simplified Arabic had introduced a binary distinction between ‘Simplified’ and ‘Traditional’ designs, the latter referring to designs that maintained larger character sets than those following the simplification principles. In internal Linotype correspondence, the question whether ‘Simplified’ or ‘Traditional’ was to be implemented on a device thus became the starting point of most new Arabic type projects.101 What 99

Seybold, The World of Digital Typesetting, 93.

100

Ibid., 89.

101

Ironically this distinction was still maintained much later when leading software com-

240

Chapter 3

Figure 3.25

Schematic diagram of the V-I-P’s optical system. From Phillips, Handbook of Computer Aided Composition.

emerged from these discussions could be described as two design philosophies, and they were most clearly embodied by Tracy, favouring ‘Simplified’, and Gabeyan who championed ‘Traditional’. Writing to Walker in late 1972 in relation to future developments of Arabic photocomposition, Tracy argued: In the Arab world the trend is likely to be more and more towards the use of Simplified Arabic and away from Traditional Arabic. It is much less complicated for the Linotype Group to manufacture for Simplified Arabic than it is for Traditional. I think it is in the interests of the Group for everyone concerned to encourage the use for Simplified Arabic wherever possible. Orders for ­Traditional Arabic should be accepted only because circumstances make it necessary: i.e. an existing user, a competitive situation, or resistance to change. Although that point of view applies to hot metal sales and manufacture, I think it is particularly important in regard to photocomposition.102 panies began to provide the first Arabic fonts for the operating systems of personal computers. See page 460 and following. 102

Walter Tracy to Arthur Henry Walker, Letter, “Arabic on Film,” 2 November 1972, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC.

Photocomposition: Towards Immaterial Type

241

Tracy, a key decision maker in the company, thus advocated a technologically compromised form of Arabic type design because of commercial expediency. Gabeyan, in contrast, repeatedly expressed the view that the emergence of photocomposition should be used to provide type conceived without the trade-offs that had been introduced for hot-metal machinery. He argued for founts with larger character sets, including an increased offer of ligatures and improved vocalisation placement. After numerous exchanges on the subject, in early 1973 Gabeyan renewed his call for a more ambitious Arabic type-making programme: Traditional Arabic on V-I-P is far more important than Simplified Arabic. None of our existing Arabic typefaces, L&M or MLCo., is good enough to compete with the Monotype design which at the present time is considered to be the best on the market. Therefore, a new Traditional Arabic typeface must be designed. […] Everything must be done, within our possibilities, to come out with the new Traditional Arabic on V-I-P before the end of this year.103 Motivated to bring back qualities which had been lost in previous technologies, Gabeyan’s stance increasingly contradicted Tracy’s (less idealistic) position. Correspondence on the subject reveals this struggle for the direction that Arabic type-making in the Linotype group should take, and is reflected in Gabeyan’s recollection: I couldn’t allow myself to yield an inch and I had to oppose them and fight back time after time doing my utmost to convince them, which wasn’t a small and an easy task.104 But although Tracy had reservations about the importance of ‘Traditional’ and sought to promote ‘Simplified’ in its stead, and despite the long delays incurred through repeated revisions of character sets, machines and priorities of the market, it appears that Gabeyan’s insistence helped to establish sound principles for Arabic type-making within the Linotype group. By late 1972, and “in anticipation of the need to put Arabics [sic] on to film”, the TDO established “a standard alignment for Arabic, taking into account the possibility of having 103

Hrant Gabeyan to Richard J. Caesar, Letter, 2 February 1973, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC.

104

Gabeyan, email message to the author, “Miscellaneous”.

Figure 3.26

“Arabic ‘Standard’ Alignment,” December 1972, reduced to 30% linear. DTGC, locker C1.

242 Chapter 3

Photocomposition: Towards Immaterial Type

243

to supply Traditional Arabic complete with diacritical signs and also taking into account Simplified and other forms of the Arabic script”.105 (Figure 3.26) Moreover, in the meantime Gabeyan had reviewed and assessed existing Arabic typefaces regarding their suitability for photocomposition. Not having found a satisfactory candidate, Gabeyan began to seek a craftsperson capable and available to produce the artwork for an entirely new typeface.106 Reporting to Tracy about his investigations in January 1973, Gabeyan suggested hiring Osman Hussein.107 Hussein had been recommended by the publishing house Dār al-Qalam, a prospective client for a Linotron 505C. Although he was not a professional calligrapher, he served as an officer in the Syrian army, Hussein had relevant credentials: according to Gabeyan, he had been a pupil of the renowned calligrapher Muḥammad Badawī ad-Dīrānī, and had written a Qurʾān at the request of the Syrian president a year earlier.108 Importantly, Hussein appears to have been knowledgeable in type design too: He follows the rules of type, he does not have his own fancy ideas about how to design a certain character. As far as I know he is the only Arab calligrapher who is well equipped and has even got a special big Russian projector for type. He writes something then he projects and applies the rules.109

105

Walter Tracy to Michael Parker, Letter, “Arabic on V I P,” 21 December 1972, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC.

106

Gabeyan had suggested licensing a typeface from the Beirut typefounder Georges Dib, but Linotype’s management was reluctant to pay the £3,000 that was asked for. In 2015 the relative income value of this sum would be approximately £70,840 – a high amount also by today’s standards. Lawrence H. Officer and Samuel H. Williamson, “Five Ways to Compute the Relative Value of a UK Pound Amount, 1270 to Present,” MeasuringWorth, 2016, http://www.measuringworth.com/ukcompare/.

107

ʿUṯmān Ḥusayn. Hrant Gabeyan to Walter Tracy, Letter, “The New Traditional Arabic Design,” 29 January 1973, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 121972, DTGC.

108

Muḥammad Badawī ad-Dīrānī, see also “8th International Calligraphy Competition Dedicated to Badawī Al-Diranī” (Research Centre for Islamic History, Art and Culture (IRCICA), 2009), http://www.ircica.org/content_images/8th-International-Calligraphy-Competition-ENG.pdf (accessed 16 May 2017). From 1972 to 2000 Hafez al-Assad was Syrian president.

109

Gabeyan to Tracy, “The New Traditional Arabic Design,” 1.

244

Chapter 3

Figure 3.27

Sample of Osman Hussein’s lettering, July 1973, reduced to 50% linear. DTGC, Hrant Gabeyan to Walter Tracy, Letter, “The New Traditional Arabic Design”.

Sample material Gabeyan attached to his letter illustrates what he might have meant by the “rules of type” (Figure 3.27). The lettering provided by Hussein appears regular and normalised, or typographic. Indeed, the uniform alignment of the vowels seems to be influenced by the vowel placement in movable type. Based on Gabeyan’s endorsement Hussein was asked to provide trial artwork for a new typeface that was to be used on both, the V-I-P and the 505C.110 The typeface design for the two devices, however, could not initially follow the same parameters. The V-I-P had limited kerning capabilities: all characters could kern to the left, and the characters located in row 12 of the fount strip could also kern to the right.111 However the 505C, although being the newer and technologically more advanced machine, had no such facility.112 Tracy thus suggested designing the 505C fount taking this limitation into account 110

Orders for both devices were obtained by December 1972.

111

Walter Tracy to Hrant Gabeyan, Letter, “Traditional Arabic on VIP,” 12 March 1973, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC. Alice Savoie points out that wide characters that used all 18 units could only kern by one unit because the exposure window of the V-I-P was limited to 19 units. Alice Savoie, “International cross-currents in typeface design: France, Britain and the USA in the phototypesetting era,” PhD thesis, University of Reading, 2014, 383.

112

“Because the scan of each character had to be completed before that of the next one could begin, the character images on the output in straightforward setting had to fit inside a series of contiguous rectangles […] This imposed a restriction on kerning which was very much like that placed on the original Linotype by the physical exigencies of the brass matrix.” Southall, Printer’s Type in the Twentieth Century, 135.

Photocomposition: Towards Immaterial Type

245

and instructing Hussein to draw the concerned characters accordingly. He argued that “the importance of kerning is often exaggerated”, and that “no great harm be done to the final appearance” if no kerning was possible in the fount.113 Yet Gabeyan insisted on a better solution.114 Reacting to Tracy’s suggestion, he acutely observed the technological step backwards this would entail: As you say, the importance of kerning is often exaggerated, but there is no doubt that when characters are kerned they look much better. We had this problem in hot metal composition and it seems now that it has come back to us with the 505.115 In the event, Gabeyan’s insistence proved worthwhile, for a workaround was found within a couple of months. The 505C was capable of placing accents above characters;116 it first exposed the base letters, and then effected a second ‘pass’ during which accents could be added. Tracy thus reasoned: If an accent can be put into position on a second ‘pass’, there seems no reason why we could not do the same with a particular letter, the ­machine leaving a space for it on its first reading and then putting the character into position on the second ‘pass’. If the character is one which ought to kern, such as ‘ra’, and the space which was left for it is slightly less than the total width of the character, then we will get the effect of

113

Walter Tracy to Hrant Gabeyan, Letter, “Arabic on 505C,” 30 January 1973, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC.

114

Gabeyan’s ambition was influenced by the demands of the client and the collaboration with Hussein. Writing to Tracy later in February 1973, he noted: “I must say that since November last year, when I met the DEQ [Dār al-Qalam] partners and then the calligrapher, I gained a lot of experience, because I have had the chance to devote more time to Arabic and also to study further the Monophoto possibilities on the spot, as well as discuss in length and in detail many aspects of the Arabic language and calligraphy with the calligrapher Osman, who, I must say, has a vast and deep knowledge specially of this language”. Hrant Gabeyan to Walter Tracy, Letter, “Traditional Arabic on V-I-P,” 22 February 1973, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC.

115

Hrant Gabeyan to Walter Tracy, Letter, “Arabic on 505C,” 7 February 1973, WT correspondence, folder 18J2 Trad Arabic VIP & 505 01-1973 - 12-1974, DTGC.

116

Arabic mark positioning on the V-I-P also employed a modified version of the machine’s standard ‘Accent Centering Routine’. Walter Tracy to Ian Houghton, Memorandum, “Arabic in VIP,” 2 May 1973, WT correspondence, folder 18J-2 Trad Arabic VIP & 505 1974-1977, DTGC.

246

Chapter 3

Figure 3.28

Artwork provided by Osman Hussein for the creation of photographic trials, n.d., reduced to 45% linear. DTGC, folder 127B.

kerning: that is to say, the tail of ‘ra’ will extend slightly under the following character.117 Although Tracy cautioned that one had to verify the feasibility of the idea, the principle proved to work and was implemented accordingly.118 Here a technical solution was found that prevented compromises in the design, as Arabic typesetting exigencies pushed the machine’s capabilities beyond its original design. In other instances the device still imposed its constraints on the typeface. The unit system in particular allowed no divergence from its principles: the minimum character width was four units and the widest eighteen units; no character could be narrower than an accent, and a ligature could not be nar117

Walter Tracy to Hrant Gabeyan, Letter, “Arabic on 505C,” 8 March 1973, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC.

118

In a letter to Richard Caesar, Tracy observed: “Hrant Gabeyan was so insistent that the appropriate characters should kern in the 505c output just as they do in the Monophoto that we have asked Jillian [Calderwood] to write a kerning routine into the programme.” Walter Tracy to Richard Caesar, Letter, “DAQ, 505c,” 30 April 1973, WT correspondence, folder 18J-2 Trad Arabic VIP & 505 01-1973 - 12-1974, DTGC.

Photocomposition: Towards Immaterial Type

Figure 3.29

247

Photographic reduction of trial characters (right) and comparison to Monotype Series 549 (left), reduced to 75% linear. DTGC, Walter Tracy to Hrant ­Gabeyan, Letter, “DAQ: 505c.

rower than the width of two accents.119 On both machines also the kerning functionality was basic. On the V-I-P characters which kerned to the left were “given a width value one unit less than face width”, and those kerning to the right had to be drawn offset, the width adjustment being made with the following character.120 Finally, by 2 May, Hussein had provided trial characters (Figure 3.28) and a first photographic proof was made at Linotype (Figure 3.29).121 It shows Hussein’s new design alongside Monotype’s Series 549, the typeface it was intended to compete with. Tracy conveyed to Gabeyan that it looked good, though he also noted issues of size and insufficient weight difference between the regular and the bold weights. Because of the pronounced height and depths of extenders, the type looked small when its overall size was matched with Monotype’s fount. The bold was not sufficiently heavy to distinguish itself from the regular weight and Tracy recommended that Hussein should review these issues.

119

Walter Tracy to Hrant Gabeyan, Letter, “D.A.Q.: 505c,” 27 April 1973, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC. Similar limitations applied to the implementation of Arabic founts on the V-I-P. Tracy to Houghton, “Arabic in VIP.”

120

Tracy to Houghton, “Arabic in VIP.” According to Ross on the 505C kerning could only be on or off, and fixed at a width of three units. Fiona Ross, email message to the author, “Arabic on Linotype Photocomp,” 17 March 2013.

121

Walter Tracy to Hrant Gabeyan, Letter, “DAQ: 505c,” 2 May 1973, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC.

248

Chapter 3

Figure 3.30

Comparison of the vertical space available on the Linotron with Monotype’s system, September 1973; (A) descender extended beyond the maximum size; (B) descender shortened to fit the scan area; (C) marks in Monotype’s design were smaller and ascenders relatively shorter; it appears that in the final fount a solution was found which did not require such compromised proportions; reduced to 50% linear. DTGC, Walter Tracy to Hrant Gabeyan, Letter, “DAQ: 505c.

Figure 3.31

Diagram comparing nominal point size to actual face size for Arabic typefaces on the Linotron 505C, September 1973, reduced to 50% linear. DTGC, folder 127B.

Photocomposition: Towards Immaterial Type

249

The development of the typeface remained complicated and advanced only gradually.122 By September, the above mentioned issues had not yet been addressed, and Tracy reiterated the importance of solving them prior to fount manufacture. Illustrated by a diagram, Tracy pointed to the vertical dimensions of the 505C, and the fact that all type sizes aligned on the same baseline. The machine’s vertical scan area was 21 units deep and extenders as well as vocalisation marks had to fit within this space. Therefore the relative size of characters and their proportions as defined by the designer were prone to being compromised during fount manufacture (Figure 3.30).123 Moreover, the combination of standard alignment with the face’s proportions and with the vertical dimensions of the scan area meant that the exposed character images were distinctly smaller than the nominal point sizes. On its 20 pt setting the actual exposed type size would be just under 16 pt, and its maximum 28 pt setting produced a printed size of around 20 pt (Figure 3.31). Only few weeks after Tracy’s comments would have arrived with Gabeyan in Beirut, the Fourth Arab-Israeli War broke out on 6 October, thus bringing further disruption to the project.124 Hussein was unavailable for weeks, and Tracy’s suggestions only reached him by November. Further corrections caused further delays, and the final artwork was eventually submitted in early 1974. Linotype’s drawing office then adapted the originals to the customary 12 inch master drawings, modifying characters where necessary to fit the 18-unit system. These drawings were then used to create negative character images, as required for fount manufacture: This was accomplished by placing a sheet of rubilith film over the outline drawing of the character. With the aid of a light table, the outline was 122

When Tracy sent his comments on 2 May skirmishes between militias of the Palestinian Liberation Organisation and the Lebanese army in Beirut resulted in a curfew and a state of emergency throughout Lebanon. The border to Syria was thus closed and Gabeyan was unable to get in contact with Hussein until July. Hrant Gabeyan to Walter Tracy, Letter, “Osman Hussein Arabic New Design,” 18 May 1973, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC.

123

The V-I-P had a larger vertical exposure area, see below.

124

Recalling this project much later, Tracy confused this conflict with the Six-Day War of 1967. Similarly, he credits Peter Pollock of the Linotype UK office for the choice of Hussein – an improbable scenario. The important role of Gabeyan shown here is diminished in Tracy’s account, in which Pollock allegedly arranged the calligrapher “through a contact there”. Walter Tracy, “Composing Room Days (and After),” Printing Historical Society Bulletin no. 40 Winter (1995): 14.

250

Chapter 3

Figure 3.32

Text sample of vocalised Arabic text composed in Osman on a Linotron 505C. From Tracy, “Arabic Without Tears”.

traced with a scalpel, the rubilith core was then peeled away revealing the character in negative; in this case white on red rather than black: the result was termed a frisket.125 As Ross pointed out in a recent publication, the use of registration aids in the form of punched holes on drawings and friskets and a transparent drawing mask helped to guarantee correct alignment, “a particularly essential requirement for joining scripts” such as Arabic.126 Manufacture of the founts was undertaken by D. Stempel AG, and proceeded quickly from then on. By late April 1974 Tracy expected the founts to be ready for use within weeks, and expressed his favourable view of the project in a letter to a colleague at Linotype: This order is a notable one. To the best of my knowledge, it is the first time that a CRT machine has been equipped to set Arabic. We have commissioned a new Arabic type design for the purpose, and independent expert opinion tells us that the type face is a remarkably good design. The computer programme is particularly an [sic] elegant one, taking care of

125

Ross, The Printed Bengali Character and Its Evolution, 192.

126

Fiona Ross, “The Type Design Process for Non-Latin Scripts,” in Non-Latin scripts, from metal to digital type (London: St Bride Library, 2012), 36.

Photocomposition: Towards Immaterial Type

251

character selection, justification process, and the automatic placement of vowel and diacritical signs.127 The development work for this first order was subsequently adapted to a number of setups. On the V-I-P, Traditional and Simplified Arabic were offered. The label Traditional stood for the new typeface, named Osman after its designer, which was paired with the new typesetting programme. The typeface featured all alphabetic characters in two weights, as well as vocalisation marks in two heights. However, in order to save character positions, vocalisation, superior figures, and punctuation signs of a single weight were shared between the regular and the bold weights. To accommodate all characters it required four fount strips; changes between the founts were automatic through the typesetting programme. In contrast, the Simplified Arabic package on offer with the machine required three fount strips as it contained few ligatures and no vocalisation marks. On the 505C only Traditional Arabic was offered initially. In its ‘luxury’ layout one weight required an entire fount grid, encompassing 256 characters (Figure 3.32). The typesetting programme had the same functionality as in the V-I-P, with a separate programme for Arabic and Latin mixing.128 The use of the computer for such wide-ranging tasks was considered a pioneering achievement by Tracy, so much so that he used its example in an article intended to highlight the benefits advances in photocomposition could offer the trade. Describing the tasks that were automated in the new machines, Tracy underlined that it was “the computer and its programs which have radically simplified the traditional processes of Arabic composition”. In contrast to his expressed stance in support of Simplified Arabic, here Tracy highlighted the large character set with a “considerable number of typographic refinement logotypes” as a benefit for the “Arabic language printer” that went beyond “the obvious advantages of speed and quality of output”.129 Although acknowledging the substantial resources Linotype had invested in the development of the system, including “typographic staff who are expert in the Arabic script”, “enthusiastic 127

Walter Tracy to P. Davies, Letter, “Arabic in 505C,” 26 April 1974, WT correspondence, folder 18J2, DTGC.

128

Walter Tracy to Michael Parker, Letter, “Arabic in Phototypesetting,” 4 February 1975, 2, Box N9902, NMAH. A dedicated layout developed for the Kayhan newspaper in Tehran contained no ligatures and could thus fit two weights on a single fount grid.

129

Walter Tracy, “Arabic Without Tears,” in The Penrose Graphic Arts International Annual, vol. 68 (London: Northwood, 1975), 122–124. Note that here Tracy refers to the simplification of the typesetting process, rather than the Arabic script.

252

Chapter 3

Figure 3.33 Sample of individual characters in regular and bold weights of the finished Osman typeface. From Tracy, “Arabic Without Tears”.

and talented computer programmers”, and “a member of the Linotype Middle East staff”, Gabeyan was not credited by name for his decisive design initiatives nor his development work.130 According to Tracy, the new typeface received “instant admiration from everyone” (Figure 3.33).131 Arguably, Linotype’s first system for Arabic photocomposition succeeded due to a number of mutually beneficial factors. By the mid-1970s the technology had reached a level of maturity that brought tangible advantages for the composition of Arabic; teething problems had been overcome, and the developers now knew how to fully take advantage of the new methods. But crucially, Linotype also benefitted from the experience it had gained in Arabic type-making over the past twenty years. The company’s work in this domain, 130

Ibid. Husseini [sic] was credited as the designer of Osman in the caption to an illustration of the article. According to Michael Fellows, the V-I-P software had been developed by Tom Sit. Michael Fellows and Fiona Ross, interview by Alice Savoie and the author, 16 February 2012, DTGC.

131

Tracy to Parker, “Arabic in Phototypesetting,” 1.

Photocomposition: Towards Immaterial Type

253

beginning with the cooperation with al-Hayat in the development of Simplified Arabic, had provided Linotype with the relevant expertise. Also, the company was able to build on prior successes; for example the typesetting programme that was developed for V-I-P and 505C was closely based on the pioneering work undertaken for al-Ahram’s computer-aided, hot-metal installation seven years earlier.132 The increasingly important roles played by the Middle East Liaison Office, and in particular by Gabeyan, as well as Tracy’s continued presence throughout this period, stand out as decisive for the sound principles on which Linotype’s Arabic photocomposition programme was founded.133 Arabic Type Developments by Linotype-Paul Ltd. The first half of the 1970s saw the most sustained increase of Arabic developments yet by Linotype-Paul Ltd. Multiple projects for clients in the Middle East were pursued in parallel, and a library of Arabic typefaces began to take shape. Existing designs were enhanced, and – as in the case of Yakout – adapted to diverse photocomposition machines, the software support for various devices was improved, and critically, additional typefaces were developed. Besides the commissioned type by Lakhdar Ghazal, (see also “Ahmed Lakhdar Ghazal and the Asv Codar Type” on page 90), Linotype-Paul also developed a new book face for quality typesetting. It was based on the same source as Monotype’s Series 549, the foundry type of the Egyptian Government Press.134 According to Tracy, “a close study of the Monotype version makes it clear that they must have worked from rather worn type; the terminals of some of the letters are rounded off”, and Linotype improved on this in Lotus.135 Because the dimensions of the original design required substantial vertical space, Linotype developed its fount in an unusual way. Initially produced exclusively 132

See “The Development of Character Selection and Kashida Insertion Routines” on page 187.

133

This exigency was not limited to Arabic, but constituted the beginning of an approach to non-Latin typeface design which for the first time aspired to the same quality standards as in the Latin script.

134

Linotype never publicly acknowledged this source of ‘inspiration’, but only referred to unspecific foundry type of allegedly Ottoman origin, a description that is maintained for the current digital version. Monotype GmbH, “Lotus® Arabic Font Family,” n.d., http://www.linotype.com/1183/LotusArabic-family.html (accessed 16 May 2017).

135

Walter Tracy to Michael Parker, Letter, “Typographic Development Programme: Arabic,” 21 June 1977, Box N9914, NMAH.

254

Chapter 3

for the V-I-P, it could take advantage of the machine’s large vertical exposure area. Characters were thus drawn to fill the entire em-square. This in turn meant that the film advance, i.e. the interlinear space, always had to be bigger than the type size; a 12D point size had to be set with at least 13D film advance, and if vowels were to be employed, this had to be increased to 17D.136 Beyond the technical aspects of new fount manufacture, the establishment of a library of Arabic typefaces also demanded a revision of its marketing. Whilst typefaces were still coupled with machinery, not requiring the promotional effort of a standalone product, more choices called for clearer distinctions between different designs.137 Therefore, and following up on his earlier naming of Yakout and others (see page 129), Gabeyan picked fitting labels for the new typefaces. The design that was based on the Egyptian Government Press type was thus called Lotus, inspired by the Lotus flowers in the pool at the main entrance of the Museum of Egyptian Antiquities in Cairo, and Osman Hussein’s typeface was renamed to Badr – ‘full moon’ in Arabic. Besides the positive cultural associations – the full moon is a symbol of beauty in the Arabic language – here Gabeyan made a subtle reference to his own involvement in the type’s development: Badr al-Ǧamālī (1015–1094) was a renowned Mamluk vizier of Armenian origin. He liberated Cairo from the Seljuq occupation and left several important buildings in the vicinity of Gabeyan’s birthplace in Cairo’s old town. According to Gabeyan, this reference to their shared Armenian ancestry remained a personal note and was never made public.138 The enlargement of Linotype-Paul’s Arabic programme coincided with the final years of Tracy’s employment. Indeed, during this period his role in Arabic developments became his primary responsibility – he estimated that in the years from 1972 to 1975 at least 80% of his time was spent “organising the typographic side of the Company’s Middle East trade”.139 The continued involvement and expertise developed over the preceding 20 years appears to

136

Ibid. Marketing material referred to this typeface as Lotus ‘Full-Face’ to emphasise this technical characteristic. A later version of the typeface for the Linotron 303 had reduced descenders and the qualifier was dropped from the name.

137

A memorandum by Tracy emphasised this point: “Now that we have a choice of faces available, it is no longer sufficient to specify ‘Traditional’. The typeface name should be used in all discussions and documents.” Walter Tracy to A. E. Genower, Letter, “Traditional Arabic Typefaces,” 16 August 1976, Box N9902, NMAH.

138

Hrant Gabeyan, email message to the author, “Thesis”, 31 July 2015.

139

Walter Tracy to Michael Parker, Letter, 2 December 1975, Box N9940, NMAH.

Photocomposition: Towards Immaterial Type

255

have paid-off, as by the end of 1975 Tracy assessed that V-I-P and 303 sales amounted to two-and-a-quarter million pounds, and were still increasing.140 Tracy’s forthcoming retirement called for the recruitment of qualified staff that could fulfil his role, specifically in regard of the Arabic programme. To this end, in 1974 Timothy Holloway was hired and trained to be Tracy’s successor.141 Within the first year of his employment, Holloway left a lasting mark on Linotype’s Arabic programme, designing the new Mitra typeface which was to become one of the most popular designs in Iran (see page 306 and following). However, and to the disappointment of the management, Holloway remained with Linotype-Paul for only 18 months.142 Apparently to pursue an independent career, Holloway left the company in March 1976.143 At this time, and probably related to the unsuccessful engagement of Holloway, Tracy wrote that he intended “to hire a young design graduate [whom he] met some time ago”.144 The graduate, who according to Tracy had studied some aspects of Arabic design, was probably Margaret Tan and was hired later in the year. In July 1976, Tracy wrote in a letter to Walter Greisner of D. Stempel AG that she had recently joined the company “as a specialist in Arabic typography”; with the letter, Tracy conveyed a separate commentary written by Tan on the design proposals by Roberto Hamm.145 Although her report could not be found, Tracy’s assessment is likely to echo Tan’s words: 140

Ibid. In 2016, the relative value of this amount would be approximately £ 38,390,000. Lawrence H. Officer and Samuel H. Williamson, “Five Ways to Compute the Relative Value of a UK Pound Amount, 1270 to Present,” MeasuringWorth, 2016, http://www.measuringworth.com/ukcompare/.

141

The decision to hire Holloway was taken by Mergenthaler’s designer Matthew Carter, Michael Parker and Tracy. Parker, expressing his agreement, stated: “He’s very young – but seems most interested in the entire business of letterdesign. While his work was not astounding – I thought it showed an even level and a certain professional approach. He seems to be sincerely interested – and I judge him a good risk for training.” Michael Parker to Walter Tracy and Matthew Carter, Letter, 6 February 1974, Box N9940, NMAH.

142

Fiona Ross, “Non-Latin Type Design at Linotype,” 2002, http://tntypography.eu/resources-list/non-latin-type-design-linotype-fiona-ross/. (accessed 16 May 2017). Tracy grudgingly remarked: “He was hired to do my job. He’s taken a year to decide he doesn’t want it. No apology, or thanks for all I put into him. He’ll have to be a very good designer …” Walter Tracy to Michael Parker, Letter, 8 October 1975, Box N9940, NMAH. Holloway successfully pursued his independent career, developing numerous renowned and awardwinning typefaces.

143

Walter Tracy to Michael Parker, Letter, 5 March 1976, Box N9940, NMAH.

144

Ibid.

145

Walter Tracy to Walter Greisner, Letter, “Arabic Design by R. Hamm,” 22 July 1976, WT cor-

256

Chapter 3

He has made the mistake of converting letter forms which are essentially calligraphic into rigid geometric forms, with loss of readability (not gain, as he thinks). And to put vowel signs after letters instead of over them means fewer characters per line.146 In September 1976, Tracy drafted an internal report on the state of LinotypePaul’s Arabic programme, reflecting the company’s “expanding trade in the Middle East”.147 Probably aimed at informing potential successors, the document outlined the type catalogue, the devices on which the founts could be used, and which software packages and language support were available. At this point five Arabic typefaces were available for photocomposition: Yakout, Badr, Mitra, Lotus and Lakhdar Ghazal. Various fount setups existed on a range of machines: the Linotron 303 and 505, the Linofilm V-I-P and the Linocomp, a small direct-entry phototypesetter.148 In 1976 the Linotype group also upgraded its V-I-P machine, switching from the old 18-unit system to the more flexible 54-unit system. Branded as the Advanced Typography Program (ATP), it allowed for more subtle spacing and kerning of founts, automatic ligature substitutions, hanging punctuation and refined justification in what became a “benchmark for imitations by many competitive suppliers”.149 Among the designs listed as future Arabic developments was “an alternative to Yakout”.150 The idea to provide a simplified typeface of a different design had respondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC. In the 1970s, the Frenchman Roberto Hamm developed a constructed Arabic typeface for the Algerian oil company Sonatrach. He subsequently published about his approach and suggestions for Arabic type design, but his work remained a one-off effort with no noticeable influence on the trade. See Roberto Hamm, Pour une typographie arabe: contribution technique à la démocratisation de la culture arabe (Paris: Éditions Sindbad, 1975); Roberto Hamm, “De la calligraphie à la typographie,” Communication et langages no. 46 (2ème trimestre 1980): 34–49. 146

Ibid.

147

Walter Tracy, Arabic. Linotype-Paul Type Designs, Fonts, Programs, Keyboards, 1 September 1976, 1, Locker C1, DTGC.

148

Introduced in 1973, the Linocomp was an inexpensive device with limited typographical capabilities. Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 48.

149

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 54. Then the V-I-P allowed a selection of characters to kern by the fixed amount of 3 units irrespective of the following character, which remained a significant limitation. Interview with Michael Fellows and Fiona by Alice Savoie and the author, 16 February 2012, DTGC.

150

Tracy, Arabic. Linotype-Paul Type Designs, Fonts, Programs, Keyboards, 8. The other pro-

Photocomposition: Towards Immaterial Type

Figure 3.34

257

Drawing for ‫ ط‬ṭāʾ, Jalal, superimposed over matching 18-unit mask, n.d, reduced to 50%. DTGC, drawings cabinet.

been considered occasionally by Linotype, and it was soon to be implemented by Tan. The design process for the resulting typeface Jalal is not documented, but in a second, revised edition of Tracy’s report about the company’s Arabic programme it features in the category of simplified faces.151 Here, Jalal was described as “strong and clear in small sizes”, and as being “designed to take advantage of [the] V-I-P optical system”. The fount was thus ‘Full-Face’ like Lotus, and not available on the 303 without design changes.152 Although following the simplification principles of Yakout, Jalal is stylistically independent jects were a new Persian text face, a new attempt at a Nastaʿlīq fount which was to be developed by the Iranian Ministry of Culture, and a bold titling face for al-Ahram. 151

Walter Tracy, Arabic and Farsi, and Other Languages Using the Arabic Script. Type Designs, Grids and Fonts, Programs. Second Edition., April 1978, DTGC.

152

Ibid., 9.

258

Chapter 3

Figure 3.35

‫ف‬ From left to right: drawings of characters ḥāʾ, ‫ �ه‬hāʾ, and the rasm of a �� fāʾ ‫ح‬ and �‫ �ي‬yāʾ ligature, n.d, reduced to 35%. DTGC, drawings cabinet.

(Figure 3.34 & Figure 3.35 overleaf). Full curves and subtle stroke modulations along the connecting stroke, as well as rounded terminals give the typeface a distinct look. Despite this successful start, Tan’s career with Linotype was brief – she left the company in 1977. Later in the year, Tracy wrote to Tan to commission an extra-bold version of the typeface, recognising that “the ideas on which it is formed are quite effective”; for this work he offered £  450 and no royalties.153 Tan refused this offer, noting that her earnings had to fund her PhD research “in Arabic Type Design”, which she intended to pursue in London. In turn, she proposed a higher flat fee of £  18 per drawing and an additional 10% in royalties.154 It appears that Linotype did not agree to this payment, and no other correspondence with Tan is documented. Linotype-Paul was initially unsuccessful in recruiting a successor for Tracy with similar capabilities. Temporarily a certain Anthony Bisley was hired for the post, but issues with his leadership soon led to another change of management.155 153

Walter Tracy to Margaret Tan, Letter, “Jalal Extra Bold,” 24 October 1977, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC. In 2016 the relative income value of this fee would be approximately £ 4,802. Lawrence H. Officer and Samuel H. Williamson, “Five Ways to Compute the Relative Value of a UK Pound Amount, 1270 to Present,” MeasuringWorth, 2016, http://www.measuringworth.com/ukcompare/.

154

Margaret Tan to Walter Tracy, Letter, “Re: Jalal Extra Bold,” 11 December 1977, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC. Later Tan published a titling typeface called Tahir. It was implemented as Letraset dry-transfer lettering and on D. Stempel AG’s Typomatic device.

155

Tracy recalled, without mentioning a name, that “he asked no questions, and preferred

Photocomposition: Towards Immaterial Type

259

In 1978 Linotype engaged Fiona Ross as research assistant. Ross, born in 1954, had grown up in an international household: her mother was Hungarian and her father half Russian, half Scottish, and following the Hungarian Revolution of 1956 her Hungarian grandparents joined the family, contributing to a rich language mix spoken at their home. Because of a keen interest in literature, which she read wherever possible in the original language, Ross applied for the course in German literature studies at the University of Kent at the age of 16. In preparation for the oral exam, Ross spent a few months in Vienna, where a serendipitous encounter made her discover Sanskrit literature, planting the seed for her later career. Ross started her degree in German literature at Kent in 1970. Three years later she had the opportunity to spend a year abroad and went to Tübingen in Germany order to pursue Sanskrit studies. Upon her return to the UK, and having completed her first degree in 1974, Ross continued her Sanskrit studies at the School of Oriental and African Studies (SOAS) at the University of London from 1975. In 1978, and whilst preparing for the final exams in her second degree, Ross was hired by Linotype-Paul.156 As research assistant Ross’ first task was to work on a system named “High Speed Arabic”. Its objective was to improve the layout of Arabic V-I-P founts for higher composition speeds. To achieve this the drum movements for character selection, the slowest aspect of composition, had to be reduced. Thus, Ross rearranged the layout of the four film strips “so that the most frequent characters occurred in the central positions of the first fonts”.157 Following the departure of Bisley in 1979, Ross led the Typographic Department, at this time consisting of only one additional staff. At the same time the responsibilities of the department greatly increased, and as part of wider company restructuring it relocated to Chelham House in Cheltenham, which became Linotype-Paul Ltd’s main office. Under Ross’ tenure (from 1985 as the first female manager of the company), the department’s size increased too; not only were new staff employed, but a new recruitment policy was applied. Now, “graduates in fine arts, rather than apprentice letter-drawers, were employed in the design studio and other graduates were employed for research and development work”.158 Amongst these new members of staff were more women than previously the to take his own course. Not long after my leaving I heard he had been dismissed in a flurry of recrimination.” Tracy, “Composing Room Days (and After),” 14. 156

Fiona Ross, interview by the author, 13 April 2013. From 1981 and part-funded by her employer, Ross pursued a PhD in Indian Palaeography in parallel to her work. She was awarded the PhD from SOAS in 1988.

157

Ross, “Non-Latin Type Design at Linotype.”

158

Ibid.

260

Chapter 3

Figure 3.36

The Department of Typographic Development, letter-drawing studio and research staff, 1983. From left to right: Georgina Surman, Lesley Sewell, Sarah Morley, Gillian Robertson, Ros Coates, Fiona Ross, and Donna Yandle. Courtesy of Fiona Ross.

case, notably Georgina Surman as the head of the Design Studio and Ros Coates and Gillian Barrett who, together with Ross, formed the Typographic Research and Development (R & D) section (Figure 3.36). At the same time Tracy retained a loose affiliation with the company after his retirement. Weekly visits to the office provided a degree of continuity, and Ross continued to consult Tracy for some projects – for example Qadi, an Arabic newspaper titling face designed by Tracy was developed during the mid-1980s.159 Although a generational change happened around 1980 and the structure and policies of Linotype-Paul changed significantly, the company was able to carry over some of the assets that had been previously established.160 A wide client base, experience with Arabic developments in a range of coun159

“Qadi” (Linotype-Hell AG, n.d.), DTGC.

160

Hrant Gabeyan left the Middle East Liaison Office in 1976. However, another Armenian employee, Dikram Panossian, had been recruited earlier by Gabeyan, effectively replacing him after his departure. In the late 1970s and 1980s, Panossian assumed a similar role to Gabeyan and developed a close working relation with Linotype’s UK office.

Photocomposition: Towards Immaterial Type

261

tries and contexts, and an approach that combined pragmatic business considerations with an appreciation of technological constraints and opportunities made Linotype well prepared for the continued challenges of the trade in the decade to come. A Widening Field, the Emergence of New Developers The spread of photocomposition technology brought about a radical change in the market. Whereas the three dominant equipment manufacturers Linotype, Monotype and Intertype had virtually held a tri-partite monopoly for the first half of the twentieth century, the advent of photocomposition provided a chance for new companies to compete for a share in the market. In lockstep with the advance and success of photocomposition and lithographic printing, computing played an increasingly important role in the printing trade. As the products changed, the industry transformed too; mechanical engineers were gradually pushed aside by electronics and computer specialists; plastic and optics replaced steel and lead, in what has been described as “radical, competence destroying technological change”.161 On the forefront of the upheaval caused by photocomposition was the Lumitype-Photon machine, which had revolutionised the principles of typographical composition when it first entered the market in 1946. By the early 1970s, the company began to expand its programme to cover Arabic. Arabic Type by the Photon Company As noted above, the Lumitype-Photon machine originated from collaboration between the French engineers Higonnet and Moyroud with the US Lithomat Corporation.162 In 1946 the parties signed a development contract that provided the inventors with funds to build a prototype while granting Lithomat the rights to manufacture and distribute the machine in the Americas; Higonnet and Moyroud kept the right to exploit their invention in the rest of the world.163 161

Tripsas, “Surviving Radical Technology Change through Dynamic Capability: Evidence from the Typesetter Industry,” 342.

162

William W. IV Garth, Entrepreneur. A Biography of William W. Garth, Jr., And the Early History of Photocomposition (Self-published, 2002), 31.

163

This initial agreement between the inventors and Lithomat was later continued with Photon; from 1952, a partnership called Photrop was formed between Moyroud, Hig-

262

Chapter 3

By 1948 a working prototype was completed and imported to the United States for public demonstrations, but research and further developments continued for another eight years before a commercial product became available.164 The technology of the Lumitype-Photon machine was marked by the backgrounds of its inventors. Neither Higonnet nor Moyroud were familiar with the printing and typesetting trade, and according to Southall they had “no preconceptions about what type-composing machinery ought to be like”.165 Indubitably, this unbiased perspective helped to develop radically new concepts for the composition of type. In the Photon 200, character images were stored on a matrix disc that rotated at a continuous 600 rpm (Figure 3.37). One disc held 16 founts of 90 characters each, thus providing a total of 1440 characters. Character selection was effectuated through synchronising an electronic flash tube with the character position on the disc at an operating speed of ten characters per second; character images were thus projected on to the photographic material through one of twelve lenses and an electro-mechanical escapement.166 Crucially, this system allowed sizes and founts to be mixed along a common baseline, irrespective of the style of the type, according to Southall “something that had not previously been possible with mechanical composition”(Figure 3.38).167 However, initially the economic fortunes of Photon did not match the promise of the new technology. Between 1957 and 1959 the company failed to achieve its sales projections, and by 1960 it was running short of capital. In consequence the board of directors asked the company’s president, co-found-

onnet and the Photon investors William Danforth, Bernard Smith and William Garth. In 1957, the partnership was incorporated, changing its name to International Photon Corporation (IPC). Marshall, Du plomb à la lumière, 194–224. 164

From 1949, the newly established Graphic Arts Research Foundation funded the research and development work; in 1952 Lithomat changed its name to Photon. Field trials of the Photon 100 machine began in 1954, and by 1956 the first mainstream product was commercialised as the Photon 200. Garth, Entrepreneur; Wallis, A Concise Chronology of Typesetting Developments 1886–1986.

165

Southall, Printer’s Type in the Twentieth Century, 81. Moyroud was trained as a mechanical engineer, Higonnet had worked in the telecommunications industry, and both men shared a keen interest in photography.

166

Phillips, Computer Peripherals and Typesetting, 515–16.

167

Southall, Printer’s Type in the Twentieth Century, 81. Christopher Burke points out that the German typefounding industry had tried something similar with a standardised baseline, introduced in 1905. This concept was, however, limited to foundry type.

Photocomposition: Towards Immaterial Type

263

Figure 3.37 Photon-Lumitype 200-500 fount matrix disc. From Phillips, Computer Peripherals and Typesetting, 516.

Figure 3.38

Mixing of styles and sizes on the Photon-Lumitype, actual size. From Southall, Printer’s Type in the Twentieth Century, 83.

er and biggest shareholder William Garth to resign.168 During the first half of the 1960s Photon solidified its position in the market, eventually becoming the world’s biggest photocomposition equipment manufacturer. It maintained this position until the early 1970s, but the second half of the 1960s witnessed the gradual demise of the company’s standing.169 Competition became increasingly fierce throughout the decade, as exemplified by the rise of the Compugraphic Corporation which began to manufacture phototypesetters at an “unprecedentedly and unimaginably low” price in 1968.170 Photon’s involvement with Arabic type only commenced towards the end of its independent existence as a company, and it was pursued by the ­International 168

Garth, Entrepreneur, 87. Later in the year Garth and Ellis Hanson, a former engineer of Photon, founded the new company Compugraphic. It was to become a revolutionary force in the trade.

169

Ibid., 307.

170

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 39.

264

Chapter 3

Photon Corporation.171 Although documentation of the Arabic developments of the company is not exhaustive, it appears that the IPC began to investigate the feasibility of Arabic composition on its machines in the early 1970s.172 Writing to Claude Lacoche, the vice-president of the IPC, Ladislas Mandel, art director of Photon’s Paris type design studio, noted that the best way of approaching the task would be to gather information directly from printers and publishers in the region; yet, as the Paris office had “no finances provided to this effect”, it would “try to obtain information [it needed] by correspondence where [its] relationships allow[ed] this”.173 Mandel further noted that the most popular typefaces seemed to be those used by the Egyptian al-Ahram newspaper (without apparent knowledge of their origins) and mentioned the simplified typeface Ghossoub by the French type foundry Deberny & Peignot, and the system developed by Lakhdar Ghazal.174 In conclusion, Mandel mentioned that IPC’s Paris office had the advantage that a member of its design staff, Miss Joyce Ini, spoke Arabic and had studied Arabic calligraphy.175 In early 1972 reports about the Arabic printing trade were submitted by Photon’s contacts. Writing to Lacoche, Walter O’Brien the head of an Argentinian business consultancy, outlined his assessment of the market for Arabic composition machinery.176 His report summarised his understanding of the requirements of Arabic typesetting, listing typical character sets, the role of vocalisation, ligatures and different styles. The report is dotted with inaccuracies, misconceptions and errors, suggesting a superficial grasp of the subject.177 171

Although nominally a separate entity, in effect the IPC was part of the US Photon Corporation which it had bought in 1967.

172

From a letter by Edward Emery of Linotype, it emerges that by August 1971 the company was “actively researching Asia and Africa to determine which Non-Roman market makes sense for the new small Photon.” Emery to Heidenreich, “V.I.P. for Right to Left Reading Composition,” 2.

173

Ladislas Mandel to Claude Lacoche, Letter, “RE: Long Range Project - Indian & Arabic Typography,” November 18, 1971, 1, Fonds Mandel, folder Dessins Lakhdar Naskhi, MI.

174

At this time al-Ahram was set in Linotype’s Yakout, composed on TTS-tape driven Elektron casters (see also “The Development of Character Selection and Kashida Insertion Routines” on page 187).

175

Mandel noted that Ghossoub “was not a great success” and that “Lakhdar seems to be regarded more favourably”. Mandel to Lacoche, “RE: Long Range Project - Indian & Arabic Typography,” 3.

176

Walter O’Brien to Claude Lacoche, Letter, 3 March 1972, 1, Fonds Mandel, box Études Arabes, MI.

177

O’Brien noted that classical Arabic (without defining this term) “uses the complete set

Photocomposition: Towards Immaterial Type

265

Nonetheless, in conclusion O’Brien found “that Arabic can be organized for setting with Photon”, and thus recommended proceeding further.178 Subsequent developments are poorly documented: only by February 1973 Paul Kauffmann of Deberny & Peignot reported to Lacoche from a visit to Tunisia: according to him there was unanimous agreement among printers in the country that the two styles ‘Nasri’ and ‘Nessih’ were the preferred and most used in Arabic typography.179 Whereas the latter may be a transliteration of the Turkish word for Nasḫ, the origin and meaning of the former is unknown. The attached newspaper sample, intended to show the ‘Nasri’ style reveals, however, that it referred to Linotype’s Yakout, correctly described as the “one [which] is used by all the newspapers” (Figure 3.39 overleaf).180 Kauffmann concluded that Photon would be well advised to develop the two styles for its machines.181 Based on this assessment Lacoche tasked Mandel to investigate the two styles with the goal to offer Arabic by the end of the year. The confusion of terms and the apparent ignorance of established typefaces in the market suggest that Photon was poorly prepared to venture into Arabic type-making; and as documentation is sparse, one may only speculate about the motivations for this late and rushed début. The fact that by this time Photon was having severe difficulties in the US and European markets may give one clue for its attempt to establish a presence in the Middle East; from 1971 Compugraphic had overtaken Photon as the world leader in phototypesetter sales, and financial woes were troubling the company.182 of images (approximately 4 images to each character), that imply [sic] 100 different characters, plus two more images very commonly used”, and that it required “something around 15” ligatures. While lacking clarity in terms, the numbers O’Brien suggests appear implausibly low for quality book typography, the potentially intended meaning of “classical Arabic”. Statements that underline the insufficient grasp of the subject include “the Arabic alphabet is uniform in all Arabic countries”, a crude dismissal of diverse regional styles, and O’Brien’s misleading description of Simplified Arabic: “character-wise, all 102 images are used; only a few are dispensed with because of their great similarity (where only one or two images per character are used is in typewritten texts).” Ibid., 1–2. 178

Ibid., 4.

179

Paul Kauffmann to Claude Lacoche, Letter, 8 February 1973, 1, Fonds Mandel, folder Nasriphot 55-65, MI.

180

Ibid. No samples of what was called “Nessih” are extant, but it was described as the standard book face.

181

Ibid., 2. He was more cautious about Lakhdar Ghazal’s type, noting that it had not been proven to be acceptable.

182

Marshall, Du plomb à la lumière, 334.

266

Figure 3.39

Chapter 3

Detail from the Tunisian newspaper as-Sabah (Aṣ-Ṣabāḥ), 30 January 1973, actual size. Columns of text marked “Nasri 65” and “Nasri 55” show Linotype Yakout in regular and bold weights, reduced to 70% linear. MI, folder Nasriphot 55-65. Courtesy of Musée de l’imprimerie et de la communication graphique (All rights reserved).

By May 1973 proofs of Photon’s first Arabic typeface Nasriphot were printed. It appears to have been an original typeface, though its design principles loosely followed those of Simplified Arabic (Figure 3.40 overleaf). A particularity of the type was its fount synopsis. On a single disc, three layouts were paired to corresponding keyboards which catered to various typesetting requirements: (1) a simplified keyboard, (2) a newspaper keyboard, and (3) a so-called classical keyboard. The first held 90 characters and was intended for “complex technical setting as in maths”. Its rationale was to employ few alphabetic characters to allow for the inclusion of additional special signs. The newspaper keyboard contained 112 characters, including 20 vowels and three ligatures, and had an ambitious goal. The documentation reasoned that because newspaper keyboards hitherto did not feature vowels, “reading [was] more difficult and limited newspaper diffusion.” Consequently, it was argued, the “improve-

Photocomposition: Towards Immaterial Type

Figure 3.40

267

Nasriphot 201.55, proof, IPC, May 1973, enlarged to 150% linear. MI, folder Nasriphot 55-65. Courtesy of Musée de l’imprimerie et de la communication graphique (All rights reserved).

ment” to provide vowels was “liable to enlarge newspaper circulation.” The third keyboard also contained 112 characters, yet offered the four conventional forms per character and was thus thought to be “more suitable for composition of all classical, literary, or religious text”.183 It appears thus that Photon’s Arabic development was led by assumptions about the requirements of the market and the most suitable typographical solutions, rather than informed by substantial research. None of the above suggestions would have held up against a thorough investigation of established practice in the trade. Similarly, the resulting typeface Nasriphot was not on a par with designs by the competition: marred by atypical proportions, inconsistent stylistic treatment and generally unsuccessful character drawings, the typeface bordered on the naïve (Figure 3.41 overleaf). Photon’s inexperience with non-Latin developments is also apparent in the choice of the designer. Joyce Ini was tasked to design Nasriphot, a decision that was probably influenced by her Arabic language skills.184 183

Anonymous, Arabic Typography, n.d., 1, Fonds Mandel, folder Projets Arabe Photon, MI.

184

In July 1973 Lacoche thanked Ini explicitly for the “dedication and determination […] shown in creating the new Nasriphot”, enclosing a cheque of $ 100 to express the company’s gratitude. Claude Lacoche to Joyce Ini, Letter, 31 July 1973, Fonds Mandel, folder

268

Chapter 3

Figure 3.41

Large-size lettering for Nasriphot, IPC, n.d., reduced to 50% linear. MI, folder Projets Arabe Photon. Courtesy of Musée de l’imprimerie et de la communication graphique (All rights reserved).

Here, the implicit assumption that language proficiency was synonymous with script-specific design excellence was not confirmed in practice, as Nasriphot must be considered a largely unsuccessful typeface. By late July 1973 company representatives visited Beirut to promote their product, and by September also a bold version with similar characteristics was made available.185 Yet, evidence about the project only continues in February 1975. In an internal report Mandel summarised Photon’s Arabic programme, and outlined its current state and future plans. According to this report, in addition to Nasriphot, two weights of Lakhdar Ghazal’s type had been manufactured and were “to be used in [the] Maghreb in the near future.”186 Regarding future developments, Mandel listed an “Enriched Nasriphot” without vowels and more “calligraphic variants, extending somewhat the scope of this

Nasriphot 55-65, MI. 185

According to Gabeyan, the type was provided on a Pacesetter and Photon quoted 300,000 French Francs for the system. Hrant Gabeyan to Richard J. Caesar, Letter, “Compugraphic & Photon,” 6 August 1973, WT correspondence, folder 18J Trad Arabic VIP & 505 07-1969 - 12-1972, DTGC.

186

Ladislas Mandel, Photon’s Arabic Typography (Paris, 4 February 1975), 1, Fonds Mandel, folder Projets Arabe Photon, MI, Paris.

Photocomposition: Towards Immaterial Type

269

a­ lphabet”, as well as a “classical Arabic alphabet” that was in the process of being designed in Paris. In conclusion, Mandel noted: Considering the range of such a market, I feel that serious investigations would be worth making BEFORE effecting the second stage of our program. To do this we should, with our trial proofs in hand, consult Lebanese, Egyptian, as well as Magrabin [sic] publishers and printers. Thus avoiding the errors made for the first version of NASRIPHOT.187 In a later letter, Mandel unambiguously described the failure of Nasriphot, noting that it “was considered inacceptable and categorically rejected” when shown in Beirut.188 He emphasised the need for feedback from potential customers, as gathered during a visit to the region. According to Mandel, the criticisms helped to revise the design, making it “the best simplified Arabic alphabet on the market.”189 The veracity of this claim remains, however, disputable: no evidence of any actual use was found for this study, suggesting that Nasriphot was not accepted by clients in the Middle East.190 Subsequently, the IPC began to develop another Arabic typeface, yet this time with the specific ambition to compete with the Monophoto for a share of the quality book typesetting market. The fount was planned to feature 448 characters and its design was “closely inspired by the Monotype face in use in Lebanon”; Mandel, however, qualified this, arguing in favour of a “distinctive but no less aesthetic style, something between Monotype and D.P. [Deberny & Peignot] Arabic Fount NESSIH.”191 In the meantime Photon’s and its affiliates’ economic difficulties became more severe. In July 1974, Deberny & Peignot was declared bankrupt and went into administration; by February 1975 also Photon officially declared ­bankruptcy.192 The following month Photon was bought by Dymo Graphics Systems. During this period, and potentially as a result of the upheaval, the Arabic type developments were put on hold. 187

Ibid., 2.

188

Ladislas Mandel to Brian Mulholland, Letter, “Design of New Faces in I.P.C. Studio - Paris - Classical Arabic,” 28 March 1975, 1, Fonds Mandel, folder Projets Arabe Photon, MI.

189

Ibid.

190

One may speculate that the politeness of the customer may well have trumped an honest assessment of the type.

191

Mandel, Photon’s Arabic Typography, 1.

192

Marshall, Du plomb à la lumière, 354, 373.

270

Chapter 3

However, under the new ownership the Parisian type design studio continued its work and by September 1975 a request from Malaysia re-initiated the last Arabic design project. Delairco, the Photon distributor in Asia, conveyed the interest of the Utusan Melayu newspaper to convert its hot-metal workflow to a photocomposition system, and requested advice about the prospect of creating a new Jawi typeface.193 Mandel thus provided samples of the new design, which the client considered to be “more suitable to their needs”, and arranged for a visit to Malaysia in January 1976.194 In a subsequent report to Lacoche Mandel summarised the meetings and outlined the plan for the project. Based on a Jawi keyboard that had been established in collaboration with the client, two additional weights of Photon’s new Arabic face were to be made. In a second stage an adaptation of an existing Linotype face was discussed, but then postponed. According to Mandel, the client was particularly “satisfied by the prospect of having an ‘original’ creation”, yet identified problems that might arise from the development of the new Arabic design. He noted that, although Ini could undertake the adaptation of a Linotype design, “she refuses to pursue in the future any creation work or program research overpassing the normal execution work of designers” [sic].195 This came, Mandel explained, as a result of the management’s decision not to revaluate designers’ salaries. He thus argued that if Dymo Graphic Systems did not reconsider their position, he did not see how the job could be implemented as it required “a very particular competence”.196 Although conclusive evidence is lacking, it appears that an agreement was found between designer and management. By August Delairco’s representative personally wrote to Ini, thanking her and Mandel for their work on the new type called Arabica, noting that the client was “very happy with the results”, and that their work on the typeface greatly helped marketing equipment in Malaysia.197 Compared to Nasriphot, Arabica was a clear ­improvement 193

Ron Simpson to John Griffiths, Letter, 4 September 1975, Fonds Mandel, folder Jawi News 55-65, MI. Simpson reported that the Nasriphot typeface had been rejected as unusable by everybody it had been shown to. Jawi is the name of the orthography used to write the Malay language with the Arabic script. Today Malay is mostly written in the Latin script.

194

Ron Simpson to Ladislas Mandel, Letter, 8 December 1975, Fonds Mandel, folder Jawi News 55-65, MI.

195

Ladislas Mandel to Claude Lacoche, Letter, “Meetings in Kuala-Lumpur - Utusan Malaya,” 27 January 1976, 2, Fonds Mandel, folder Jawi News 55-65, MI.

196

Ibid.

197

Ron Simpson to Joyce Ini, Letter, 24 August 1976, Fonds Mandel, folder Jawi News 55-65, MI.

271

Figure 3.42

“Arabica,” specimen, Dymo Graphic Systems, n.d., reduced to 65% linear. MI, folder Arabica. Courtesy of ­Musée de l’imprimerie et de la communication graphique (All rights reserved).

Photocomposition: Towards Immaterial Type

(­ Figure 3.42). Proportions and overall styling are close to those found in typical book faces such as the acknowledged model Monotype Series 549, yet Arabica is not a clone. Rather, it reveals diverse influences, and appears to mix elements of a number of existing typefaces. Whereas individual characters betray the model of Monotype Naskh, the overall appearance is distinctly different.

272

Chapter 3

Unfortunately, where Arabica diverges, it is not very successful. Text composition lacks cohesion because of erratic, often loose spacing, and the indecisive placement of diacritical dots. In contrast to Series 549, only two ligatures are included, contributing to a more elongated and less dynamic look. In addition, the two ligatures look out of place when used, as they stand out as oddities, not matching the remainder of the fount. On the level of character shapes, examples of unsuccessful designs are the initial and medial ‫ ك‬kāf which have awkward, overly condensed proportions, and its isolated form which drops unnecessarily low under the baseline, causing a rupture of the line. Having borrowed from established typefaces, Arabica passes as an acceptable design, but fails to convince fully as an original creation. In January 1977 Mandel received a letter from Delairco that summarised the setup of the system; it thus appears that the Paris office provided only the artwork, and the fount and keyboard layout were finalised locally. As the customer was replacing a Linotype hot-metal installation, the keyboards were kept identical to those of the old equipment, undoubtedly to ease the change for the operators. But whereas the linecaster employed a side magazine and a Pi tray for the manual insertion of individual matrices, these were now organised as a second fount on the matrix disc. In total the disc held eight founts of 112 characters each, allowing the inclusion of Arabica’s bold version as founts three and four, and the inclusion of regular and bold Latin with the remaining four founts.198 This setup proved successful and the client placed a second order for a Pacesetter with an identical matrix disc.199 Indeed, considering the change of printing quality photocomposition enabled, satisfaction of the customer was justified. Initially, Utusan Melayu was printed mixing hot-metal composition with the new photoset type, unambiguously demonstrating the advantages of the new technology (Figure 3.43). In contrast to the smudged, barely legible print obtained from the stereotyped hot-metal type, the photoset column had a clean and precise reproduction of the type, showing unquestionable advantages of legibility. Soon after the successful conclusion of this project, in 1977 the Paris type design office ceased to exist, thus marking the definite end of Photon.200 The company’s late involvement with Arabic type-making remained a curiosity of little impact on the market, or indeed future developments. As such, its case 198

Ron Simpson to Ladislas Mandel, Letter, 20 January 1977, 1–2, Fonds Mandel, folder Jawi News 55-65, MI.

199

Ibid., 2.

200 Marshall, Du plomb à la lumière, 373.

Photocomposition: Towards Immaterial Type

Figure 3.43

273

Detail of a page of Utusan Melayu, 24 August 1976. (above) actual size (below) enlargement, showing the quality difference between hot-metal and photocomposition. MI, folder Arabica. Courtesy of ­Musée de l’imprimerie et de la communication graphique (All rights ­reserved).

274

Chapter 3

illustrates the difficulties type-makers who lacked specialised expertise encountered when faced with non-Latin developments. Photon did not have the competencies required for successful Arabic type design, and failed to provide the necessary resources to gather expert advice; the language skills of an Arabic speaker among its staff may have been mistakenly interpreted as expertise in Arabic type-making – an untenable equation – and little profound research was undertaken to foster knowledge in the domain. In conclusion, Photon’s Arabic type development did not break new ground. Established practices were repeated with seemingly little understanding of their origins, and changes that were proposed failed to become successful. Although Photon’s technology had been revolutionary, none of it was applied to Arabic typography. Without thorough comprehension of the Arabic script and its associated typographical problems, it comes as no surprise that no new approaches could be put forward. Endeavours in Arabic Type by Dr Edward Plooij Similar to Higonnet and Moyroud, Dr Edward Bernard Plooij was an outsider to the printing trade. A Dutch citizen, he described himself as “mathematician, orientalist specializing in medieval Arabic science, and mechanical inventor”.201 In the latter role, Plooij entered the typographical stage in the late 1950s when he approached the leading typesetting manufacturers Linotype and Monotype, trying to interest them in new concepts for the typographical composition of Arabic. In 1958 Plooij had visited Monotype’s booth at the DRUPA printing fair, criticising the company’s Monophoto Arabic.202 By summer 1959 he sent Monotype a copy of a patent he had filed in the Netherlands, enquiring whether the Corporation was interested in his invention. According to the accompanying report by Monotype’s Continental Office, Plooij claimed “to have found a method of simplifying Arabic composition”, and was in contact with

201

Edward B. Plooy, “A New System of Photo-Composing Arabic Script,” Quaerendo IV, no. 4 (1974): 1. In some sources the author signs Plooij, in others Plooy. No systematic use of either spelling is apparent, here Plooij is generally used.

202 G. C. Barrett to D. Weller, Letter, “RE: Arabic/Dr. Edward Bernard Plooy, Alkmaar, Holland,” 11 August 1959, correspondence folder Arabic (Egyptian 3), MT. The DRUPA is the largest print and typesetting technology fair in the world. It is held every four years in Düsseldorf, Germany. The name DRUPA is a portmanteau for the German words Druck (print) and Papier (paper).

Photocomposition: Towards Immaterial Type

Figure 3.44

275

“Inrichting voor het machineschrijven of zetten van Arabisch en daarmee verwant schrift,” Dutch Patent NL113873, patent drawing, reduced to 50% l­ inear.

the Amsterdam Typefoundry, as well as the publishing house Brill in Leiden.203 The patent, subsequently published as Dutch patent Nr. 1138743, featured an analysis of conventional Arabic founts and suggested new principles of composition that were illustrated by drawings (Figure 3.44).204 Monotype’s UK Office reviewed the material and returned it to Plooij, noting that it was found to be interesting but that the company was “so committed with other languages […] that it would be several years” before his proposal could be reviewed in detail;205 moreover, Monotype’s management expressed the opinion that simplified Arabic could not be considered before the Egyptian Government approved it. Further exchanges ensued in which Plooij emphasised that his intention was by no means to devise a simplified Arabic; yet, Monotype reiterated their position that “until the learned authorities in Egypt are prepared to accept any version of Arabic other than that which they now find to be acceptable, we should be wasting time and limited capacity in carrying out a lengthy investigation”.206 203 Ibid. 204 Plooy, Dr. Edward Bernard. Inrichting voor het machineschrijven of zetten van Arabisch en daarmee verwant schrift. Dutch Patent NL113873, filed 17 July 1959, and issued 15 June 1967. For an English translation see Plooij, Dr. Edward Bernard. A Method and Apparatus for Typewriting or Composing Arabic or Related Writing. UK Patent 1,176,523, filed 10 January 1967, and issued 7 January 1970. 205 D. Weller to G. C. Barrett, Letter, “Simplified Arabic Composition, Dr Edward B. Plooij Alkmaar, Holland,” 14 August 1959, correspondence folder Arabic (Egyptian 3), MT. 206 C. Fellows to G. C. Barrett, Letter, “Simplified Arabic Composition, Dr. Edward B. Plooy,

276

Chapter 3

At the same time Plooij contacted Linotype, where his ideas and associated invention were reviewed by Tracy. Writing to Tage Bolander of ­Mergenthaler International, the group’s continental European office, Tracy outlined his opinion about Plooij’s system. As Tracy was in the process of becoming the Linotype group’s expert in Arabic, his views became the official company line. In contrast to Monotype’s staff, Tracy understood Plooij’s intentions, and articulated them thus: The object of the scheme is to restore to printing the true aesthetic quality of calligraphic Arabic script. This is to be done by – (1) Ensuring that characters overlap each other, not necessarily at a point midway between them. In order to achieve this, an elaborate and careful system of kerning is necessary, which, in the case of the typewriter is to be achieved by better spacing, and in the case of mechanical composition is to be achieved by use of photocomposition devices. (2) The vertical position of characters is to be alterable at will so that ligatures may be formed (i.e. two characters may appear as one above the other), or characters may be set higher than normal according to the character which follows. This effect is to be achieved by means of an escapement device which is described and illustrated. The authentic nature of characters is to be arrived at by the provision of a variety of design elements which are in some cases not complete characters but parts of characters which may be combined together at will. One aspect of this is the provision of a separate array of the dots which are in some cases the only difference between one character and another. It is stated that the scheme requires less than 100 elements with which letters, ligatures, points, vowel signs and figures can be achieved. It is therefore plain that this total is within the range of such a keyboard as on the I.B.M Executive typewriter.207 Tracy thus appreciated the potential of Plooij’s ideas, but retained reservations about their practicability. He observed that the principles required “a complicated system of character manufacture, an elaborate system of spacing

Alkmaar, Holland,” 22 September 1959, correspondence folder Arabic (Egyptian 3), MT. 207 Walter Tracy to Tage Bolander, Letter, “Scheme by Dr. E.B. Plooij for Composition of Arabic,” 6 October 1959, 1, Box P3640, NMAH.

Photocomposition: Towards Immaterial Type

277

adjustment, and very limited tolerances in mechanism.”208 Moreover, in 1959 the selection and positioning of characters would have been manual, rather than computer-aided, and in Tracy’s view it would have been severely taxing for the operator, significantly reducing the output of work. He thus considered “the scheme to be not feasible from the point of view of manufacture or of normal printing production requirements.”209 Despite the discouraging feedback Plooij received from the industry, he continued his work and extended its scope, and by December 1959 he filed a patent for a photocomposition device.210 One may speculate that, lacking interest from established manufacturers, Plooij hoped to apply his new principles for Arabic composition to a phototypesetter he developed himself. No further traces of Plooij’s endeavours in the coming years were found; in 1966, however, Monotype received a new reference to his work. This time, the printers Stephen Austin & Sons conveyed the information. Monotype reviewed the system, coming again to the conclusion that although it had “attractive features”, the company felt that the inventor “totally underestimates the difficulty and cost of providing sufficient mechanical accuracy in a productive machine to enable the various component parts of letters to be superimposed satisfactorily”.211 Now, Monotype grasped the principles suggested, but maintained that it would be too technically challenging to promise success. When the DRUPA fair was held again the following year, Plooij renewed his efforts to interest the leading companies in his invention. At the fair he met Linotype representative Ralph Goodman. In a brief report, Goodman stated that Plooij had been appointed UNESCO adviser “to develop low-cost Arabic photo-composition for the Middle East”, and that he had spent 17 years in developing his system.212 Concluding, Goodman remarked: He may be a crank or a genius, a Mark Twain or a Mergenthaler, but we must stay close to him as: 208 Ibid., 2. 209 Ibid. 210

Plooij, Dr. Edward Bernard. Inrichting voor het optisch projecteren van tekens op een opnemer. Dutch Patent NL123701, filed 14 December 1959, and issued 15 September 1967. For an English translation see Plooij, Dr. Edward Bernard. Apparatus for Optically Projecting Characters onto a Light Sensitive Surface. UK Patent 972,071, filed 13 December 1960, and issued 7 October 1964.

211

R.L.S. Harrison to Anonymous, Letter, “M/S/1805,” 17 November 1966, 1, correspondence folder Arabic (Egyptian 3), MT.

212

Ralph Goodman, Dr. E.B. Plooij, 7 June 1967, WT correspondence, folder 18a, DTGC.

278

Chapter 3

1) Arabic has just been adopted as the 5th official language for U.N.O. 2) He claims his system, whatever it may be, could be married to Linofilm Quick, and could be suitable for low-cost village printing! 213 Tracy, however, seemed less impressed. Although he acknowledged Plooij’s “thorough knowledge of the Arabic language and script”, he also discredited his attempts as being driven by a “‘romantic’ attitude towards mechanical composition of any kind”.214 Nonetheless, Tracy agreed to obtain further information in order to assess whether Plooij’s ideas had evolved in a more “practical” way. By February further documentation was gathered, and Tracy felt that his unfavourable views were confirmed. Returning the material to Linotype’s patent officer, Frank Bryant, he reaffirmed that Plooij’s ideas were “wrong”, as “the Job [sic] of the printer is first, to produce as many copies as is needed as quickly and as economically as possible; and second, to make the results as pleasing in appearance as possible, given the methods of production”.215 According to Tracy, the equipment available from Linotype and Monotype met these needs “very well”, and he concluded that “it is a pity that Plooij seems to be in a position where he can put forward his impracticable ideas to UNESCO officials who are willing to be impressed by experts, however wrong headed”.216 Plooij continued his endeavours for another two years before approaching Bolander again. In his renewed attempt he explained that a first experimental typewriter-composing machine had been built by the Science Department of the Board of Education and Science in The Hague. According to Plooij it followed the principles of his patented system, and the typeface it employed was based on the hand of the Egyptian calligrapher ʿAbd ar-Rāziq Muḥammad Sālim. Although Plooij claimed wide support for his project in the Arab world, his repeated attempts to interest Linotype in his work suggest that there was no prospect of any adoption of his system. As had previously been the case, Plooij’s message was handed on to Tracy, triggering again a similar reaction as before:

213

Ibid.

214

Walter Tracy to Arthur Henry Walker, Letter, “Mr. Plooij,” 20 October 1967, WT correspondence, folder 18a, DTGC.

215

Walter Tracy to Frank Bryant, Letter, “Dr. E.B. Plooij,” 28 February 1968, WT correspondence, folder 18a, DTGC.

216

Ibid.

Photocomposition: Towards Immaterial Type

279

His desire is admirable; but I am afraid it is unrealistic in terms of the practical requirements of modern printing. His principles of type design are the reverse of those embodied in simplified Arabic where, as you know, the reduction of character forms has been the way to tape composition. My own view is that just as swash characters have no place in high speed text composition in European languages, there is no place for the elegancies of calligraphy in Arabic composition. Only by making a truly typographic adaptation of calligraphic letter forms (and ultimately of separating the letters as in roman) will Arabic printing be able to take full advantage of modern technology. I believe this is the view of progressive Arabic publishers and printers.217 However, interested by the composing machine mentioned, Linotype pursued further exchanges with the inventor. Through D. Stempel AG Tracy obtained drawings of Plooij’s typeface (Figure 3.45 overleaf); judging from these, he realised that the character set included not only the four characters required for the basic typographical representation, but further context-dependent variations of letters. Intrigued, Tracy sought clarification from Plooij. Writing to Bolander on the subject, he noted that if the system relied on more than four shapes, the operator would need to select them manually – the concern Monotype’s review had first articulated – or that “a computer will have to perform the discriminatory function and the character array in the photosetting equipment will spread extravagantly over two or more grids or film strips.”218 Later in the year, after having reviewed Plooij’s UK patent, Tracy amplified the point about computer-aided typesetting in a letter to Bryant: My own belief is that to make printing imitate handwriting is to limit its development. For typesetting to be fast, cheap and versatile the character system must be as simple as possible. Dr. Plooij’s desire is not worth pursuing. Paradoxically, though, it might be viable in say, fifteen years time, when computerised Arabic composition may be widespread and a book printer may turn up with a strong desire (and cash) to produce elegant typesetting.219 217

Walter Tracy and Arthur Henry Walker, “Arabic Composition: Dr. Plooij,” 11 May 1970, WT correspondence, folder 18d Dib-Aoun, DTGC.

218

Walter Tracy to Tage Bolander, Letter, “Arabic: Dr. Plooij,” 14 July 1970, WT correspondence, folder 18d Dib-Aoun, DTGC. Ironically, five years later at the insistence of Gabeyan, Linotype produced Badr with a character set that required two film strips per weight.

219

Walter Tracy to Frank Bryant, Letter, “Dr Plooij: Patent 1176523 for Arabic Composition,”

280

Chapter 3

Figure 3.45

Above: Specimen fragment, Plooij, 1969, reduced to 40% linear. DTGC, folder 18a. ­Below: Detail, actual size.

Eventually, in November 1970 a meeting was organised between Tracy and Plooij. Here, it seems improbable that Linotype had any interest in pursuing a collaboration with the inventor; indeed it appears that the meeting was arranged to bring the subject to a definitive end. Writing to Plooij after their encounter, Tracy made the company’s position clear. After expressing that Linotype understood and respected Plooij’s aspirations, he reasoned:

21 August 1970, WT correspondence, folder 18 Al Ahram, DTGC.

Photocomposition: Towards Immaterial Type

281

However we are very conscious of the fact that that principle is contrary to the idea which is animating many people actually engaged in the practicalities of publishing and printing in the Arab world. I mean the idea of creating character forms especially for printing, which are necessarily different from those which occur in calligraphy and which are designed to simplify the style and number of characters needed for printed communication. After all, the Simplified Arabic which we first developed 15 years ego, and which is increasing in popularity, was first proposed to us by a printer in the Arab world; and Mr. Nasri Khattar [one of the proponents of Arabic script reform, see page 86 and following] is not the only person to advocate the idea of a single form for each alphabet character.220 In conclusion, Tracy noted that although the approval of politicians and religious leaders was positive, it would have been more pertinent to see a newspaper publisher or government printing office investing in the equipment proposed by Plooij. Moreover, he emphasised the economic concern, pointing out that the business case would have to be excellent to justify the license fee of £ 10,000 that Plooij requested. Despite the unambiguous letter, Plooij appears to have been unreceptive to the message, writing again to Tracy with an offer. In the reply, his disappointment and disillusionment are evident: I have no time to explain you why I do not agree with the reasoning in your letter. Besides the Arabic script using world is much larger than the printing community that, nolens volens, has accepted for the time being your Simplified Arabic. Arabic for typesetting will be simplified in future, but in another way, and by the Arabs themselves. With typewriters and fotocomposers [sic] the Gutenberg (Caxton) way is not the only way. And for Arabic it is not a good way too.221 Plooij concluded that “at my age I have to see to the completion of the work of my life”, making a final offer and ultimatum: requesting a decision within two weeks, Plooij presented an option in his invention against a fee of £3,000 and the cover of his expenses of a sales trip to the Middle East.222 In due time Lino220 Walter Tracy to Dr. Edward Bernard Plooij, Letter, 20 November 1970, WT correspondence, folder 18 Al Ahram, DTGC. 221

Dr. Edward Bernard Plooij to Walter Tracy, Letter, 23 November 1970, WT correspondence, folder 18 Al Ahram, DTGC.

222

Ibid.

282

Chapter 3

type refused the proposition, thus bringing an end to exchanges with Plooij which had been on-going for eleven years.223 By 1975 Plooij was still pursuing his idea, and wrote that “machines, suitably adapted for the new system, have been devised by me and are on the verge of being manufactured”.224 Plooij’s case demonstrates that although technological change offered an opportunity for new developers to propose new approaches to the printing trade, success still depended on other factors. The conservatism of established companies and the substantial capital investment photocomposition required proved to be insurmountable obstacles for Plooij. One may only speculate whether his fortunes, and indeed the development of Arabic type-making, would have taken a different direction had he approached an entrepreneurial newcomer like William Garth, rather than the dominant companies of the trade. The fact remains that the ideas first articulated by Plooij during this period were applied successfully in later years.225 Plooij’s underlying analysis of the Arabic script, and the shortcomings of its typographical representation still appear valid from a contemporary perspective, and the concise summary of Arabic typographical history he offered in 1975 remains more lucid than some current accounts. Writing about the beginnings of Arabic typography in the late 18th and 19th centuries, he stated: However, while in the West type designing had already developed into an independent art of high nobility, Arabic typography never became an artistically satisfying activity. To meet the Gutenberg requirements the running-hand writing had to be misformed [sic] somehow and the number of individual types necessary for achieving a reasonably close approach to calligraphic quality was extremely high, amounting in fact to several hundreds. Original types were costly and counterfeits often of low quality. The introduction of hot metal composing machines with a limited keyboard meant that the number of individual types had to be drastically diminished. At the same time the need for printed matter increased, and so did the difficulties for the printers.226 223

Walter Tracy to Dr. Edward Bernard Plooij, Letter, 1 December 1970, WT correspondence, folder 18 Al Ahram, DTGC.

224

Plooy, “A New System of Photo-Composing Arabic Script,” 332. No evidence of such an implementation has been found for the current study.

225

See also “The KITAB Software for the Hell Digiset” on page 327, and “Linotype Sheeraz, a New Approach to Arabic Type” on page 368.

226

Plooy, “A New System of Photo-Composing Arabic Script,” 331.

Photocomposition: Towards Immaterial Type

283

In 1975 Plooij still maintained some optimism when he wrote that “fortunately the deadlock has been broken by the advent of photo-composition”, suggesting that the increased flexibility of the technology would finally enable an uncompromised typographical representation of Arabic.227 Although in principle correct, further technological developments, in particular increased computerisation, were necessary to achieve a system that provided the functions Plooij envisaged whilst remaining economically viable. With hindsight, and indeed as had eventually been recognised by Tracy, it appears that Plooij’s ideas for Arabic typographical composition were ahead of their time. Arguably, the reasons that kept Plooij’s system from implementation were a lack of flexibility and vision by decision makers within established companies, and their emphasis on economic and short-term practical considerations.

227

Ibid.

284

Figure 4.1

“The fainting of Laylah and Majnun”, probably Shiraz, approximately 1550–1600, 12,2 × 16,7 cm. 1-86-154.123, Selections of Arabic, Persian, and Ottoman Calligraphy, African and Middle Eastern Division, Library of Congress, Washington, D.C.

Chapter 4

Persian Type and Typography Early Persian Printing and Typography In the Arabic script world the region of current Iran historically held a position of particular importance. It developed a distinct culture of text production, epitomised by the Nastaʿlīq writing style that emerged from the fifteenth century, and which became a characteristic feature of the highly developed Persian book art (Figure 4.1).1 Indeed, Nastaʿlīq became the dominant style with estimates suggesting that 75 % of all written output starting from the midfifteenth century in Iran was written in this hand.2 It is worthwhile to bear in mind that the Persian language is not a Semitic language like Arabic, but part of the Indo-European language family.3 Thus, and despite numerous Arabic loan words, linguistically speaking Persian is structurally closer to most European languages than to Arabic, creating text patterns that differ substantially from those of Arabic. These differencesَ are exemplified by the use of the definite article in both languages. In Arabic ‫ ا ل‬al is prefixed to each of the noun’s modifying adjectives, resulting in frequent repetitions.4 By َ contrast the Persian definite article ‫ را‬rā does not follow this rule and its lower frequency combined with varying letter shapes reduces the vertical rhythm found in Arabic. Moreover, Persian orthographic rules prescribe breaks within words that would be joined in Arabic orthography. Visually, Persian text is thus more fragmented with a higher frequency of mid-word breaks and generally shorter letter groups. These linguistic differences have been suggested as explaining the emergence and preference of the Nastaʿlīq style for this language.5 1 For a detailed account of the history of the so-called hanging scripts see Blair, Islamic Calligraphy, 270–86. 2 William L. Hanaway and Brian Spooner, Reading Nastaʻliq: Persian and Urdu Hands from 1500 to the Present (Costa Mesa, CA: Mazda Publishers, 2007), 3. 3 “Ethnologue. Languages of the World.,” n.d., http://www.ethnologue.com/ (accessed 16 May 2017).

4 For example ‘the new Simplified design’ could be translated as َْ َ ْ َّArabic َ ُ ‫َْ َ َ َ ْ ُْ َ َ ة‬ ‫’ ا � ت�ل�����ص�ِ��مي� ا �جل� ِ�د ي��د �ِل��ل��ح‬, at-taṣmīm al-ǧadīd li l-ḥurūfi al-ʿarabiyyä al-muḫtaṣarä, ‘ �‫�ُر و�فِ� ا �ل�عَر�يّ���ة ا لم�� خ�ت�����ص َر‬ ِ‫ب‬ ‫م‬ literally meaning ‘the design the new for the letters the Arabic the simplified’. 5 Vlad Atanasiu, De La Fréquence Des Lettres et de Son Influence En Calligraphie Arabe (Paris: L’Harmattan, 1999); Blair, Islamic Calligraphy. According to Blair, Nastaʿlīq was never popu-

286

Chapter 4

As in other regions of the Arabic script world, typography arrived late in Iran and was slow to find acceptance as a means of text production. Printing was introduced to Iran with the political changes initiated during the Qajar dynasty (1785–1925), which opened up the country to European influences. The first book known to have been printed in Iran dates from 1817.6 European technology – specifically letterpress printing – was promoted as an agent of modernity.7 Yet, the incunable period of Persian typography was brief and of marginal reach. The first half of the nineteenth century only saw a total of 55 printed editions and manual text manufacture remained the norm. Extant samples of these early letterpress printed books show poor type and printing quality, factors that undoubtedly played a role in marring the success and further spread of Persian typography at this point.8 It comes as no surprise that a readership used to the degree of perfection found in manuscript production would have rejected literary works of such an inferior appearance. Indeed, the introduction of a different printing technology that allowed more fidelity to the aesthetic norms found rapid acceptance and success. The lithographic printing process was brought to Iran from Russia in 1832– 33, when a first print shop was set up in Tabriz and a Qurʾān printed.9 By the middle of the century, the technique had spread to other cities in Iran and it dominated the trade until the 1870s, when improved typographical printing reappeared. The success of lithography was such that by the end of the nineteenth century almost five times as many books were printed by lithography than by letterpress.10 Ulrich Marzolph has pointed out that a number of factors favoured the new printing process.11 Compared to the cost and complexity of the equipment required for a letterpress workshop, the initial investment in material lar for writing Arabic. This statement appears too categorical. Whilst it is true that it is rarely used for longer texts, it is frequently found in use for short display purposes such as shop signs, titles and logos. 6 Ulrich Marzolph, “Early Printing History in Iran (1817 – ca. 1900),” in Middle Eastern languages and the print revolution, 251. 7 See also Peter Avery, “Printing, the Press and Literature in Modern Iran,” in The Cambridge History of Iran, vol. 7 (Cambridge University Press, 1991), 819. 8 For examples of such early Persian typography see Marzolph, “Early Printing History in Iran (1817 – ca. 1900).” 9 Ibid., 256–57. According to Avery, lithographic printing began already a decade earlier. Avery, “Printing, the Press and Literature in Modern Iran,” 819. 10 Marzolph, “Early Printing History in Iran (1817 – ca. 1900),” 257–61. 11 Ibid., 263–68.

Persian Type and Typography

287

and ­training was reasonable. The influential calligraphers’ guild, which understandably feared typography as jeopardising their craft, recognised that lithography did not pose the same threat as letterpress printing. This process required, and potentially increased, demand for their skills. However, besides these economic considerations, another decisive factor was the quality of the final product. The appearance of European-made Arabic printing types in the nineteenth century “was too crude and unshapely to satisfy the aesthetic sensibilities of the reading public”, an issue that did not arise in lithographic printing.12 Moreover, Marzolph argues that “this dissatisfaction was further reinforced by the rather irritating fact that inevitably all Persian incunabula were printed using the only letters in existence – which were Arabic”.13 According to Marzolph, the lack of Persian characters forced printers to approximate spellings with the closest matching Arabic letters. Yet, given the ease by which the Persian accented letters could be added to existing founts, this does not seem to be the most plausible scenario. Many early type founts contained Persian characters but, despite their nominal completeness, their style might have been perceived as being Arabic and therefore rejected by a Persian readership. Also newspapers, the first published in Iran in 1837, relied largely on lithography for most of the nineteenth century.14 The history of Iran’s Press was closely linked with political changes, and increasing publishing freedoms in the early twentieth century meant that newspapers blossomed.15 A turning point came in 1925 when the last Qajar Shāh was overthrown by Rezā Shāh Pahlavi. Although he is widely credited as a modernising force, the number of newspapers plummeted under his reign from around 150 to 50 by 1940, indicating the erosion of the publishing sector through authoritarian leadership.16 However, two newspapers published in the capital Tehran, Iran (founded in

12

Ibid., 265. Marzolph is a rare case that points out the aesthetic issues of printing type, as printing-historical and palaeographic accounts tend to emphasise religious and/or political reasons for the late acceptance of printing in the Arabic script world.

13

Ibid.

14

Avery, “Printing, the Press and Literature in Modern Iran.” Avery’s detailed account of the development of the Press contains barely any observations about printing techniques, but mentions the change over from lithography to letterpress printing of the important Ḥabl al-Matīn newspaper in 1905.

15

Ibid., 839.

16

Under his reign, the University of Tehran was established and large construction projects initiated. Encyclopædia Britannica Online, “Reza Shah Pahlavi,” https://www.britannica.com/biography/Reza-Shah-Pahlavi (accessed 16 May 2017).

288

Chapter 4

1915), and Ettelaʿat (founded in 1925), survived and became dominant.17 From the 1940s Kayhan, another Tehran-based newspaper became Ettelaʿat’s principal competitor, but after 13 years of competition the two papers joined forces in the mid-1950s “to form a partnership with which none could compete. They had the plant, experience and financial ability that enabled them to become a publishing empire without precedent in Iran.”18 Their influential position was also to be felt in the field of typography. The beginnings of mechanical typesetting in Iran are poorly documented in the literature. From the account of Reginald Orcutt it emerges that by 1927 Arabic linecasting machines were not yet known in Iran, and that the majority of printers used German equipment and type, reflecting the strong economic ties of the two countries during that period.19 Evidence for mechanical composition in Iran begins in the 1950s. According to Tracy, Ettelaʿat Publications commissioned the manufacture of matrices in 9 pt and 10 pt from Linotype in 1957.20 The following year an article in Linotype Matrix publicly announced the project, noting that this installation was the first for mechanical composition of Persian by a newspaper in Iran.21 The article also acknowledged the cultural preference for Nastaʿlīq, and explained how Linotype approached the task. A Nasḫ style typeface “with certain characteristics of nastaliq” was thus taken as model and a corresponding fount manufactured.22 According to the article this was done “so successfully […] that columns of Linotype-set text appearing adjacent to columns set by hand were virtually indistinguishable”.23 Indeed, it appears that this earliest recorded collaboration between the publishing house and Linotype was successful and thus commenced a lasting business relationship that continued into the 1980s.24 17

Ṭṭilāʿāt (information / knowledge). Avery, “Printing, the Press and Literature in Modern Iran,” 845.

18

Kayhān (universe). Ibid., 859.

19

Orcutt, Merchant of Alphabets, 62.

20

Walter Tracy to E. Gabrielian, Letter, “Iran Chap : Type Faces.,” 28 March 1968, WT correspondence, folder 18c Persian and Pushtu, DTGC.

21

Linotype & Machinery Ltd, “Ettela’at, First Persian-Language Newspaper to Employ Mechanical Composition in Iran,” Linotype Matrix II, no. 29 (June 1958): 2.

22

In contrast to other type developments from the same period, no documentation of the design and manufacturing process could be found.

23

Linotype & Machinery Ltd, “Ettela’at, First Persian-Language Newspaper to Employ Mechanical Composition in Iran,” 2.

24

Also the Intertype Corporation is known to have supplied typesetting equipment and corresponding Persian founts during the same period, yet detailed evidence and records

Persian Type and Typography

289

Figure 4.2 Alahram specimen (detail). From Zapf, About Alphabets: Some Marginal Notes on Type Design, reduced to 70% linear, apparent kerning pairs marked.

At around the same time another, unrelated, contribution to Persian typography came from the German calligrapher and type designer Hermann Zapf. According to his personal account, in 1954 the Stempel foundry commissioned a typeface in which “the customary kerned characters were to be avoided, to make the new Arabic more useful for newspaper composition than the founts hitherto used”.25 Judging from the accompanying illustration, this technical remark is curious, as the type seems to show kerned character pairs (Figure 4.2); indeed, although D. Stempel AG manufactured matrices for linecasters, Zapf’s design appears to have been conceived as foundry type, making kerning possible and desirable.26 According to the designer’s recollection the type was finished and sent to Tehran in 1956 and met with great approval. Yet no actual use of it could be traced and other typefaces of documented success in Iran have little in common with Zapf’s design. The type was subsequently revised and named Alahram, in an apparent reference of the Egyptian newspaper. Indeed, close examination of the specimen reveals that its title Persian by Hermann Zapf was conspicuously extended to include Arabic, suggesting little success in the initially intended Persian market (Figure 4.3 overleaf).27

of this company’s work in Arabic-script typography have yet to emerge. 25

Hermann Zapf, About Alphabets: Some Marginal Notes on Type Design (Cambridge & London: M.I.T. Press, 1960), 50.

26

No evidence of an implementation on any of the then current typesetting machines was found.

27

According to a recent publication by Zapf the typeface was mainly used by al-Ahram, but no evidence of this use could be found for the current study. Hermann Zapf, Alphabet Stories: A Chronicle of Technical Developments (Bad Homburg: Mergenthaler Edition, Linotype GmbH, 2007), 37. Recently the design was revived digitally and is now distributed as Palatino Arabic. Monotype GmbH, “Palatino® Arabic Font Family,” n.d., http:// www.linotype.com/286269/PalatinoArabic-family.html (accessed 16 May 2017).

290

Chapter 4

Figure 4.3 “Persian by Hermann Zapf & Arabic,” specimen. From Hermann Zapf: calligrapher, type-designer and typographer. An exhibition arranged and circulated by The Contemporary Arts Center, Cincinnati Art Museum 1960–1961. The Contemporary Arts Center/ Cincinnati Art Museum, Cincinnati, 1960, reduced to 60% linear.

A Simplified Typeface for Persian Soon after the introduction of Simplified Arabic and its rapid success in Lebanon and Egypt, the traditional centres of publishing in the Middle East, interest in its principles extended outside the Arabic speaking world. By 1963 the principles of Simplified Arabic had come to the attention of Linotype’s long-standing customer Ettelaʿat Publications in Tehran.28 The firm printed amongst other publications Persian, English and French language newspapers and ranked among the leading printing establishments in Iran. At their request Linotype manufactured a small quantity of Persian matrices according to the principles of Simplified Arabic and provided them in early 1964 for tests in Tehran.29 These matrices were directly based on Simplified Arabic 8 with 9 and adapted for Persian only insofar as the required diacritical signs had been 28

The publishing house’s name was later changed to Iran Chap.

29

Walter Tracy to Michael A. Nahas, Letter, “Persian (Simplified),” 16 December 1963, WT correspondence, folder 18c Persian and Pushtu, DTGC.

Persian Type and Typography

291

Figure 4.4 Proof of Linotype Simplified Persian fount, composed by Ettelaʿat, June 1964, actual size. DTGC. WT correspondence, folder 18c Persian and Pushtu.

Figure 4.5

“Specimen of 10-pt (German) loose type selected by Ettela’at ( for simplifying),” 1964, actual size. DTGC. WT correspondence, folder 18c Persian and Pushtu.

292

Chapter 4

added. By June Ettelaʿat had tested the fount for Persian composition and concluded that the design was of no use to them (Figure 4.4 on page 291). In the reply the company’s general manager S. Moshkriz stressed that the design was Arabic and reiterated the need for a Persian fount based on the same principles.30 The report of Linotype’s representative Michael Nahas echoed the client’s view that a distinct design was necessary: They all liked very much the Simplified system, and wanted to use it soon but with a new design based on Persian style. They suggested to copy from a new Intertype Traditional face name ‘Saadi’ which they liked and is now used on their two Intertypes. However, I advised them to better select or design a new good face, than to imitate the Competitor’s.31 Although the client was unambiguous about the typeface’s perceived Arabic character, a definition of what features would make a design look Persian was not provided. Instead, examples of typefaces which were considered to suit the local preferences were suggested to the manufacturer. Besides the above mentioned Intertype fount, Moshkriz provided a sample of German foundry type in 10 pt as being particularly suited, asking Linotype to produce a matrix fount in its likeness (Figure 4.5 on page 291). For Tracy this feedback seemed hard to comprehend; the sample that was provided was not Nastaʿlīq, the style most commonly associated with Persian writing culture, as he noted in a letter to Nahas: “the German loose type is certainly a traditional Naskh design, the descenders being shorter than usual because of the small body size. It is not Persian (Nastaliq) in any way.”32 Although in principle correct, this reasoning also demonstrates unfamiliarity with the distinct typographical identity that developed in Iran. Here, Tracy discovered that some designs based on the Nasḫ style were perceived as being more Persian than others; at this time, however, the subtle differences which defined this national identity remained opaque to Linotype’s typographical adviser. Traces of the project only continue from 1966 when Linotype approached their client with the suggestion to adopt their newly finished Series 10 with 11 (later called Raed) for the Persian language (Figure 4.6). It was the second 30

Moshkriz to Linotype & Machinery Ltd, Letter, “Re- Simplified Persian Matrices,” 3 June 1964, WT correspondence, folder 18c Persian and Pushtu, DTGC.

31

Michael A. Nahas to Walter Tracy, Letter, “Letter/Report Simplified Persian,” 9 June 1964, WT correspondence, folder 18c Persian and Pushtu, DTGC.

32

Walter Tracy to Michael A. Nahas, Letter, “Simplified Persian,” 23 June 1964, WT correspondence, folder 18c Persian and Pushtu, DTGC.

Persian Type and Typography

293

Figure 4.6

Left: Raed specimen, not dated, Arabic language cover, reduced to 40% linear. DTGC, box Arabic general.

Figure 4.7

Right: Preliminary specimen Raed Persian, May 1966, annotations by ­Massoudi, reduced to 75% linear. DTGC. WT correspondence, folder 18c Persian and Pushtu.

t­ ypeface designed by Nabih Jaroudi and had been in development for a few years. Although the typeface was not popular, its more conventional appearance might have suggested it to Tracy as a suitable design for the Iranian client.33 Moreover, the addition of Persian characters to the fount was a minor undertaking for Linotype, and it appears that Tracy remained reluctant to comply with Ettelaʿat’s wish for a new typeface. By May 1966 Persian trial settings of Raed were printed and sent for evaluation to Tehran, and in early June Nahas reported back from a meeting with the prospective customer. The typeface was slightly better received than Simplified Arabic, yet the client remained critical of the design.34 At this point 33

The typeface had no commercial success in hot-metal composition and it was not adapted for later technologies. Hrant Gabeyan, interview by the author, 11 March 2011, DTGC.

34

Michael A. Nahas to Walter Tracy, Letter, “Persian Type-Face No.10 with 11,” 2 June 1966,

294

Chapter 4

Ettelaʿat provided for the first time some indications of design features that were expected from Persian type. According to Nahas, the �‫  ب‬bāʾ and the bowl of ‫ ��س‬sīn and related letters were too narrow, and the initial form of ‫ ك‬kāf was considered excessively angular and should be more open in shape. Further to the report by Nahas, the managing director of Ettelaʿat, Farhad Massoudi, personally wrote to Tracy to explain the problems of the typeface. In what he termed “glaring irregularities”, he described characters as “odd shaped”, of erroneous proportions or “too rigid and not Persian”, to conclude that “the big problem is that the new type face is too Arabic” (Figure 4.7 on page 293).35 As an illustration and to underline the stylistic differences Massoudi enclosed a sample of what he considered a successful design (Figure 4.8). The typeface features a pronounced horizontal joining stroke, which contributes to proportions that are more removed from the manuscript form than those found in Raed. Overall it is noticeably rounded, with a prominence of smooth curves and an avoidance of sharp corners, cuts and steeps angles. Letter shapes which tend to strongly reflect the influence of the traditional writing tool show a high degree of abstraction and rationalisation, as exemplified in the four characters of mīm. The design appears formally consistent and commends itself for well ‫م‬ executed letter shapes, yet the fount’s manufacture lessens its overall quality. Spacing is erratic, a shortcoming exacerbated by an apparent lack of kerning. Given these faults, and what appears as traces of matrix side-walls, the fount was probably made by Intertype. Raed, on the other hand, has different design features. Its overall colour is rather lighter than the norm and the bold version appears unusable for headlines as it has barely sufficient weight to stand out against the regular. Stylistically the typeface resembles the trade lettering found in Middle Eastern newspaper titles of the period, yet lacks their spontaneous quality. Its spikey look echoes the broad-nib pen and stands diametrically opposed to the gentle softness of the fount Massoudi recommended. Raed’s proportions and treatment of related letter shapes such as ‫ و‬wāw and ‫ ر‬rāʾ lack consistency and contribute to a spotty text that is marred by the poor spacing of the linecaster. Where the sample provided by Ettelaʿat conveys a calm and balanced impression, Raed appears frantic and unsteady. But while the quality of the design itself is debatable, for Ettelaʿat the decisive problem was the cultural association of the style. Paramount in the client’s rejection of the typeface was the identity that was WT correspondence, folder 18c Persian and Pushtu, DTGC. 35

Farhad Massoudi to Michael A. Nahas, Letter, 27 June 1966, WT correspondence, folder 18c Persian and Pushtu, DTGC.

Persian Type and Typography

Figure 4.8

295

Sample of Persian fount provided by Massoudi (detail), actual size. The most probable source of this typeface was the Intertype Corporation. DTGC. WT ­correspondence, folder 18c Persian and Pushtu.

conveyed through certain formal characteristics, an issue that seems to have outweighed the concern for typographic quality. Despite these repeated exchanges, no immediate order of equipment, nor indeed the manufacture of a new typeface ensued. In autumn 1967 steps were taken to continue the project of a simplified Persian typeface when a meeting between Ettelaʿat and Linotype was held in Tehran. As a result a document detailing a number of changes suggested for the adaptation of Simplified Arabic was drafted and sent to Linotype (Figure 4.9 overleaf). A memorandum reiterated remarks similar to those made in relation to Raed, and outlined additional changes to characters. Although the feedback applied to two different typefaces, some common themes emerge from Ettelaʿat’s comments. Generally, wider and more open proportions were sought, and shapes of calmer, more rounded appearance preferred. Design aspects that derived from the simplification principles such as the flat medial ḥāʾ were considered awkward, and ‫ح‬ the closed and angular counters of mīm were not appreciated by the client. ‫م‬

296

Figure 4.9

Chapter 4

Simplified Persian - Points discussed, page from Douglas Turner to Walter Tracy, “Simplified Persian”, Letter, 29 November 1967, reduced to 70% linear. DTGC, WT corres­pondence, folder 18c Persian and Pushtu.

Persian Type and Typography

297

Tracy, however, still seemed unconvinced. In a reply to the aforementioned report he underlined that the design was the “work of an expert calligrapher”, and articulated his scepticism as to the feasibility of suggested diversions from the simplification scheme.36 Although three years had passed since the client’s first appeal for a new, specifically Persian design, Linotype was still reluctant to accept and appreciate the particular typographical identity sought by the client. This stance was only to change with firm orders of equipment. By April 1968, Linotype representative Ralph Goodman assessed the situation as largely unchanged: I am afraid that Iran Chap [the former Ettelaʿat] have not yet got to the position where they are willing to accept Simplified Persian, although they realise that, for the long term, they will need to change their views […] Still the only real convert to Simplified is Mr. Dariush Homayoun, who is still unable to persuade his Board of Directors either to buy our machines, or to accept Simplified – they are at present composing Traditional on Intertypes.37 Further protracted exchanges persisted for months, with both parties reiterating their respective views – Tracy trying to convince the client of the merits of the existing types, and Iran Chap maintaining that none of the choices was acceptable. At this time for Linotype new typeface developments were subject to machine and equipment sales and the long-term interest of the company. By the 1960s the end of hot-metal composition was imminent and consequently the manufacturer was hesitant to invest more in this soon-to-be obsolete technology. The customer’s insistence remained insufficient for as long as it was not underlined with purchasing power, a fact that changed in July 1968.38 As part of a large order of equipment from Linotype and the commitment of continuous purchases over a number of years, Iran Chap reiterated its need for a newly conceived simplified Persian typeface. Apparently aware of its improved bargaining position, and with the assurance to purchase a minimum of 20 founts upon completion, the publishing house proposed that its in-house

36

Tracy to Nahas, “Simplified Persian.”

37

Ralph Goodman to Walter Tracy, Letter, “Iran Chap – Tehran,” 1 April 1968, WT correspondence, folder 18c Persian and Pushtu, DTGC.

38

The client had already been a large customer with its 14 Linotype casters, but the new type development seemed bound to new orders.

298

Chapter 4

expert, Mr Haghighi, would draw a new typeface.39 As exchanges continued amid apparent indecision by Linotype, eventually Iran Chap commissioned the new design unilaterally.40 From this point on developments attained a different pace. By 22 September a first set of drawings was finished by Haghighi and sent to the United Kingdom for examination by Tracy. In acknowledging receipt, Goodman outlined the terms of manufacture to the company’s agent in Tehran: 1. A minimum of 15 founts must be ordered in each fount size required, and three fount sizes will be the maximum for this face. 2. Price per matrix would be 2/11d. 3. To cover our development costs, a sum of £500 must be deposited by Iran Chap, and this fixed development change would be deducted from the cost of matrices when purchased. 4. As regards Mr. Haghighi’s fees, for each 25 founts of this new Simplified Persian bought in Iran, we are prepared to pay £100 to Iran Chap, towards their design fees. 5. Due to our extremely heavy commitments, we could not promise delivery of this new face before two years from 1st July 1969, of course if we could improve on this delivery we would do so, as we are as anxious as Mr. Massoudi is to establish this new face, with its great advantages using ‘Elektrons’ and a computer.41 The project was pursued irrespective of these demanding terms, demonstrating the urgency and importance the client assigned to a typeface that was acceptable for the Persian market. The drawings received by Linotype show a high degree of craftsmanship, suggesting that Haghighi was acutely aware of the general requirements and the precision matrix manufacture called for (Figure 4.10). Width of strokes and distribution of contrast are consistent, 39

Ralph Goodman to Walter Tracy, Letter, “Persian Faces,” 5 August 1968, WT correspondence, folder 18c Persian and Pushtu, DTGC. According to contemporary Persian type designer Masoud Mohammadzadeh, his full name was Hossein Abdollahzadeh Haghighi � ‫ح��س��ی�ن �ع ب���د ا �ل�ل�ه �ز ا د ه‬ �), email message to the author, “Re: Haghighi,” 9 September 2013. (‫ح��ق��ی��ق��ی‬

40

In August 1968, Goodman, by then seemingly convinced of the necessity for this undertaking and realising the potential economic interest, reported to Tracy that Iran Chap “have jumped the gun on commissioning the face from Mr. Haghighi”. Ralph Goodman, “Persian Faces,” 13 August 1968, WT correspondence, folder 18c Persian and Pushtu, DTGC.

41

Ralph Goodman to E. Gabrielian, Letter, “Iran Chap - Simplified Persian,” 26 October 1968, WT correspondence, folder 18c Persian and Pushtu, DTGC.

Persian Type and Typography

Figure 4.10

299

Artwork for Simplified Persian, 1968, reduced to 70% linear. DGTC, box Simplified Arabic.

300

Chapter 4

alignment is carefully observed and the general appearance seems to follow a coherent and consistent approach.42 Haghighi’s design is particularly informative in terms of its aesthetics, as it demonstrates the characteristics that were considered typically Persian by the client. The character repertoire followed the principles of Simplified Arabic, reducing the number of characters required for a joined appearance. All letters of the Persian alphabet in a regular and a bold weight were contained in the drawings. Although some drawings appear similar to Simplified Arabic, reflecting the influence of the simplification scheme on the design (see initial and medial �‫ ب‬bāʾ), the overall impression is one of a distinctly different voice. Letter shapes appear softer because of rounded edges, smooth curves and less frequent cuts. Most angles are shallower than in Simplified Arabic, resulting in more open half-counters and larger, more spacious counters. In general, the style of the drawings is more strongly removed from the manuscript form than typically the case in comparable Arabic designs, and contributes to a distinct typographical appearance. While both weights appear largely consistent by themselves, a comparison between regular and bold reveals some discrepancies. Most characters have similar configurations, but ḥāʾ, ʿayn, �‫ �ي‬yāʾ and mīm were drawn differ‫ح‬ ‫ع‬ ‫م‬ ently in the two weights (Figure 4.11). Final and isolated ḥāʾ are marked by a ‫ح‬ differing weight distribution in the bowl, contributing to a change of stress and increased contrast, also seen in the respective characters of ʿayn. Similarly, ‫ع‬ the proportions and dynamic of isolated �‫ �ي‬yāʾ differ substantially, yet the most noticeable inconsistency between the two weights is found in initial mīm. The ‫م‬ regular has the conventional, though visibly rationalised three-stroke configuration of the manuscript form. By contrast, the bold retains no trace of the writing tool in its highly synthetic typeform. Its elliptic shape of geometric perfection features an unconventional stress and contrast distribution and retains barely any stylistic proximity to the other weight. These inconsistencies were, however, neither accidental nor resulting from lack of skill. A close examination of the drawings reveals that these substantial variations were prescribed by the linecaster mechanics. Because both weights were duplexed, i.e. struck into the same matrix, every character had to be designed to the same width in regular and bold (Figure 4.12). Therefore unconventional solutions had to be found in the configuration of the bold letterforms; the same matrix widths had to accommodate strokes sufficiently thick to distinguish the two weights. Conversely, the design of the regular might also have been widened to provide enough space for the bold characters. Numerous letter shapes in the lighter weight feature 42

Linotype may have briefed Haghighi for his task.

Persian Type and Typography

301

Figure 4.11

Left: Notional indication of the matrix widths of selected characters, demonstrating the effect of duplexing on the design of letterforms. Adapted from Figure 4.10.

Figure 4.12

Right: Illustration of a composed line of double letter matrices, demonstrating why duplexed founts had to assign the same matrix widths to different weights. From Legros and Grant, Typographical Printing Surfaces, 426.

a­ typical elongations which are suggestive of the design compromises that duplexing necessitated (e.g. see initial and medial ‫ �ه‬hāʾ and final ʿayn). ‫ع‬ By January 1969 Tracy had reviewed the drawings and suggested a number of improvements such as an increase of the bold’s weight, improvements of individual character shapes and the inclusion of numerals specifically conceived for the new typeface.43 In parallel to the revision process, the parties agreed to the terms of the typeface development. Iran Chap was granted 18 months exclusivity and the design was given the name Ettelaʿat.44 By February Haghighi had redrawn all bold characters and implemented some changes in the regular and the design was sent to Linotype for manufacture. Haghighi’s drawings, mounted on six cardboard panels, show a modestly heavier bold face and minor adjustments to individual characters, but generally retain the characteristics of the earlier draft (Figure 4.13 overleaf). With the design settled, Linotype commenced manufacture in due course and by July 1969, trial proofs in 10 pt of 43

Walter Tracy to Farhad Massoudi, Letter, “Simplified Characters,” 15 January 1969, WT correspondence, folder 18c Persian and Pushtu, DTGC. Haghighi followed the suggested increase of weight for the bold design, and used his discretion regarding Tracy’s comments on individual character shapes.

44

Farhad Massoudi to Ralph Goodman, Letter, “Simplified Persian,” 12 February 1969, WT correspondence, folder 18c Persian and Pushtu, DTGC.

Chapter 4

Figure 4.13

Ettelaʿat Simplified revised artwork, 1969, reduced to 30% linear. DTGC, box Simplified Arabic.

302

a selection of characters were printed and approved by Massoudi (Figure 4.14). Linotype thus proceeded with the remainder of the fount, including the development of auxiliary components such as keyboards and their documentation (Figure 4.15 on page 304). Yet, another year passed before the product was

Persian Type and Typography

303

Figure 4.14 “10D-point Trial Characters for Ettela’at Simplified,” proof on newsprint, August 1969, actual size. DTGC, box Simplified Arabic.

finished because its manufacture was hampered by quality issues: poor alignment and inconsistent depth of the matrices required repeated revisions. Approximately six years had passed from the customer’s formal request for a new Persian typeface to the point that a usable product was made available in 1970. While practicalities like manufacturing capacity, long-distance communication, and financial considerations were undoubtedly decisive in the length of this process, Linotype’s ignorance and incomprehension also contributed to the delays.45 The notion of a typically Persian typographical identity was a novelty for the company’s decision makers, and its formal expression remained intangible for years. Haghighi’s Ettelaʿat typeface thus marked an important point in Persian typographical history. Here the specific cultural preferences found an expression in a distinct and articulate form. Designed with local expertise and in accordance with a scheme conceived for the technical requirements of the Linotype, it is another example of the fruitful collaboration between Middle Eastern clients and Western manufacturers. As an original Persian typeface, 45

In the meantime Linotype & Machinery Ltd also merged with K. S. Paul Ltd to become Linotype-Paul Ltd in 1967.

304

Chapter 4

Figure 4.15

“Linotype Simplified Persian Keyboard 387 with French Keyboard 103,” 1969, ­reduced to 50% linear. DTGC, box Simplified Arabic.

it may have served as a model for subsequent designs. The experience of this project also taught Tracy a lesson of cultural sensitivity, as illustrated by a letter he wrote to fellow Linotype staff: “They allege that the shape of the characters is typical Arabic and not Persian at all…” This quotation from a recent letter from Gabrielian […] is typical of statements made on a number of previous occasions. This is rather like an Italian printer or publisher looking at some Italian text matter and complaining that it looks too English – because it is set in Baskerville, a style of letter which has never been popular in countries which prefer Garamond and Bodoni. That is to say, the people in Iran are not saying they cannot read the type: they are simply expressing an irrational prejudice – strengthened, perhaps, by nationalistic factors of political or emotional origin. Nevertheless, the prejudice is very strong, and will only be dispelled as the appetite for modern equipment increases. I think that for quite a long time we may have to face demands from printers in Iran for type faces particular to that country, and we will have to accept the fact that these designs will have no sales value outside that territory.46 46

Walter Tracy to Arthur Henry Walker, Letter, “Iran: Typography,” 27 February 1970, WT correspondence, folder 18c Persian and Pushtu, DTGC.

Persian Type and Typography

305

Besides Tracy’s belated recognition of Persian typographical identity, the prospect of change he suggests is emblematic for his views on typographical progress. The suggestion that the need for new technology will overrule, and eventually “dispel” cultural preferences is at once an assessment, and a self-fulfilling prophecy. Tracy was aware of the continuous need for new equipment and the limitations it, in turn, imposed on Arabic type. It appears that Tracy accepted technologically imposed change and the compromises it entailed for aesthetics, style, and regional preference as unavoidable side-effects of progress. The long delay in the implementation of the Simplified Persian project, paired with rapid progress in the field of typesetting technology, meant that by the time the founts were finished hot-metal composition had almost reached obsolescence.47 By the late 1960s photocomposition became increasingly demanded by clients in the Middle East, indicating that the industry had reached maturity. The newspapers of the most developed countries attained unprecedented reader numbers, and the correspondingly bigger budgets were reflected in large orders of machinery and the call for the latest technology.48 As a result, competition for this increasingly interesting market intensified and new manufacturers tried to gain a foothold in the region, forcing the established companies to upscale their involvement. A case in point is Iran, where the publishing industry attracted only minor interest from typesetting equipment manufacturers prior to the 1960s. Following the radical reform policies of the Pahlavi regime at the beginning of the decade, encompassing land reforms and the support of industrial expansion in combination with increasing oil revenues, Iran’s economy grew dramatically.49 These considerable societal changes came to be echoed in the printing trade, where publishers and printing houses solidified their position as important clients of Western manufacturers. As demonstrated above, large orders provided a negotiating position which allowed clients to influence type developments to a hitherto unknown extent. In 1970 Iran Chap ordered Linotype equipment worth £57,000 and ­additionally 47

A few years later, when Ettelaʿat became one of the first Linotron 505C customers, Simplified Persian was foregone in favour of the new Osman type. Peter Watkins to Richard J. Caesar, Letter, 13 March 1974, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC.

48

See also “The Development of Character Selection and Kashida Insertion Routines” on page 187.

49

From 1965 to 1972 the country experienced an average annual rate of growth of GNP of 10.4 %, and, despite its significant population growth a parallel increase of per capita GNP of 7.2 %. Hossein Askari and Shohreh Majin, “Recent Economic Growth in Iran,” Middle Eastern Studies 12, no. 3 (1 October 1976): 105.

306

Chapter 4

disposed of a revolving credit of £18,000 per year to spend on machinery.50 Undoubtedly the magnitude of such orders had substantial leverage on the decision making of equipment manufacturers, and in the 1970s this new-found position translated into initiatives for typographical developments from these customers. To accompany the modernisation of machinery, a new visual language was sought and, as existing typefaces did not fulfil all the requirements, entirely new designs were commissioned from Linotype. Mitra and Nazanin, the Consolidation of a Persian Typographical Identity In April 1975 it emerged that Kayhan, a leading newspaper in Iran, was interested in purchasing a Linotron 303 photocomposition system (Figure 4.16– 18).51 The prospective order remained subject to Linotype-Paul providing the newspaper with a new typeface of its own choosing, demonstrating the importance assigned to the typographical voice of the paper. Considerations about appropriateness, aesthetics, identity and distinctiveness, hitherto secondary to more practical issues, became more important, and the call for their implementation was supported by economic arguments. At the same time, growing demand in the Middle East and numerous new developments for the increasingly competitive photocomposition market stretched Linotype’s manufacturing capacities. Against this background, Tracy assessed Kayhan’s request as uneconomical: I think it is unlikely that the typeface they ask for will be saleable outside of Iran. That is to say, we are unlikely to find any more than two or three customers for the design.

50

Goodman to Tracy, “Persian Faces.” In 2016, the relative economic power value of the order would be approximately £ 1,952,000, and the annual budget for typesetting equipment £ 616,400. Lawrence H. Officer and Samuel H. Williamson, “Five Ways to Compute the Relative Value of a UK Pound Amount, 1270 to Present,” MeasuringWorth, 2016, http://www.measuringworth.com/ukcompare/.

51

Walter Tracy to Richard J. Caesar, Letter, “Keyhan - Tehran,” 7 April 1975, WT correspondence, folder 18c/1 Persian, DTGC. The 303, announced in 1974, replaced the 505. It was based on the same principles, although in a slightly simplified and improved arrangement. It was able to generate type sizes from 4 to 72 pt, in increments of ½ pt, at an output speed of 150 newspaper lines per minute. Arthur H. Phillips, Handbook of Computer Aided Composition (New York and Basel: Marcel Dekker Inc, 1980), 384–89.

Persian Type and Typography

307

Figure 4.16

Components of the Linotron 303 system. From Phillips, Handbook of Computer-Aided Composition.

Figure 4.17

The character-generating system of the Linotron 303, showing the fount grid and the CRT. The machine could hold either 11 or 24 grids. From Phillips, ­Handbook of Computer-Aided Composition.

Figure 4.18 Fount grid of the Linotron 303 holding 144 characters. From Phillips, Handbook of Computer-Aided Composition.

308

Chapter 4

Any typeface takes many weeks to produce in drawing and frisket form, and our type drawing office is limited in its capacity. I do not think we should put into the department a type design which will not produce sales sufficient to justify the effort. It is better for us to spend our time on type development projects which have a wide application in sales terms.52 In another letter to his colleague E. Vesey, Tracy recommended to “steer” the client towards existing designs in order to avoid the making of a new Persian typeface.53 Yet, Kayhan did not accept any of the proposed types and insisted on a newly conceived Persian face. Having experienced the complicated and long-winded developments of Osman and Ettelaʿat Simplified, Tracy advanced a new strategy to satisfy the customer without the “costly and tedious business of commissioning a design from a local designer”.54 Thus, Linotype-Paul prepared two new proposals: a slightly modified version of Osman, and an entirely new design, conceived for the specific task at hand: a Persian newspaper face. Timothy Holloway, who had recently joined Linotype-Paul was given this task. The notion of a Persian typeface specifically conceived for newspaper composition is likely to have been influenced by another recent order from Iran. In 1973, Iran Chap ordered the new Linotron 505C for English, French and Persian composition.55 The machine was intended for book composition and the customer required a suitable typeface. Iran Chap therefore requested a new typeface to be made, and provided a model on which it should be based. At this time Linotype was in the process of finishing the Osman typeface (see page 243 and following) and, as this was conceived for book typography, it was sug� gested to the client. Linotype thus sent samples comparing the two typefaces to Tehran for evaluation. Iran Chap accepted Osman, yet made this subject to the revision of issues which had been noticed. The principal shortcomings pointed out in the client’s assessment were related to size. Osman had been designed with pronounced extenders, making its face appear small on the body. According to the client, 52

Tracy to Caesar, “Keyhan - Tehran.”

53

Walter Tracy to E. Vesey, Memorandum, “Keyhan - Tehran,” 30 April 1975, WT correspondence, folder 18c/1 Persian, DTGC. The proposals included Zapf’s 1956 Alahram typeface; the client’s disapproval underlines the aforementioned uncertainty about the alleged success and popularity of the design.

54

Walter Tracy to Martin Boothman, Letter, “Keyhan: New Typeface,” 16 July 1975, WT correspondence, folder 18c/1 Persian, DTGC.

55

E. Gabrielian to Messrs. Linotype, 13 December 1973, WT correspondence, folder 18c/1 Persian, DTGC.

309

Figure 4.19

Illustration of the maximum character area available on the 505C with a superimposed sketch of two characters of ­Osman. Linotype, November 1974, reduced to 75% linear. DTGC, WT correspondence, folder 18c/1 Persian.

Persian Type and Typography

the small counters of some characters filled-in during printing, decreasing legibility and contributing to an uneven appearance.56 Farhad Massoudi, the 56

Farhad Massoudi to Walter Tracy, Letter, 11 December 1973, WT correspondence, folder 18c/1 Persian, DTGC.

310

Chapter 4

publisher’s managing director, therefore proposed to increase the size of the typeface on the body, as well as the relative size of counters for a more successful rendering in small print on low-quality paper. But while Massoudi’s suggestions were in principle sound, they could not be implemented because the size of the face was determined by the phototypesetter. In reply to Massoudi’s concerns, Tracy provided a drawing to illustrate the dimensions within which the characters had to be accommodated (Figure 4.19 on page 309). Its height was divided into three zones: the main character area and space above and below for accents and deep descenders. All characters had to align on the fixed baseline to ensure different typefaces could be mixed in the same line of text. Therefore, the configuration of characters had to follow these proportions, with noticeable effects for the Persian design. As Tracy pointed out, Linotype had used the entire depth for descenders, “more, in fact, than in a normal roman type”.57 The difficulty, however, arose when trying to maintain the correct proportions of Arabic letterforms on a system conceived around Latin script characteristics: The space above the baseline is for the ascending parts of such letters as ‘h’ in Roman and ‫ ط‬in Farsi. The ascenders must, of course, be strictly in proportion to the descenders. A letter ‘h’ normally occupies nearly all the available space because its total height is about three times the depth of the descender. The proportion is quite different in the face you have for Farsi, where the ascender is about 1½ times the depth of the descender. Hence the small size of the overall effect compared with Roman faces. It is the natural proportions in relation to the fixed baseline which produces the visual result.58 Thus, the assumptions underlying the machine’s setup were based on characteristics of the Latin script and were not adequate for the differing features of Arabic. It follows that despite technical advances and the increased flexibility of photocomposition, the making of non-Latin type with different proportions and higher formal complexity remained problematic. ∵ 57

Walter Tracy to Farhad Massoudi, Letter, 27 November 1974, WT correspondence, folder 18c/1 Persian, DTGC.

58

Ibid.

Persian Type and Typography

311

When Kayhan commissioned Linotype in July 1975 to develop a new Persian newspaper typeface, the questions of relative size on the body and the fillingin of counters could be re-considered in a different context. Here an entirely new typeface provided the opportunity to conceive a design in response to the specificities of newspaper production with phototypesetters. The typeface could now be tailored towards its use in small sizes on absorbent newsprint paper, where the filling-in of counters and the smaller perceived size of Arabic type were among the most pressing concerns.59 Holloway was tasked to design this typeface and he proceeded quickly and successfully.60 By August 1975 drawings of the new typeface were sufficiently advanced to be presented to the client, and it was met with immediate

Figure 4.20 Character drawing for mīm, Mitra bold, n.d. reduced ‫م‬ to 25% linear. DTGC, drawings cabinet. 59

The proportions were unusual enough for Tracy to point them out to his colleagues at Stempel who were in charge of the grid-manufacture: “The proportions of the ascenders, descenders and middle range characters in this typeface is quite different from that in the Osman and Simplified Arabic design. The descenders are fairly short, and this has enabled us to put the design onto the standard z-line”. Walter Tracy to Rene Kerfante, Letter, “Keyhan Newspapers: Farsi Typeface for 303,” 18 December 1975, WT correspondence, folder 18c/1 Persian, DTGC.

60

Holloway developed the design in an unconventional way as the bold weight, inspired by Persian Nasḫ lettering, was drawn before the regular. “Mitra” (Linotype-Hell AG, n.d.), DTGC. According to Ross, defining the appropriate proportions of the more complex bold weight avoids difficulties that could arise if the regular was designed first. Today, Ross and Holloway still use this approach where applicable. Fiona Ross, email message to the author, “Re: Remaining Elements,” 30 August 2013.

312

Chapter 4

Figure 4.21

Mitra, proof (detail), composed on a Linotronic 300 in 18 pt, 1991, enlarged to 200% linear. DTGC, folder PostScript bromide proofs.

a­ pproval.61 A later specimen of the design states that the typeface featured “angular, clipped stroke-endings to give a brighter effect in print”, and that “for more compact setting, the descenders were made short and the fitting kept close” (Figure 4.21).62 These characteristics reflect the typeface’s intended use in newspaper composition: it had to be economical, clear and feature character shapes that withstood the conditions of newspaper printing during this period. The deliberate design of the characters is underlined explicitly in the specimen, noting that “for greater legibility, the normally small counters and teeth were enlarged” and that “particular attention was given to the design of certain characters to avoid filling in during printing” (Figure 4.22).63 These design principles mark an important change of approach in the conception of Arabic typefaces. In earlier Arabic type developments other factors stood in the foreground: the size of the character set, the technical implementation of kerning and the inclusion of vocalisation marks and their positioning were some of the most urgent concerns. All of these design problems arose primarily because of the physical characteristics of founts 61

Unfortunately no traces of the original artwork, nor proofs of the typeface development could be located in the collections of the DTGC.

62

Linotype-Hell AG, “Mitra.”

63

Ibid.

Persian Type and Typography

313

Figure 4.22 Character drawing of the ���‫ ش‬šīn �‫ �ي‬yāʾ ligature of Mitra, inked-in character superimposed over unit-grid, 1984 & 1986, reduced to 30% linear. DTGC, drawings ­cabinet.

and ­machinery. ­Indeed, the phrase “Arabic has to be put on [insert name of a typesetting machine]”, commonly used in internal documents of typesetting equipment manufacturers, speaks volumes for the primacy of the engineering challenge. In other words, the design of the typeface in the more narrow sense of conceiving specific, original character shapes had been secondary to the design of the fount as an element of a composition machine. But by the mid-1970s Arabic type-making and the limitations of equipment were well understood, and standards of Arabic composition became established, arguably even entrenched. As seen in the case of Plooij, innovation beyond these

314

Chapter 4

principles of Arabic ­type-making and composition was not acceptable for the dominant manufacturers. Gradual technical improvements were sought, but the market’s acceptance of the existing approaches gave little incentive for a re-evaluation of fundamental concepts. Here, typeface design in its narrow sense began to come to the fore. Clients no longer contented themselves with functional Arabic composition, but looked for visual distinction. The development of Kayhan’s new typeface exemplifies this change in priorities. Kayhan refused existing typefaces that would have provided the same functionality, and insisted on a new design which replied to a specific brief and conveyed the typographical identity the client intended.64 With the working title Linotron Farsi No.2 the new design for Kayhan was put into production in autumn 1975, and its completion was given the highest priority by Linotype-Paul. Because the generic name posed problems of ambiguity, the new typeface was called Mitra in December of the same year.65 In December friskets for the bold version were dispatched to D. Stempel AG for grid-manufacture, with the regular weight following in January 1976. The customer exerted substantial pressure for rapid completion of the typeface and by late February first tests were conducted at Linotype in the presence of an Iran Chap representative. These trials confirmed the client’s approval of Holloway’s design, but rejected the fount layout as it required too many changes between the four fount grids, rendering composition speed uneconomical. As a consequence, numerous elements of the system had to be revised and the delivery was delayed. Yet, the entire project was a success for both, Linotype-Paul and Iran Chap, as their continued and increased collaboration demonstrates. In January 1976, when the manufacture of Mitra was in full flow, the newspaper’s managers enquired about the possibility of an additional typeface commission. With its manufacturing capacities already stretched, this additional development posed further difficulties for Linotype-Paul. Kayhan was one of its most important customers in this period, and refusal or delay of the project could have had a detrimental effect on the business relation. Yet Linotype-Paul was not able to commit more resources, and Tracy advanced a different strategy. Contrary to his stance in relation to the client’s earlier commission, Tracy now appealed for cooperation by Kayhan in providing Linotype-Paul with the 64

Interestingly this new approach emerged in the context of commissions from Persianlanguage customers, a detail that may be indicative of variations of regional typographical preference. Whilst Arabic designs like Osman remained closely attached to the traditional forms of manuscript production, there appears to have been more leeway for interpretation and rationalisation in Persian typefaces of the same period.

65

Tracy to Kerfante, “Keyhan Newspapers: Farsi Typeface for 303.”

Persian Type and Typography

315

new design “in size and form suitable for immediate manufacture”.66 To facilitate this sometimes complicated process, Tracy briefed the client about the specifications of grid manufacture. In addition to a sample drawing, illustrating the exigencies of curve quality, the client was supplied with the following: (1) sheets of draughtsmen’s film material cut to size, and ready punched at the top; (2) a grid showing the unit system to which the characters must conform; (3) a pin bar on which the drawing material can be positioned.67 The suggestion to externalise the design process relied on sharing manufacturing methods, as the two processes were so tightly interwoven that the separation into individual tasks was problematic and time-consuming. Earlier projects had shown that designers who were not sufficiently familiar with the specificities of type manufacture failed to provide usable artwork, requiring long-winded corrections and adaptations.68 Linotype in turn provided Kayhan with high-quality drafting material, as well as a short guide instructing the designer about rules that had to be followed in producing new type designs for the Linotron 303.69 The notes for the designer contained general advice on type design and technical specifications with which the artwork had to comply. The technical advice was primarily related to the unit system and character alignment, arguably the least intuitive and most constraining aspects of the composition equipment at that time. Drawings of individual characters had to be fitted within the available units, and, in case of joining characters, had to conform to overlap and fitting requirements. When Tracy provided these instructions, Kayhan again commissioned Haghighi, the designer of Ettelaʿat Simplified, to create the new typeface that 66

Walter Tracy to E. Vesey, Letter, “Keyhan Newspapers Ltd - Tehran,” 28 January 1976, WT correspondence, folder 18c/1 Persian, DTGC. Tracy might have considered this procedure because the previous experience made with Ettelaʿat’s Simplified design proved the skill of their in-house designer: “If the work is entrusted to Mr. Haghighi, or an artist of similar distinction, we can be sure that he will find no difficulty in producing the characters in that way.” Ibid.

67

Tracy to Vesey, “Keyhan Newspapers Ltd - Tehran.”

68

Hitherto the terms ‘calligrapher’ or ‘artist’ were commonly used at Linotype, implying more clearly the removal from manufacturing processes and technical concerns.

69

Walter Tracy to E. Lachini, Letter, “New Type Designs for the 303 System,” 18 February 1976, WT correspondence, folder 18c/1 Persian, DTGC.

316

Chapter 4

was to be used on the Linotron 303 in addition to Mitra. Work started in the course of the year and a meeting between the designer and Tracy in September 1976 proved that despite the guidance, the principles of character design in accordance with predefined unit widths proved problematic.70 Summarising the meeting in a letter to the client, Tracy emphasised the importance of the unit system for computerised justification, and expressed his concern that the artist “does not seem to understand why all characters must be designed to a known number of units” (Figure 4.22 on page 313).71 Tracy therefore reiterated the principles governing the character drawings: a/ If the character is to join others, at one side, or both sides, he must draw it to the edge of the unit area, and then add a little to ensure a precise join. b/ If a character is to have space on one side or both sides he must not draw beyond the dotted line on the grid sheet we supplied.72 Haghighi revised his drafts and by January 1977 the drawings were ready for manufacture. Documentation of the production ends in March of the same year, suggesting that all obstacles had been overcome and the founts delivered to the client in due course. The entire process of this type development thus only took between 12 and 18 months, significantly shorter than the production times of comparable earlier projects. This increase in speed is indicative of a changing industry in which delays known from the hot-metal era were no longer acceptable. The acceleration of processes entailed a substantial reduction of documentation and correspondence which may explain the lack of drawings and proofs of the typeface’s development. Moreover, the new design had no proper name yet, with Linotype records referring to it either as the New Kayhan Face or simply Haghighi. This was only to change in a later revision of the design and its adaptation to new typesetting technology in the 1980s and 90s, when the typeface was called Nazanin after the granddaughter of the newspaper owner (Figure 4.23 on page 317).73 70

No visual records of the original development of the typeface could be found in the collections of the DTGC. As the artwork was provided by Haghighi, it was probably returned to the designer after fount manufacture.

71

Walter Tracy to E. Lachini, Letter, 11 October 1976, WT correspondence, folder 18c/1 Persian, DTGC.

72

Ibid.

73

As recalled by Fiona Ross in a conversation with the author, 10 November 2012.

Persian Type and Typography

317

Figure 4.23 Nazanin, specimen, Linotype Library, n.d., reduced to 60% linear. DTGC, L­ ocker C1.

318

Chapter 4

Mitra and Nazanin became two of the most popular and widely used typefaces in Iran, lastingly influencing the visual culture and typographical identity of Persian design. Today the Kayhan newspaper still uses both typefaces prominently in its printed edition, and the acclaimed poster art from Iran frequently employs both designs.74 Their success can be related to various, mutually beneficial factors. The publishing house Iran Chap had the vision and economic means to commission new designs, insisting on the importance of typographical voice and cultural identity. Although Linotype was initially reluctant to develop new type, Tracy eventually recognised the value and commercial prospect of designs conceived for a specific market. Collaboration between the client, giving input on stylistic matters, and Linotype, providing the infrastructure and technical expertise, proved decisive for the successful completion of the new Persian typefaces Linotype developed in the 1970s.75 The role of Haghighi as the designer of Ettelaʿat Simplified and Nazanin was significant for the consolidation of Persian typeface design. Holloway’s Mitra, in turn, demonstrates that design with sensitivity to cultural preferences is not the exclusive domain of the native artist.76

74

As typeface piracy is rampant in the Middle East, many instances of cloned digital fonts are used instead of the originals. Judging from the downloadable PDF documents on the newspaper’s website, at the time of writing Kayhan is among the users of freely available versions of Mitra and Nazanin which are probably illegitimate copies of the originals. Moreover, Kayhan’s website currently employs pirated fonts of the Nassim typeface, developed by the author for the online presence of the BBC World Service. “Rūznāmah Kayhān”, n.d., http://www.kayhan.ir/ (accessed 16 May 2017).

75

In the 1980s, as a part of Linotype’s continued policy of collaborative work, Haghighi was again involved in the adaptation of Nazanin for digital composition.

76

Today, although both typefaces remain popular in Iran, their continued use may also reflect a paucity of choice, and the lack of new original developments. Whereas the 1970s marked a particularly active time for Persian type-making, the political and societal upheaval of the Islamic revolution in 1979, and the subsequent First Persian Gulf War of 1980–88 certainly disadvantaged further typographical activity. Recently, Persian type design appears to experience renewed interest from a new generation of designers.

Chapter 5

Beginnings of Digital Arabic Type Type Loses its Material Properties The evolution of typesetting machinery in the 1960s and 1970s brought about the increasing use of computers to control tasks of the composition process. From the simple, hard-wired logic employed by the Justape, to the advanced capabilities of the programmes written for the mini-computers in the V-I-P and the Linotron 505C, technological leaps happened within less than a decade. The rise of computers in typesetting caused a synchronous reduction of mechanical and eventually photomechanical aspects. Whereas the V-I-P employed a computer to facilitate composition, it still generated type with direct photography. Character images were stored as negative film strips and projected on to photosensitive material. Its contemporary, the Linotron 505C also used photographic masters but reproduced the character images on a CRT, rather than directly on film or paper. Although in this process, as Southall notes, “the character image is dissected electronically to produce the information from which the output image is constructed, and the pulses of voltage produced when the matrix is scanned are shaped in the machine’s internal circuitry so that the spot of light on the writing CRT is either on or off”, it did not constitute a “binary mode representation of the image”.1 The 505C and similar ‘scanned-matrix’ machines could thus be described as hybrids, blending photomechanical and electronic technologies. In a subsequent evolutionary step the remaining photomechanical elements were abandoned. The next generation of machines no longer held the character images in the form of film matrices, but stored them as digital representations. In these devices type no longer existed as a physical entity, but became entirely immaterial, a specification rather than a visible, tangible artefact. Digital fonts stored information that specified the intended appearance on the output material, but had lost all material properties. As Southall summarised: This breakdown of the mechanical link between type and image, coupled with the shift from relief to planographic printing surfaces that came with the rise of offset lithography, called into question all the conceptual 1

Southall, Printer’s Type in the Twentieth Century, 150.

320

Chapter 5

Figure 5.1

Above left: Diagrammatic illustration of character-generation by strokes. From Otmar Hoefer, Schriftqualität auf dem PC (Bad Homburg: Linotype Library, 1988).

Figure 5.2

Right: Comparison of output from Linotron 404 (top) and Linofilm V-I-P ­(bottom), enlargement to 240% linear. Note the jagged edges on the 404 sample. From Seybold, “Digitized Type: What Is It? What Does It Mean for Typesetting and Word Processing?”.

Figure 5.3

Below left: Bitmap edit sheet, marked up for corrections of the scanned i­ mage. From Charles Bigelow and John W. Seybold, “Aesthetics vs. ­Technology Part II,” The Seybold Report on Publishing Systems 11, no. 11 (8 February 1982).

and dimensional frameworks on which five hundred years of received wisdom about type and typography had been built.2 2

Ibid., 140.

Beginnings of Digital Arabic Type

321

In the first fully digital devices, an electron beam ‘painted’ the characters in a sequence of vertical, overlapping strokes on the raster of the output tube. The character specifications consisted of coordinates at which the beam was switched on and off as it traversed the grid, thus creating the character image that was then projected on the photosensitive material (Figure 5.1).3 Because length, spacing and angle of the strokes could be altered, a single master could be used for a wide range of sizes and the slanting and distorting of characters. Although this provided unprecedented flexibility, it also entailed a number of compromises in quality, the most noticeable being issues related to resolution. When a master specification was scaled up beyond a certain limit, in effect reducing its resolution, the edges of characters became uneven and showed the typical saw-tooth effect of early digital output. Moreover, the frequency of the stroke – higher for small sizes and lower for big sizes – resulted in different light intensity, making small type look too dark and large type appear washed-out. Equipment manufacturers thus provided corrective logic to vary the beam’s intensity according to size. Initially, the quality of digital output was thus often inferior to the one obtained from photomechanical typesetting machines (Figure 5.2).4 Font manufacture also changed markedly. At the Linotype group the digitisation of character images was based on the process established for the manufacture of photomechanical founts.5 Character images were made as friskets, identical as for the V-I-P, and subsequently scanned at 3000 by 3000 elements per em.6 The scanned character images were then output on proofing printers for corrections by learned draughtspeople.7 In a process called ‘bitmap editing’ every character image was reviewed, and where applicable, marked for corrections, “indicating which superfluous or missing picture elements (pixels) 3

Seybold, The World of Digital Typesetting, 123–24.

4

The quality relied on multiple, often interconnected variables, not resolution alone.

5

Seybold noted that although “every manufacturer has its own techniques for digitizing type faces […] most have found that they have to go through similar processes.” John W. Seybold, “Digitized Type: What Is It? What Does It Mean for Typesetting and Word Processing?,” The Seybold Report on Publishing Systems 8, no. 24 (August 27, 1979): 12.

6

Ibid.

7

Seybold reported that some companies made corrections directly on video editing terminals, but at Linotype Mergenthaler hard copies were preferred as they enabled the draughtspeople to see characters side by side. Ibid. Later, when Linotype Ltd sent friskets of non-Latin designs to D. Stempel AG for fount manufacture, it became customary to also send staff with script specific expertise to instruct and guide the bitmap editor in the digitisation. Ross, interview.

322

Chapter 5

Figure 5.4

Hell-Digiset filmsetter. From Phillips, Computer Peripherals and ­Typesetting.

Figure 5.5

Bitmaps for the typeface Edison, designed by Hermann Zapf for a later generation of the Hell Digiset. Resolutions from left to right: 100 lines per em, 200 lines per em, 400 lines per em. From Bigelow and Seybold, “Technology and the Aesthetics of Type. Maintaining the ‘Tradition’ in the Age of Electronics.”

should be removed or added” (Figure 5.3 on page 320).8 Editing of the digital representations was then done on a visual display unit in a process which depended on the resolution the machine offered: 8

Ross, The Printed Bengali Character and Its Evolution, 197. Ross stated that 202 digitisations often used the same friskets as had been made for the V-I-P founts.

Beginnings of Digital Arabic Type

323

In general, there is an inverse relationship between the resolution of a digital raster and the amount of design processing necessary to obtain an acceptable letterform in that raster. That is, the coarser the raster, the more bit-map editing and redesign is necessary.9 The first commercial typesetting machine that implemented the step to a fully electronic character reproduction was the Digiset, developed by Dr.-Ing. Rudolf Hell GmbH in Kiel, Germany. It was announced in 1965, and first commercially installed in 1967 (Figure 5.4).10 The speed of the prototype Digiset was high at 600 characters per second, which, according to Wallis, was “the prime stimulus for shifting to digital CRT technology”.11 In other respects, however, the machine had considerable constraints. The built-in memory of 196 kilobits only held information of 180 characters up to a size of 14 pt because every size required a different image data set, consuming large amounts of expensive memory. The Digiset thus offered a typographical range too limited for most applications.12 Moreover, low resolution of the output resulted in jagged edges of characters, attracting negative reviews in the trade.13 These shortcomings, paired with a steep price restricted the device’s initial use to telephone directory and newspaper text setting (Figure 5.5).14 Pierre MacKay and Computer-generated Arabic Type From 1965 an agreement with Hell allowed the Radio Corporation of America to develop and market a device based on the Digiset in the United States.15 It was promoted under the name VideoComp and featured a small computer instead of the first Digiset’s hard-wired front end, making it an interesting

9

Charles Bigelow and John W. Seybold, “Technology and the Aesthetics of Type. Maintaining the ‘Tradition’ in the Age of Electronics,” The Seybold Report on Publishing Systems 10, no. 24 (24 August 1981): 15.

10

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 35–38.

11

Wallis, “From Frame to Desktop: a Century of Typesetting,” 75.

12

Southall, Printer’s Type in the Twentieth Century, 145.

13

Wallis, “From Frame to Desktop: a Century of Typesetting,” 75.

14

Later models improved on these limitations and were offered at lower prices.

15

From 1969 the company was known as RCA Corporation, and the RCA Graphic Systems Division was in charge of the development of the VideoComp.

324

Chapter 5

option for the more complex tasks involved in Arabic composition.16 In the early 1970s the University of Washington embarked on a project that aspired to develop a computer typesetting programme for Arabic. The Arabic Script Publications Project was led by the late Pierre MacKay, (1933–2015), professor emeritus of Classics, Near Eastern Languages and Civilization, and Comparative Literature. In an account of the project’s development, MacKay stated that the VideoComp was chosen after a survey of existing photocomposition methods. Because of “its software-generated typefaces, requiring no costly development of optical or electronic masks” the system appeared “ideally suited to Arabic script”.17 Similar to Plooij, MacKay realised that a division of Arabic letters into smaller elements than entire letters, and their computer-assisted assembly into words and text could result in a more efficient and more elegant system. In collaboration with fellow professor Walter Andrews, MacKay studied the script morphology and assembled “a repertory, not of characters, but of pen-strokes, which could be assembled into characters”.18 The rationale MacKay advanced was simple, yet stood in stark contrast to the established practise of the trade. He reasoned: A single font or set of character descriptions for the VideoComp can provide for only 128 distinct images, and although occasional changes of font will appear to be almost instantaneous, frequent changes would noticeably slow down text generation. By filling our ‘font’ with pen-strokes rather than entire letters, we are able to generate an almost limitless range of letter-shapes out of a mere 120 ‘characters’ in the VideoComp sense.19 Hitherto, the irreconcilability of limited character sets and authentic Arabic script rendering generally led to the latter’s simplification. But here a diametrically opposed approach was put forward. The simplification of the script no longer appeared necessary as the new means of photocomposition and computer logic allowed different solutions. Interpreted in a new way, the limited character set posed less of an obstacle, as hundreds of combinations could be obtained from a small number of efficiently designed pen-strokes. The

16

Seybold, The World of Digital Typesetting, 127.

17

Pierre A. MacKay, “Setting Arabic with a Computer,” Scholarly Publishing 8, no. 2 (January 1977): 145.

18

Ibid.

19

Ibid.

Beginnings of Digital Arabic Type

325

i­ mplementation of the system was made possible by the advent of more potent computers which could perform the required calculations at high speed. The system MacKay conceived thus took advantage of the new technology. It consisted of two complementary programmes to execute the composition tasks. In reference to Arabic terminology, the programmes were called Katib and Hattat, a reflection of their respective functionality.20 Whereas all pageformatting operations were executed by Katib, Hattat was designed to perform “the final task of selecting and assembling the appropriate pen-strokes and generating a tape to drive the VideoComp.”21 Besides the specific Arabic functionality, the system was unusual as the final tape contained the entire font information, making it independent from the fonts that were installed on the machine. This specific aspect of the system is noteworthy: while it was conceptually progressive, some of its assumptions were problematic. Type design did not seem to feature prominently in the investigations of the developers, and technical rationality substituted aesthetic sensibilities in MacKay’s system. For example, character spacing, a key aspect of quality typography, was left to approximation: Since Katib does not deal with the actual character shapes of Arabic script, it cannot use exact spacing values to determine the content of the eventual printed line, but must work with the average probable length of each letter.22 The system was put to a first practical test in April 1976 in the publication of a scholarly edition of Diocles’ On Burning Mirrors.23 An illustration of the text shows at once the severe aesthetic shortcomings (Figure 5.6 overleaf). Where the inexperience in typographical practice had the advantage of no conceptual bias in the design of the system, it was detrimental to the appearance of the text. The character drawings, proportions and spacing do not comply with the lowest of typographical standards. In stark contradiction to the novelty of the means, here the final result was inferior to the earliest Arabic typefaces on Linotype hot-metal machines.

20

kātib (scribe / clerk), ḫaṭṭāṭ (calligrapher).

21

MacKay, “Setting Arabic with a Computer,” 146.

22

Ibid., 149.

23

Diocles, On Burning Mirrors, ed. G. J. Toomer, Sources in the History of Mathematics and the Physical Sciences 1 (New York: Springer Verlag, 1976).

326

Figure 5.6

Chapter 5

Page 20 from On Burning Mirrors, reduced to 90% linear. From MacKay, ­“Setting Arabic with a Computer.”

Beginnings of Digital Arabic Type

327

No further use of the system was made as the programmes were written for the specific computer the VideoComp employed, and thus became obsolete once the hardware was decommissioned.24 MacKay, however, remained actively involved in the development of early computer typesetting systems and contributed to the development of the Arabic functionalities of TeX, the software typesetting system developed by Donald E. Knuth.25 Although TeX attained an important role in pioneering principles of desktop publishing, its impact on Arabic typography remained marginal. Occasional use remained largely restricted to academia, and the various developers’ focus on solving issues related to computing, rather than typographical excellence, appear to have perpetuated problems already apparent in MacKay’s work for the Video­ Comp. Design quality of Arabic TeX output never achieved parity with other, more typographically informed systems, which may have contributed to its ongoing niche-position. The KITAB Software for the Hell Digiset Apparently unrelated and uncoordinated, another Arabic typesetting project was conducted at around the same time for the largely similar Hell Digiset machine in Germany. The context for this development was, however, altogether different from MacKay’s project. Here, the impetus to conceive a new Arabic typesetting programme came from the Ernst Klett printing and publishing house in Stuttgart, an established specialist of educational literature.26 In the mid-1970s Klett, probably against the prospect of a substantial order, became interested in Arabic textbook production. The house was an early adopter of 24

Pierre A. MacKay, interview by Dave Walden, 2007, http://tug.org/interviews/mackay-p. html (accessed 16 May 2017).

25

Donald E. Knuth and Pierre A. MacKay, “Mixing Right-to-Left Texts with Left-to-Right Texts,” TUGboat 8, no. 1 (April 1987): 14–25. There is a substantial body of literature about TeX and the related Metafont software. See for example Donald E. Knuth, TEX and METAFONT : New Directions in Typesetting (Bedford, Massachusetts: Digital Press and American Mathematical Society, 1979); Donald E. Knuth, The METAFONT Book (Reading, Massachusetts: Addison Wesley, 1986); Donald E. Knuth, The TeXbook (Reading, Massachusetts: Addison Wesley and the American Mathematical Society, 1986); Donald E. Knuth, Digital Typography, CSLI lecture notes 78 (Stanford, CA: CSLI Publications, 1999).

26

The Klett Group is still active today. “Klett Gruppe - The Klett Group - Dedicated to Education,” Klett Gruppe, n.d., http://www.klett-gruppe.de/home/the+klett+group/ dedicated+to+education.765.htm (accessed 16 May 2017).

328

Chapter 5

the Digiset, but at this time the machine had no Arabic typesetting capability and Klett therefore sought to acquire relevant expertise in the field. To this end the Interpart company, a Stuttgart-based affiliate of Klett, contacted Hrant Gabeyan in 1975, and engaged him the following year as its representative for the Middle East.27 Aware of Gabeyan’s work for Linotype’s character selection and Kashida insertion routines, in autumn 1976 Klett-Interpart tasked Gabeyan to help in the conception of an Arabic typesetting programme for the Digiset.28 This time, and in contrast to the attitude that prevailed at Linotype, Gabeyan recalls that he was given carte blanche as Klett-Interpart sought to develop the best possible product. In what appears to have been a project with substantial economic backing, Gabeyan could freely specify the features that he considered necessary for a competitive offer on one of the technologically most advanced typesetting machines of its age, the Digiset 400 T2.29 The Digiset 400 was a series of phototypesetters released by Hell in 1974. They had fully digital font storage and could produce type in up to five size ranges, and like Linotype’s V-I-P and 505C the new device featured a computer which facilitated advanced typesetting functions.30 In lack of existing Arabic type available for this machine Gabeyan requested the creation of a new typeface specifically for this project. Resorting to his established network in Egypt, he had no difficulty in identifying a suitable

27

Initially Gabeyan worked for both companies but left Linotype-Paul on March 1, 1978 (he continued his duties for Linotype & Machinery Ltd, the branch of the Linotype group which maintained its dwindling hot-metal sales, until 1979). This move was partly motivated by the significant differences in salary: what he received from Interpart was two and a half times the salary he had earned at Linotype. The substantial difference in salary could be read as indicative of the aggressive policies new companies employed to secure expertise and thus an entry to the market. Gabeyan emphasises that he was employed because of his established customer relations in the region. Gabeyan, interview.

28

Ernst Klett and Interpart were nominally two distinct companies. However, in the presentation of its Arabic typography programme reference is made to Klett-Interpart as the developer. For convenience this denomination is adopted here.

29

The project may even have benefitted from political backing: The first production for which Klett-Interpart used their new system was a large order of Arabic text books commissioned by the Deutsche Gesellschaft für Technische Zusammenarbeit GmbH (GTZ) (German Agency for Technical Cooperation, Ltd) within a cooperation agreement between the Kingdom of Saudi Arabia and the Federal Republic of Germany.

30

4–16 pt, 8–32 pt, 16–64 pt, 32–112 pt and 64–112 pt. Arthur H Phillips, Handbook of Computer Aided Composition (New York and Basel: Marcel Dekker Inc, 1980), 372.

Beginnings of Digital Arabic Type

329

Figure 5.7 Adly Boulos (left) and Hrant Gabeyan, Germany 1977. Courtesy of Hrant ­Gabeyan.

designer: Adly Boulos (Figure 5.7).31 Boulos, an Egyptian Copt, had been alAhram’s long-time principal lettering artist and appears to have been designing Arabic type at least once before this project.32 Following Gabeyan’s recommendation, Boulos was hired to design the artwork for a typeface suitable for the educational and scientific publications of Klett. In parallel to the design work Gabeyan, now in a leading conceptual role, directed Boulos, specified the character set covering multiple languages and ligatures, and guided the programming team led by Gerhard Lieser in the development of the software.33 31

ʿAdliī Būlūs.

32

From a note by Tracy, recording a conversation with Gabeyan, it emerges that Boulos had received 320 Egyptian pounds for the design of a simplified Arabic typeface for the Nebitype machine. Walter Tracy, handwritten note, 4 October 1966, WT correspondence, folder 18a Simplified Arabic (Al Ahram) Lakhdar Ghazal, DTGC. The Nebitype was a manual linecasting machine that operated with similar principles as the Ludlow Typograph. Because of its ability to cast large type sizes it was particularly popular for jobbing and display work.

33

Gabeyan, interview. Although Gabeyan was employed by Interpart, the company offered an additional payment for the consultation work on the system; asked for a price

330

Chapter 5

Figure 5.8 Left: The calligram for KITAB was also designed by Boulos. From Lieser, “Production of High Quality Arabic Texts on a CRT Filmsetting Machine.”

Figure 5.9 Right: Bitmap of the isolated character of ‫ ك‬kāf, Klett-Interpart KITAB system. From Lieser, “Production of High Quality Arabic Texts on a CRT Filmsetting Machine.”

For example, taking advantage of the advanced options of the Digiset, Gabeyan specified the substitution routines for contextually varying tooth characters, as well as mark positions which took the dimensions of the character that they modified into account.34 Detailed information about the further development process is lacking, but according to Gabeyan, the development had been completed and was unveiled at a function in Stuttgart in autumn 1977. By April 1978 Lieser presented a paper about the project at the 5th Symposium on ‘Computers in Literary and Linguistic Research’ in Birmingham, UK, referring to the new Arabic typesetting system for the Digiset with the acronym KITAB: Klett-Interpart Typography for for his work, Gabeyan did not know how much to charge. Eventually, a Lebanese printer to whom Linotype had sold a 505C three years earlier – possibly Dār al-Qalam – provided Klett-Interpart with a quote for the same development work. Klett-Interpart then offered the same amount to Gabeyan, in his recollection “hundreds of thousands of Deutschmark”, which he gladly accepted. Ibid. 34

Gabeyan, “Miscellaneous”, email message to the author.

Beginnings of Digital Arabic Type

331

Arabic Basis DP-Program.35 (Figure 5.8) In Lieser’s presentation of KITAB he explained its development within the wider historical picture and arrived at an analysis that echoed notions Plooij had articulated earlier. Referring to the contextual variations of Arabic letters, he noted that “the exact shape of every character depends upon its position in the word, the graphic or blank which precedes or follows”. In contrast, he argued, Arabic type changed this “harmony of each line”, as “the rectangular lead types of a relatively limited character set were not suitable for context sensitive scripts”. Lieser’s conclusion of the script’s typographical evolution was unambiguous in its judgement: Arabic script has been treated so far in typesetting as the sequential putting together of single graphic forms instead of an unbroken stream of joined characters as they are produced by a calligrapher’s reed. The technical constraints of typesetting have forced so many Arabic readers to accept poor typefaces and unaccented texts and resulted in a noticeable degeneration of the former elegance of Arabic script.36 Lieser further identified a range of compromises and shortcomings which technical constraints imposed on Arabic typography. He listed simplification, incorrect positioning of vowels, and the indiscriminate elongation of letters as characteristic of the poor quality of Arabic type, and then continued to explain how KITAB overcame these issues. The software automatically selected contextual variations of characters, formed ligatures, inserted Kashida elongation strokes in appropriate positions, placed vowels on various heights, depending on the proportions of the base character, and added elongations to accommodate vowels in narrow contexts. According to Lieser, the new typeface was created with consideration of the “principles of character generation by the CRT of the Digiset”, yet this was not further explained.37 (Figure 5.9) The KITAB system found its first use in a series of vocational training books published jointly by the Ministry of Education of Saudi Arabia and the Ernst Klett publishing house. An example of this series is found in Tiknūlūǧīā al35

Gerhard Lieser, “Production of High Quality Arabic Texts on a CRT Filmsetting Machine,” Advances in computer-aided literary and linguistic research: proceedings of the Fifth International Symposium on Computers in Literary and Linguistic Research held at the University of Aston in Birmingham, UK from 3–7 April 1978, Department of Modern Languages, University of Aston in Birmingham, 1979: 1. The acronym also referenced the Arabic word for book: kitāb.

36

Lieser, “Production of High Quality Arabic Texts on a CRT Filmsetting Machine,” 4–5.

37

Ibid., 9.

332

Chapter 5

Figure 5.10

Above: Page composed with Klett-Interpart KITAB system, Tiknūlūǧīā alKahrabāʾ, 159, reduced to 45% linear. Courtesy of Hrant Gabeyan.

Figure 5.11

Below: Detail of Figure 5.10, enlarged to 200% linear.

Beginnings of Digital Arabic Type

333

Kahrabāʾ (Figure 5.10).38 Here, the typeface designed by Boulos is used in two weights and multiple sizes, combined with a Latin sans serif typeface. It is a classically-proportioned typographical Nasḫ design, with a somewhat more zestful and energetic appearance than Monotype’s Series 549, its principal competitor for quality book typography. This is achieved through compact letterspacing (by contrast the word spacing seems unnecessarily loose), a large number of ligatures and dynamic individual typeforms such as the initial and medial ‫ ك‬kāf. Gabeyan’s influence is apparent in the inclusion of contextually raised teeth characters, which add liveliness and clearly distinguish similar letterforms. The typeface, however, also has numerous shortcomings, perhaps because Tiknūlūǧīā al-Kahrabāʾ was one of the earliest publications using it and at this stage the type and the composition system had not been tested enough. The secondary weight lacks darkness to distinguish it sufficiently from the text weight, and could only qualify as a semi bold. Word spacing is too loose, and while most letter combinations are tight, some pairs suggest an apparent lack of kerning. A related problem occurs when diacritical signs come uncomfortably close to adjacent characters, as for example when a tooth with dots above is preceded by a ‫ ك‬kāf – a problem that should have been addressed by the software. On the level of letter drawings some characters such as the final ‫ لا‬lām ālif  ligature appear idiosyncratic and stand out on the page, while some marks are too light and lack definition (Figure 5.11). Nonetheless, compared to the contemporaneous Badr typeface composed on the 505C with similar software support, the improved placement of vowels is apparent: no marks are trailing the characters they modify and their positions relative to base characters are generally better. Overall the type appears promising but unfinished, an impression that is confirmed by Gabeyan: the Kitab typeface has several shortcomings, as we didn’t have time to make various tests and modify […] Klett was happy to have finally an Arabic typeface, jumped at the chance and used it immediately as it was, and then I left, and since, I don’t think that the typeface has been much modified, if they still use it.39

38

Tiknūlūǧīā al-Kahrabāʾ, The Ministry of Education, Saudi Arabia, Ernst Klett, Stuttgart, and Interpart, Stuttgart, 1979. The author is indebted to Hrant Gabeyan for sharing this source.

39

Hrant Gabeyan, email message to the author, “Writing up/Compugraphic,” 6 October 2016.

334

Chapter 5

Excepting these drawbacks, the typeface designed by Boulos served its purpose well. Some details lend it liveliness, whilst its overall appearance is conventional and unobtrusive, suitably so for its intended use in schoolbooks and other educational material. In KITAB it complemented a powerful typesetting system, and the volume Tiknūlūǧīā al-Kahrabāʾ – which was translated and produced by Interpart – is testimony to the high quality standards of the companies involved in its development. The difference in approach between MacKay and Klett-Interpart is marked, and it is indicative of changes that technological progress brought to the trade. The new methods allowed individuals and new companies without background in typography to venture into the development of new typesetting devices without the prohibitive costs of earlier machinery. MacKay, by training a classics scholar, typified a newly emerging kind of developer who explored the new technology’s potential for his own typographical needs. No longer was a slow, labour-intensive and expensive workflow required, but it became feasible for individuals without relevant training in the field to explore typography and type-making in a non-professional context.40 But although the unbiased position of the newcomer provided opportunities for innovative approaches to typesetting and type-making, it came with the risk of amateurism. In the case of MacKay it appears that the emphasis on the computer’s role overshadowed equally relevant aspects of typographical composition, much to the detriment of the final result. By contrast, Klett-Interpart, an established publishing house, appreciated the need to obtain external expertise in a domain in which it had little or no experience. The hiring of Gabeyan and the uncompromising use of resources for the system’s development demonstrate that the publisher was acutely aware of the value expertise warranted. A comparison of the results of the two projects shows the importance typographical training and experience retained in the face of radical technological change. Whereas MacKay’s development had pioneering qualities, its technological innovation could not remedy its typographical, aesthetic and cultural shortcomings. Klett-Interpart, on the other hand, combined progressive use of technological advances with an appreciation of typographical conventions and traditions, in a sound, if somewhat conservative approach to quality typography.41 40

MacKay’s work relied, however, on the substantial computer infrastructure a major US university could provide at the time, an asset which remained beyond the reach of individuals without institutional affiliation.

41

The author’s attempts to find further documentation about Klett-Interpart’s Arabic typography were unsuccessful. Although Hermann Hörl, the project manager for KITAB was

Beginnings of Digital Arabic Type

335

Compugraphic’s Venture into Arabic Type A few years after the completion of KITAB, Hrant Gabeyan once more played an important role in the evolution of Arabic type-making and typesetting technology. In 1977 the Compugraphic Corporation introduced the EditWriter 7500, a competitive direct-entry phototypesetter. Its hardware consisted of a keyboard, a 15-line video screen, two floppy drives and crucially, two Intel 8080 microprocessors, enabling potent editing features (Figure 5.12). This configura�tion allowed for the parallel operation of the photo-unit and the editing terminal which contributed to the instant success of the machine.42 In early 1978 Rory Cowan, International Product Specialist of Compugraphic, went to Cairo on a fact-finding mission for the company’s planned entry into Arabic photocomposition. Amongst other visits, Cowan met Gabeyan, who had still been known at the company because of his work for the character selection and Kashida insertion routines Linotype had developed jointly with Compugraphic more than ten years before (see page 187 and following).

Figure 5.12

The Compugraphic EditWriter 7500, video terminal, keyboard and floppy drives in the foreground, photo-unit in the background to the left. Courtesy of JeanMax Rocanin-Borraz.

contacted, little additional information could be obtained. Later Hörl and colleagues who had been involved with the development of KITAB founded the company Tradigital. Here, a substantial textbook project for Saudi Arabia provided the impetus to develop further tools for high-quality Arabic typesetting. Hermann Hörl, email message to the author, “AW: Hintergrund Und Geschichte von Tradigital,” 13 March 2013. Today Tradigital is still active in the field. Tradigital Stuttgart GmbH, “TRADIGITAL - Philosophy,” 2012, http://www.tradigital.de/about.htm (accessed 16 May 2017). 42

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 55.

336

Chapter 5

Figure 5.13

Video screen of the EditWriter 7500. Above full view of the screen, below detail of the status line. Compugraphic Corporation, Manuel d’utilisation: ­Photocomposeuse CG 7500, not dated.

From this meeting emerged a collaboration in which Compugraphic engaged Gabeyan to help in the development of new Arabic composition software for the company’s range of photocomposition machines.43 After a lengthy administrative process Compugraphic and Interpart, then Gabeyan’s employer, agreed to a four-day consultancy against a fee of 10,000 Deutsche Mark, with every additional day charged at 1,000 USD.44 Gabeyan 43

Rory Cowan to Hrant Gabeyan, Letter, 16 February 1978, Private collection Hrant Gabeyan.

44

Holger Brée to Rory Cowan, Telex, 30 June 1978, Private collection Hrant Gabeyan. KlettInterpart had initially quoted 30,000 DM for a three-week consultancy. Compugraphic eventually decided not to require so much time and requested for the fee to be adjusted accordingly.

Beginnings of Digital Arabic Type

337

thus visited the Compugraphic headquarters in Wilmington, Massachusetts, from Thursday 13 to Tuesday 18 July 1978. In preparation for the meeting, Compugraphic staff developed a comprehensive agenda, detailing, amongst others, the following discreet tasks for Gabeyan: finishing of the keyboard layouts and fount schemes including mark placement, ligature selection, and the status line (Figure 5.13), review of the type samples, design of the display characters for the video screen, and review of preliminary product specification, including the repeated review of all elements.45 In a densely packed visit, Gabeyan was thus briefed about the specifications of the machinery and the company’s needs and expectations, and then set-out to develop a comprehensive document. It contained distinct sections for Traditional and Simplified Arabic, and sections for their respective ligature substitution routines, Kashida insertion rules, as well as three discreet keyboard layouts for different regions using the Arabic script. The resulting “Arabic Documentation Package” became the basis for Compugraphic’s Arabic composition software (Figure 5.14 overleaf). But whereas Gabeyan was central in the planning of the founts, nothing in the documentation suggests that he was much involved in the design and revision of the artwork. Whilst his initial estimate listed the selection and supervision of a calligrapher for the design of the typeface as one of the duties he would perform, this item was later dropped by Compugraphic, perhaps to reduce costs and save time.46 In Gabeyan’s recollection Compugraphic was pressed for time, seeking to release an Arabic product as quickly as possible, and therefore dispensed with the development of an original typeface.47 Instead, it appears that existing models from the competition were closely studied: in reply to an enquiry from Peter Schacht of Compugraphic, Gabeyan explained which characters of the planned fount scheme corresponded to 45

Steven Reef to Carey et. al., Memorandum, “Second Revision for Meeting with Hrant Gabeyan,” 14 July 1978, Private collection Hrant Gabeyan. The status line was the first line of the video display. It was divided into three areas, providing the operator with (1) information about the typographical setup (fount, type size, interlinear space, remaining space of the composed line, automatic or manual justification), (2) remaining memory available for further text input (either 200 lines or 6.000 characters in total), and (3) an indication for whether the machine is set to text entry or edit mode, in the latter the advance widths of characters and the remaining space in a line were not counted. Compugraphic Corporation, Manuel d’utilisation: Photocomposeuse CG 7500, not dated, digitised and edited by Jean-Max Rocanin-Borraz, 2015, http://rocbor.net/typo/PhotoCompo/ Compugraphic.htm, (accessed 16 May 2017).

46

Hrant Gabeyan to Rory Cowan, Letter, 5 March 1978, Private collection Hrant Gabeyan.

47

Hrant Gabeyan, email message to the author, “Writing up/Compugraphic,” 6 October 2016.

338

Chapter 5

Figure 5.14

Cover page of the “Arabic Documentation Package” Gabeyan compiled for the Compugraphic Corporation, reduced to 50% linear. Courtesy of Hrant Gabeyan.

characters shown in published Monotype and Linotype specimens, suggesting a considerable influence (if not outright copying) of existing typefaces.48 Over the course of the following months, and through continued correspondence, the software was developed and refined. By January 1979, Schacht 48

Hrant Gabeyan to Peter Schacht, Letter, 28 December 1978, Private collection Hrant Gabeyan. The typefaces that appear to have been studied were Monotype’s Series 589 and Linotype’s Yakout.

Beginnings of Digital Arabic Type

339

reported to Gabeyan that some additional ligatures could be implemented, and requested advice for 15 more characters, their selection routine and mark placement. From this letter it also emerges that Arabic fraction characters were included in the founts, indicating a typographically advanced product. Indeed, when the EditWriter 7800 was announced in 1979, it provided powerful composition features.49 In addition to the capabilities of the Latin phototypesetter, the 7800 offered two Arabic typesetting programmes and typefaces, a Simplified and a Traditional design. The Traditional fount scheme clearly showed the influence of Gabeyan’s consultation: it comprised more than 350 characters, including numerous stylistically preferable ligatures and contextual variants that were automatically selected, and marks were automatically positioned on one of four levels that reflected the height of the typeform.50 Furthermore, the EditWriter 7800 allowed for the mixing of Arabic and Latin script type in one line, both displayed on its video screen, and justification was fully automated.51 The success of Compugraphic’s Arabic offer is not known, but it may be surmised that the low price of the phototypesetters in conjunction with their technological prowess might well have given Compugraphic a sizable share of the market.52 Whilst the typefaces that were shipped with the EditWriter were derivatives of models by Linotype and Monotype, later additional designs, notably by Ahmed Lakhdar Ghazal, were added to its library. The influence of its first two Arabic typefaces, however, eventually went well beyond the company’s potential reach in phototypesetting, as will be discussed later (see page 409 and following). 49

The EditWriter 7800 was essentially an Arabic-enabled version of the 7700, released in 1978. According to Wallis, the 7700 was an upgrade of the 7500 that provided an increased composition speed of 40 characters per second. Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 56.

50

A recent account of the development of the EditWriter 7800 does not mention Gabe­ yan’s involvement. Boutros et. al., Talking About Arabic, 42–43. Beyond its testimony to the shortcomings of existing literature on the history of Arabic type-making, Boutros et al’s account illustrates the recurring issue of unmerited claims of authorship or the concealment of originators of key contributions in the field. The relevant sections of the publication were probably written by the listed contributor Steven Reef, a former Compu­graphic employee.

51

Compugraphic Corporation, “Coming soon!,” announcement from promotional material, not dated. Private collection Hrant Gabeyan.

52

The company’s disregard of intellectual property rights in type design is widely acknowledged, and probably extended also to its Arabic founts. See Savoie, “International crosscurrents in typeface design: France, Britain and the USA in the phototypesetting era,” 193.

340

Chapter 5

Digital Tools in the Service of the ‘Hanging’ Styles Nastaʿlīq and Ruqʿä By the late 1970s the trend towards fully digital typesetting became increasingly clear. In August 1979 The Seybold Report on Publishing Systems dedicated an entire issue to digitised type where Seybold, in recognition of the changing technology, articulated a forecast for the trade: We believe that we are in the midst of a transition from phototypesetting to digital typesetting. It will impact everyone from the smallest direct input user to the largest newspaper. Full page output of text and graphics will be done in digital form. High speed output is being done and will be done digitally. Low quality output will be done in digital form. Direct output to paper and direct output to plate will also be done in digital form.53 Digital typesetting had advanced substantially since Klett’s Digiset was introduced. Whilst initially exclusive to the domains of high-speed and high-volume printing, digital machinery began to permeate all areas of the trade from the second half of the 1970s. Linotype’s first digital CRT machine, the Linotron 606, was released in 1975 and was still designed as a “top-of-the-line model” which remained exclusive because of its high price.54 In 1978 the range of digital devices was enlarged with the introduction of the Linotron 404 and the Linotron 202, the latter a device “which attained pre-eminence in its market sector” (Figure 5.15).55 Whereas the 606 and its successor the 404 were designed as replacements for the high-end model 505, the 202 replaced the V-I-P which was popular with quality trade typesetters.56 Although slower than the 606 (which boasted up to 1050 lines per minute), the 202’s 450 lines per minute 53

Seybold, “Digitized Type: What Is It? What Does It Mean for Typesetting and Word Processing?,” 1. Although digital composition had been in use for a few years, here it appears that the concepts and notions of the new technology were still sufficiently novel to merit detailed explanation of resolutions, character specifications and the processes employed in digitising characters.

54

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 52.

55

Ibid., 56.

56

At under $50,000 this high-speed CRT machine was positioned in the same price range as slower second generation devices, thus contributing to its success in that market. Seybold, The World of Digital Typesetting, 134. The relative real price of the machine in 2015 would be approximately $ 182,000. Lawrence H. Officer and Samuel H. Williamson, “Seven Ways to Compute the Relative Value of a U.S. Dollar Amount - 1774 to Present,” MeasuringWorth, 2017, http://www.measuringworth.com/uscompare/ (accessed 16 May 2017).

Beginnings of Digital Arabic Type

Figure 5.15

341

Schematic diagram of the Linotron 202. From Phillips, Handbook of ­Computer-Aided Composition.

still remained substantially faster than second-generation phototypesetters.57 An important difference of the 202 lay in the way in which fonts were stored. Where earlier digital machines stored character specifications as dot patterns, a memory-intensive approach that required multiple digitisations, the 202’s fonts were defined as character outlines (Figure 5.16 overleaf). These outlines constituted specifications of a “character shape without regard to the particular pattern of dots which will be required to reproduce that shape at a particular size and writing resolution.”58 Depending on the size of the output, the outline was thus used to generate corresponding dot patterns ‘on the fly’. The primary advantages of outlines lay in using a single master for every point size, the provision of a constant density of writing strokes, and the flexibility for font transformations by the user (size, slant, width, height). In contrast to the 606, which relied on four masters to generate all sizes, the 202 could ­therefore cover the entire size range from 4½ pt to 72 pt with a single master.59 The out57

Phillips, Handbook of Computer Aided Composition, 384, 392.

58

Seybold, “Digitized Type: What Is It? What Does It Mean for Typesetting and Word Processing?,” 12.

59

Phillips, Handbook of Computer Aided Composition, 384. Another advantage over some earlier systems was that output resolution was constant at 975 strokes per inch.

342

Chapter 5

Create

Create

Create

Destroy Destroy Destroy

Figure 5.16 Diagrammatic illustration of outline description of character image as used for Linotron 202 fonts. Adapted from Joe Condon, Brian Kernighan and Ken Thompson, “Experience with the Mergenthaler Linotron 202 Phototypesetter, or, How We Spent Our Summer Vacation,” Digital restoration and typesetter forensics, 13 December 2013, (accessed 16 May 2017) http://www.cs.princeton. edu/%7Ebwk/202/.

line fonts required little memory and could be stored on cheap floppy discs, each holding 30 different fonts.60 Yet, the first outline formats also imposed their own constraints. As no curved outlines could be specified, character images were made from straight-line segments which approximated the curves of letterforms. As Ross explained, “with the breakdown of each letterform into discrete elements, the edge quality of the characters suffered; at large sizes, the straight-line segments became visible on curves and diagonals” (Figure 5.17).61 The spread of digital typesetting equipment reached one of the lowest market segments when in 1979 the German Linotype GmbH introduced the ­CRTronic. It featured a keyboard, a small video editing screen, two floppy disc drives, and was the first small and inexpensive direct-entry phototypesetter to employ CRT technology (Figure 5.18). It was described as a “table-top unified small typesetting system”.62 With a resolution of 1270 lines per inch, and the ability to create 46 sizes from 4 pt to 36 pt in 0.7 pt steps, the CRTronic “offered the greatest versatility of any direct-entry typesetting machine” when it was introduced.63 The device became an immediate commercial success with over 1,000 unit sales within the first one and a half years.64

60

According to Ross, this was a conservative estimate; in practice around 60 fonts could be held by one disc. Ross, The Printed Bengali Character and Its Evolution, 197.

61

Ibid., 197–98.

62

Phillips, Handbook of Computer Aided Composition, 392.

63

Seybold, The World of Digital Typesetting, 136.

64

John W. Seybold, “Mergenthaler Linotype’s CRTronic: A Review and Status Report,” The Seybold Report on Publishing Systems 10, no. 8 (22 December 1980): 1.

Beginnings of Digital Arabic Type

Figure 5.17

CRTronic output of Jalal, 1991, enlarged to 150% linear. DTGC, folder ­softdressing proofs.

Figure 5.18

Arabic CRTronic S, 1983. DTGC, folder Arabic CRTronic.

343

344

Chapter 5

Throughout the 1950s and 1960s, the Monotype Corporation lagged behind other manufacturers in its adoption of new technologies.65 Monotype continued its series of second-generation Monophoto machines throughout the 1960s and up to 1974 when the last Monophoto 400/8 was released.66 Although the Monophoto series was known for its “type of extremely good quality”, by the 1970s its limits of speed turned into a liability.67 At this point even thirdgeneration CRT machines had reached the end of their lifecycle as market interest began to shift to fully digital typesetting machines. It became clear that Monotype had to develop a competitive offer for high-speed composition. The Corporation did this by getting ahead of the trade with the introduction of a new and revolutionary technology. The Lasercomp, first released in 1976, no longer featured a CRT, but, as the name suggests, employed a laser to write out text and graphics to the medium (Figure 5.19).68 It did so in “raster scan fashion”, meaning the construction of all visual elements on a page by means of dots, or pixels, that are laid down by a sweeping movement of the laser.69 It thus qualified as what came to be called an ‘imagesetter’, distinguished from earlier photocomposition devices by the fact that “its unit of output is an entire page rather than a single character.”70 The Lasercomp was faster than any of Monotype’s earlier machines with an output of approximately 1,100 newspaper lines per minute at a constant, size-independent resolution of 1000 lines per inch.71 Digital fonts were loaded from floppy discs to the rigid disc that held a maximum of 300 fonts of 128 characters each, which could be output in sizes from 4 to 256 pt.72 Whereas CRT machines effected changes of type size by a change of resolution, the Lasercomp’s constant resolution required a different approach. Here every type 65

Boag noted that until the introduction of the Lasercomp, Monotype only “responded” to the new technological developments of competitors. Boag, “Monotype and Phototypesetting,” 72–73.

66

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 51.

67

Seybold, The World of Digital Typesetting, 386.

68

Seybold points out that it was not the first machine to use this technology, but it was the first that became a viable commercial proposition. Ibid., 386–87.

69

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 51.

70

Southall, Printer’s Type in the Twentieth Century, 161.

71

Boag explained that because the laser always swept the full page width it suffered speed drawbacks for galley composition: “it took as long to expose a single-column galley of 11 picas as it did to expose the full 58-pica page”. Boag, “Monotype and Phototypesetting,” 69.

72

Phillips, Handbook of Computer Aided Composition, 398.

Beginnings of Digital Arabic Type

345

Figure 5.19 Monotype System 3000 Lasercomp. From Phillips, Handbook of ComputerAided Composition.

size demanded a distinct font that provided descriptions of every character, consuming many of the available fonts.73 The Lasercomp has been described as “truly revolutionary” and “premature” respectively, indicating the novelty of approach.74 Its capabilities, notably the combination of text and images on a page and features like “the ability to set white text on a black background [and] to lay down tints”, contributed to the interest the machine received.75 Yet its success was hampered by multiple factors: whereas it was capable of outputting whole pages, no good graphics software that would facilitate the layout was available before 1979, and Monotype provided a graphics terminal only in 1982.76 Furthermore, its high-cost of £ 40,000, paired with the inflexibility of early models made the Lasercomp a proposition with little practical interest for large parts of the trade.77 But although its characteristics suggested applicability to the newspaper market, 73

Southall, Printer’s Type in the Twentieth Century, 163. Southall notes that this limitation only applied in the beginning of the technology, and that by 1984 models of the Lasercomp were able to generate the size-specific masters from a single outline master.

74

Boag, “Monotype and Phototypesetting,” 73, and Seybold, The World of Digital Typesetting, 390.

75

Ibid., 389.

76

Boag qualified this, stating that “even then these systems did not allow full page make up”. Boag, “Monotype and Phototypesetting,” 73.

77

The Lasercomp’s relative real price in 2011 would be approximately £ 236,200. Lawrence H. Officer and Samuel H. Williamson, “Five Ways to Compute the Relative Value of a UK Pound Amount, 1270 to Present,” MeasuringWorth, 2016, http://www.measuringworth. com/ukcompare/.

346

Chapter 5

“many key staff failed to appreciate the very different requirements of the new market Monotype now aimed itself at, and this resulted in ineffective marketing, wasteful internal power struggles and unhealthy lengthened mourning of the gradual decline in Monotype’s traditional market base.”78 Nastaʿlīq on the Monotype Lasercomp Against this background, the prospect of establishing a presence with an important newspaper would have been sought by the Corporation, and by the late 1970s an ambitious project in Pakistan provided the opportunity. Here, Monotype was approached by two authorities of “printing and the graphic arts”, Mr Ahmad Mirza Jamil, and Mr Matlubul Hasan Saiyid (Figure 5.20).79 Jamil was then managing director of Elite Publishers Ltd, Karachi, and the acting chairman of the Pakistan Association of Printing and Graphic Arts Industry; Saiyid had been the association’s chairman for over 12 years and was a recognised scholar and author. The two men had conducted a study in which they assembled a comprehensive lexicon of the Urdu language and collected numerous specimens of Nastaʿlīq calligraphy. According to Jamil Naqvi, while collecting the samples they “examined each one in detail, the lettering, word setting, plane and plinth, rise and fall, attraction, angularity, clarity, precision, inherent character, and shapeliness from every aesthetic and artistic aspect”.80

Figure 5.20 Matlubul Hasan Saiyid (left) and Ahmad Mirza Jamil (right). From Naqvi, “Nastaliq – the Elegant Urdu Script – Its Origin and Progress from the Sixth Century to the Present Day.”

78

Boag, “Monotype and Phototypesetting,” 73.

79

Jamil A. Naqvi, “Nastaliq – the Elegant Urdu Script – Its Origin and Progress from the Sixth Century to the Present Day,” The Monotype Recorder no. 3, New Series (October 1981): 4. Naqvi himself had been the librarian in the Government of India and the Muslim University in Aligarh, India.

80

Ibid.

Beginnings of Digital Arabic Type

347

Based on this corpus, it appears that Jamil and Saiyid tried to establish what amounted to a graphemic inventory of Urdu written in the Nastaʿlīq hand ­(Figure 5.21 overleaf).81 Naqvi noted that having gathered the samples, “composite blue prints of the integrated appearance and shapes were then prepared”, leaving considerable room for interpretation.82 Lacking further evidence, one may speculate that these “composite blueprints” were a synthesis of all samples, conceived as a generic, ‘ideal’ reproduction of the style. It is not documented what system was applied in arranging this inventory as Naqvi only refers to the aesthetic quality of the drawings, but it appears that semantic and graphic aspects were fused in a loose fashion. Thus, the writing style was not analysed as a language-independent sign system, but the Nastaʿlīq morphology and the Urdu lexicon were seen as a single, unified and, consequently less flexible, whole. This underlying assumption, in turn, had substantial influence on the approach chosen in the following steps. When Jamil and Saiyid reviewed the latest typesetting technology, Monotype’s new Lasercomp promised an opportunity to successfully implement Nastaʿlīq with typographic means, and the two men approached the Corporation with their concept. According to Graham Sheppard, who was to become Monotype’s employee responsible for the ensuing project, the proposed “idea seemed practicable”, implying that the Corporation agreed with its development.83 Presumably in search of financial support, Jamil and Saiyid then approached a leading Pakistani newspaper, The Daily Jang, seeking its cooperation in the project. Sheppard recalled that “the proposal was for a Nastaliq font that could be set on the Lasercomp, and which would resemble closely the existing text in The Daily Jang as written by scribes, and which would be set by straightforward keyboarding.”84 According to Sheppard, the idea convinced the newspaper, and the font was commissioned from Monotype. However, the company’s role was limited to the production only, as concept and artwork were provided by Jamil and Saiyid. While Sheppard went to Karachi to assist in 81

A plain text version of the corpus has recently been published. It provides insight into the principles that were applied in its compilation. Center for Language Engineering, “Valid Ligatures of Urdu,” 13 January 2011, http://www.cle.org.pk/software/ling_resources/ UrduLigatures.htm, (accessed 16 May 2017).

82

Naqvi, “Nastaliq – the Elegant Urdu Script – Its Origin and Progress from the Sixth Century to the Present Day,” 4.

83

Graham Sheppard, “Urdu Nastaliq, 1981,” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 31.

84

Ibid.

‫ ‪348‬‬

‫‪Chapter 5‬‬ ‫ض‬ ‫��ا�ل�� ض‬ ‫�س��ط�����‬ ‫ت �ی‬ ‫ض�ح���ع��ض ت‬ ‫ت�����ھ�‬ ‫ے‬ ‫چ‬ ‫گ�‬ ‫�ل��ھ��س��ٹ���ئ‬ ‫ت ھ�‬ ‫ے‬

‫����� ضٹ ض‬ ‫ض عت�������‬ ‫�ی‬ ‫����ع��ضٹت‬ ‫چ ت�����ھ�‬ ‫ے‬ ‫�‬ ‫ل���ع��ضس ض‬ ‫ت ��ح��ھ�‬ ‫ے‬ ‫چ‬ ‫گ�‬ ‫ل���ع��ض ت‬ ‫ت�����ھ�‬ ‫ے‬ ‫چ‬

‫�� تشض‬ ‫���س������‬ ‫�ی‬ ‫����ع��ضٹ ض‬ ‫چ ت�����ھ�‬ ‫ے‬ ‫�‬ ‫ل���ع��ضس ئ‬ ‫ت ��ح��ھ�‬ ‫ے‬ ‫چ‬ ‫گ�‬ ‫ل���ع��ض ض‬ ‫ت�����ھ�‬ ‫ے‬ ‫چ‬

‫�ض‬ ‫چح���ھ�س���� ض‬ ‫چ��ع���تا�‬ ‫�� ض‬ ‫�ت‬ ‫��‬ ‫چع�ت�� ا�������‬ ‫ت� ضس‬

‫گ� گ ض‬ ‫�ض‬ ‫�ض‬ ‫�‬ ‫ضح���ھ�س���� ٹ‬ ‫ل���ھ��ا�ل� ض‬ ‫ضح���ھ�س���� ض‬ ‫���چ ا���س�ت����ضلٹ‬ ‫��ع���تا�‬ ‫ض��ع���تا�‬ ‫ض��ع���تا�‬ ‫ت ��ر‬ ‫�‬ ‫ض‬ ‫گ�‬ ‫����ع���ضٹت‬ ‫��������� ات‬ ‫چح���ھ������ ات‬ ‫ل���ع���ض�ت‬ ‫�‬ ‫ت چ���ت���� ضس چع����ت����سض‬ ‫ض ع����ت���� ضس چ ت ���ت����سض‬

‫تض ض‬ ‫�����س������ع���‬ ‫�ی‬ ‫����ع��ضٹ ئ‬ ‫چ ت�����ھ�‬ ‫ے‬ ‫گ�‬ ‫�ل��ھ��س ٹت‬ ‫�ت����ھ�‬ ‫ے‬ ‫�ح‬ ‫چ���ھ���� ت‬ ‫چ��عا����ھ�‬ ‫ے‬

‫���� ض‬ ‫�‬ ‫ہ ت ا�ل��س� ا�ل�‬ ‫�ی‬ ‫�� ض‬ ‫�ئ‬ ‫��ع�‬ ‫چ ت�� ا����ھ�‬ ‫ے‬ ‫گ�‬ ‫�ل��ھ��س ٹ ض‬ ‫�ت����ھ�‬ ‫ے‬ ‫�ح‬ ‫چ���ھ���� ض‬ ‫چ��عا����ھ�‬ ‫ے‬

‫����ع��ضٹت‬ ‫ض ت�����ھ�‬ ‫ے‬

‫ض‬ ‫ٹ���ا �‬ ‫ت�ل��� ا�ل�‬ ‫ت �ی‬ ‫ض�ح���ع��ض ض‬ ‫ت�����ھ�‬ ‫ے‬ ‫چ‬ ‫ض�ح���ع�� ٹت‬ ‫�ت����ھ�‬ ‫ے‬ ‫چ‬ ‫����ع� ضٹ ض‬ ‫ض ت������ھ‬ ‫ے�‬

‫ل� ضت‬ ‫���س���‬ ‫ہ �ی‬ ‫ض�ح���ع��ض ئ‬ ‫ت�����ھ�‬ ‫ے‬ ‫چ‬

‫����ھ��س��‬ ‫چ چ���ھ�����‬ ‫�ی‬ ‫�ح� ض‬ ‫�ت‬ ‫��ع�‬ ‫ض ت�� ا����ھ�‬ ‫ے‬

‫����ع��ضست‬ ‫ض ت ��ح��ھ�‬ ‫ے‬ ‫چ‬ ‫�ح� ض‬ ‫�ض‬ ‫��ع�‬ ‫ض ت�� ا����ھ‬ ‫ے�‬ ‫ض‬ ‫������� ت‬ ‫ض��عا����ھ�‬ ‫ے‬

‫ض�ح���ع�� ٹ ض‬ ‫�ت����ھ�‬ ‫ے‬ ‫چ‬ ‫����ع� ضٹ ئ‬ ‫ض ت������ھ‬ ‫ے�‬

‫ض�ح���ع�� ٹ ئ‬ ‫�ت����ھ�‬ ‫ے‬ ‫چ‬ ‫�� ض‬ ‫�ت‬ ‫��ع�‬ ‫چ ت�� ا����ھ‬ ‫ے�‬

‫�� ض‬ ‫�ض‬ ‫��ع�‬ ‫چ ت�� ا����ھ�‬ ‫ے‬

‫ش‬ ‫����‬ ‫�ا�ل����لئ‬ ‫ت‬ ‫چ��ر‬ ‫����ع� ضٹت‬ ‫ض ت���������‬ ‫ت� ضس‬

‫ل ضض‬ ‫��م��ط����������‬ ‫ت� ضس‬ ‫�ح���ع� ٹت‬ ‫ض ��ت�������‬ ‫چ ت� ضس‬

‫�ح� ض‬ ‫�ت‬ ‫�‬ ‫ض�ع�ت�� ا�������‬ ‫ت�س‬ ‫ض‬ ‫��ا�ل���س�� ض‬ ‫ط�ت���و‬

‫�ح� ض‬ ‫�ت‬ ‫�‬ ‫چ�ع�ت�� ا�������‬ ‫ت� ضس‬ ‫ٹ‬ ‫ض������س�� ض‬ ‫ت �ت���و‬

‫ض���‬ ‫ے‬ ‫گ�‬ ‫���‬ ‫ے‬

‫����ع��ضست‬ ‫چ ت ��ح��ھ�‬ ‫ے‬ ‫چ‬ ‫�ح� ض‬ ‫�ض‬ ‫��ع�‬ ‫چ ت�� ا����ھ‬ ‫ے�‬ ‫ٹ‬ ‫����ع���� ٹ‬ ‫چ چ���ھ��ھ�‬ ‫ے‬

‫����ع��ضس ض‬ ‫ض ت ��ح��ھ�‬ ‫ے‬ ‫چ‬ ‫�ح� ض‬ ‫�ئ‬ ‫��ع�‬ ‫ض ت�� ا����ھ‬ ‫ے�‬ ‫ض‬ ‫������� ض‬ ‫ض��عا����ھ�‬ ‫ے‬

‫����ع��ضس ئ‬ ‫ض ت ��ح��ھ�‬ ‫ے‬ ‫چ‬ ‫�ح� ض‬ ‫�ت‬ ‫��ع�‬ ‫چ ت�� ا����ھ‬ ‫ے�‬ ‫ض‬ ‫������� ئ‬ ‫ض��عا����ھ�‬ ‫ے‬

‫�ل� ضت‬ ‫ض��ت�����ہ‬

‫ل�م����ت ض‬ ‫س��ح���ہ‬

‫����تس�ض‬ ‫�ح�ت�ل�ہ‬

‫����ع� ض ت‬ ‫ض ت�س��ح�����‬ ‫چ ت� ضس‬ ‫ض‬ ‫������ ٹ‬ ‫چ��ت ا�ل�ت����‬ ‫�‬ ‫ی‬

‫����ع� ض ت‬ ‫چ ت�س��ح�����‬ ‫چ ت� ضس‬ ‫ض‬ ‫�����شض ٹ‬ ‫ت ��� ا�����‬ ‫ت �ی‬

‫�ح���ع� ض ت‬ ‫ض ت� �������‬ ‫چ ت� ضس‬ ‫ت‬ ‫���� ض‬ ‫س��ط��م��ط�ت��ہ‬

‫����ع��ضس ض‬ ‫چ ت ��ح��ھ�‬ ‫ے‬ ‫چ‬ ‫�ح� ض‬ ‫�ئ‬ ‫��ع�‬ ‫چ ت�� ا����ھ‬ ‫ے�‬ ‫�‬ ‫چح���ھ���� ئ‬ ‫چ��عا����ھ�‬ ‫ے‬

‫تض‬ ‫����ھ��ض�‬ ‫��ت���ت��ہ‬

‫����ع��ضس ئ‬ ‫چ ت ��ح��ھ�‬ ‫ے‬ ‫چ‬ ‫�‬ ‫ل���ع��ض ت‬ ‫تس��ح��ھ‬ ‫ے�‬ ‫چ‬ ‫گ�‬ ‫�ل��ع��ض ئ‬ ‫ت�����ھ�‬ ‫ے‬ ‫چ‬

‫����س�ست‬ ‫چت �ض�ح��ھ�‬ ‫ے‬

‫‪8 letter combinations‬‬ ‫�‬ ‫گ�‬ ‫ل���ھ ٹت‬ ‫ل���ع� ض ت‬ ‫��س�ت�������‬ ‫ت�س��ح�����‬ ‫ت� ضس‬ ‫چ ت� ضس‬

‫‪Bari Yeh endings‬‬ ‫‪Two Character Ligatures all possible Combinations:‬‬ ‫ض���‬ ‫ے‬

‫���‬ ‫ے‬ ‫چ‬

‫���‬ ‫ے‬

‫ض‬ ‫���‬ ‫ے‬

‫ض�ض��ھ�‬ ‫ے‬

‫���ھ�‬ ‫ے‬ ‫چض‬

‫ت‬ ‫���‬ ‫ے‬

‫ٹ‬ ‫���‬ ‫ے‬

‫ش‬ ‫���‬ ‫ے‬

‫ض‬ ‫���‬ ‫ے‬

‫ت‬ ‫���‬ ‫ے‬

‫�‬ ‫���‬ ‫ے‬

‫ت‬ ‫�ض��ھ�‬ ‫ے‬

‫ٹ‬ ‫�ض��ھ�‬ ‫ے‬

‫ش‬ ‫�ض��ھ�‬ ‫ے‬

‫���‬ ‫ے‬ ‫چ‬

‫���‬ ‫ے‬

‫ض‬ ‫���‬ ‫ے‬

‫���ھ�‬ ‫ے‬

‫ش‬ ‫���ھ�‬ ‫ے‬

‫���‬ ‫ے‬

‫�م�‬ ‫ے‬

‫ض‬ ‫���‬ ‫ے‬

‫�ت��‬ ‫ے‬

‫ں‬ ‫���‬ ‫ے‬

‫���ھ�‬ ‫ے‬

‫ض‬ ‫���ھ�‬ ‫ے‬

‫�ط�‬ ‫ے‬

‫ض‬ ‫�ط�‬ ‫ے‬

‫‪Three Character Ligatures (ending at Bari-Yeh) all possible Combinations:‬‬

‫�عض��ھ�‬ ‫ے‬

‫ض‬ ‫�عض��ھ�‬ ‫ے‬ ‫ش‬ ‫���ھ�‬ ‫ے‬ ‫چ‬

‫ض‬ ‫�ض��ھ�‬ ‫ے‬

‫ت‬ ‫�ض��ھ�‬ ‫ے‬

‫�‬ ‫�لض��ھ�‬ ‫ے‬

‫�ح��ھ�‬ ‫ے‬ ‫چچ‬

‫�ح��ھ�‬ ‫ے‬ ‫چ‬ ‫�����ھ�‬ ‫ے‬ ‫چ‬

‫ٹ‬ ‫���ھ�‬ ‫ے‬ ‫چ‬ ‫ت‬ ‫���ھ�‬ ‫ے‬ ‫چ‬

‫�‬ ‫�ل��ھ�‬ ‫ے‬ ‫چ‬

‫ض�ح��ھ�‬ ‫ے‬ ‫چ‬ ‫گ�‬ ‫�ل��ھ�‬ ‫ے‬ ‫چ‬

‫�ل��ھ�‬ ‫ے‬ ‫چ‬

‫ض�حض��ھ�‬ ‫ے‬ ‫گ�‬ ‫�لض��ھ�‬ ‫ے‬ ‫ض‬ ‫�ح��ھ�‬ ‫ے‬ ‫چ‬

‫�ح��ھ�‬ ‫ے‬ ‫چض‬

‫�حض��ھ�‬ ‫ے‬

‫ض‬ ‫�حض��ھ�‬ ‫ے‬

‫���ض��ھ�‬ ‫ے‬

‫ش‬ ‫���ض��ھ�‬ ‫ے‬

‫���ض��ھ�‬ ‫ے‬

‫ض‬ ‫���ض��ھ�‬ ‫ے‬

‫�لض��ھ�‬ ‫ے‬

‫���ض��ھ�‬ ‫ے‬

‫ض‬ ‫�ض��ھ�‬ ‫ے‬

‫ہ���ض��ھ�‬ ‫ے‬

‫��ھض��ھ�‬ ‫ے‬

‫ت���ھ�‬ ‫ے‬ ‫ض‬

‫�����ھ�‬ ‫ے‬ ‫چ‬

‫ش‬ ‫�����ھ�‬ ‫ے‬ ‫چ‬

‫ض���ھ�‬ ‫ے‬ ‫چ‬

‫�����ھ�‬ ‫ے‬ ‫چ‬

‫ہ�����ھ�‬ ‫ے‬ ‫چ‬

‫��ھ��ھ�‬ ‫ے‬ ‫چ‬

‫ت���ھ�‬ ‫ے‬ ‫چ‬

‫ض‬ ‫�����ھ�‬ ‫ے‬ ‫چ‬

‫ٹت‬ ‫���ھ�‬ ‫ے‬

‫شت‬ ‫���ھ�‬ ‫ے‬

‫ت‬ ‫�����ھ�‬ ‫ے‬

‫شت‬ ‫�����ھ�‬ ‫ے‬

‫ت‬ ‫�����ھ�‬ ‫ے‬

‫ضت‬ ‫�����ھ�‬ ‫ے‬

‫ت‬ ‫�ط��ھ�‬ ‫ے‬

‫ضت‬ ‫�ط��ھ�‬ ‫ے‬

‫ت‬ ‫�ع��ھ�‬ ‫ے‬

‫ضت‬ ‫�ع��ھ�‬ ‫ے‬

‫ضت‬ ‫���ھ�‬ ‫ے‬

‫تت‬ ‫���ھ�‬ ‫ے‬

‫�ت‬ ‫�ل��ھ�‬ ‫ے‬

‫ت‬ ‫�ل��ھ�‬ ‫ے‬

‫ت‬ ‫�����ھ�‬ ‫ے‬

‫ضت‬ ‫���ھ�‬ ‫ے‬

‫ت‬ ‫ہ�����ھ�‬ ‫ے‬

‫ت‬ ‫��ھ��ھ�‬ ‫ے‬

‫ت‬ ‫ت���ھ�‬ ‫ے‬

‫ٹ‬ ‫ض���ھ�‬ ‫ے‬

‫ٹ‬ ‫���ھ�‬ ‫ے‬ ‫چ‬

‫تٹ‬ ‫���ھ�‬ ‫ے‬

‫ٹٹ‬ ‫���ھ�‬ ‫ے‬

‫شٹ‬ ‫���ھ�‬ ‫ے‬

‫ضٹ‬ ‫���ھ�‬ ‫ے‬

‫تٹ‬ ‫���ھ�‬ ‫ے‬

‫�ٹ‬ ‫�ل��ھ�‬ ‫ے‬

‫ٹ‬ ‫ض�ح��ھ�‬ ‫ے‬ ‫گ‬ ‫�ٹ‬ ‫�ل��ھ�‬ ‫ے‬

‫ٹ‬ ‫�ح��ھ�‬ ‫ے‬ ‫چ‬

‫ٹ‬ ‫�ح��ھ�‬ ‫ے‬

‫ضٹ‬ ‫�ح��ھ�‬ ‫ے‬

‫ٹ‬ ‫�ل��ھ�‬ ‫ے‬

‫ٹ‬ ‫�����ھ�‬ ‫ے‬

‫ضٹ‬ ‫���ھ�‬ ‫ے‬

‫ش‬ ‫ض�ح��ھ�‬ ‫ے‬ ‫گ‬ ‫�ش‬ ‫�ل��ھ�‬ ‫ے‬

‫ش‬ ‫�ح��ھ�‬ ‫ے‬ ‫چ‬

‫ش‬ ‫�ح��ھ�‬ ‫ے‬

‫ضش‬ ‫�ح��ھ�‬ ‫ے‬

‫ش‬ ‫�����ھ�‬ ‫ے‬

‫شش‬ ‫�����ھ�‬ ‫ے‬

‫ش‬ ‫�����ھ�‬ ‫ے‬

‫ش‬ ‫�ل��ھ�‬ ‫ے‬

‫ش‬ ‫�����ھ�‬ ‫ے‬

‫ضش‬ ‫���ھ�‬ ‫ے‬

‫ش‬ ‫ہ�����ھ�‬ ‫ے‬

‫ت‬ ‫�ح��ھ�‬ ‫ے‬ ‫چ‬

‫ت‬ ‫�ح��ھ�‬ ‫ے‬

‫ضت‬ ‫�ح��ھ�‬ ‫ے‬

‫ت‬ ‫���ھ�‬ ‫ے‬ ‫چ‬

‫تت‬ ‫���ھ�‬ ‫ے‬

‫�ع��ھ�‬ ‫ے‬ ‫چ‬

‫���ھ�‬ ‫ے‬ ‫چچ‬ ‫ض‬ ‫�ع��ھ�‬ ‫ے‬ ‫چ‬

‫ت‬ ‫���ھ�‬ ‫ے‬ ‫چ‬ ‫ض‬ ‫���ھ�‬ ‫ے‬ ‫چ‬

‫ت‬ ‫ض�ح��ھ�‬ ‫ے‬ ‫گ‬ ‫�ت‬ ‫�ل��ھ�‬ ‫ے‬

‫ض‬ ‫���ھ�‬ ‫ے‬ ‫چ‬

‫ت‬ ‫ض���ھ�‬ ‫ے‬

‫�ط��ھ�‬ ‫ے‬ ‫چ‬

‫ض‬ ‫�ط��ھ�‬ ‫ے‬ ‫چ‬

‫�طض��ھ�‬ ‫ے‬

‫ض‬ ‫�طض��ھ�‬ ‫ے‬

‫ٹ‬ ‫�����ھ�‬ ‫ے‬

‫شٹ‬ ‫�����ھ�‬ ‫ے‬

‫ٹ‬ ‫�����ھ�‬ ‫ے‬

‫ضٹ‬ ‫�����ھ�‬ ‫ے‬

‫ٹ‬ ‫�ط��ھ�‬ ‫ے‬

‫ضٹ‬ ‫�ط��ھ�‬ ‫ے‬

‫ٹ‬ ‫�ع��ھ�‬ ‫ے‬

‫ضٹ‬ ‫�ع��ھ�‬ ‫ے‬

‫ٹ‬ ‫ہ�����ھ�‬ ‫ے‬

‫ٹ‬ ‫��ھ��ھ�‬ ‫ے‬

‫ٹ‬ ‫ت���ھ�‬ ‫ے‬

‫ش‬ ‫ض���ھ�‬ ‫ے‬

‫ضش‬ ‫�����ھ�‬ ‫ے‬

‫ش‬ ‫���ھ�‬ ‫ے‬ ‫چ‬

‫تش‬ ‫���ھ�‬ ‫ے‬

‫ٹش‬ ‫���ھ�‬ ‫ے‬

‫شش‬ ‫���ھ�‬ ‫ے‬

‫ش‬ ‫�ط��ھ�‬ ‫ے‬

‫ضش‬ ‫���ھ�‬ ‫ے‬

‫تش‬ ‫���ھ�‬ ‫ے‬

‫�ش‬ ‫�ل��ھ�‬ ‫ے‬

‫ض���ح�‬ ‫ے‬ ‫ض‬

‫ش‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫‪the‬‬ ‫‪Typeset‬‬ ‫‪by‬‬ ‫‪� digital‬ضس��ح�‬ ‫‪� corpus.‬ضس��ح�‬ ‫ضح��ح�‬ ‫ضح��ح�‬ ‫‪in DecoType‬ح��ح�‬ ‫ضح��ح�‬ ‫‪ Nastaliq‬ض���ح�‬ ‫ض���ح�‬ ‫‪ the‬ض���ح�‬ ‫���ح�‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫چ‬ ‫چ‬ ‫ض‬

‫ض���ح�‬ ‫ے‬ ‫ٹ‬ ‫���‬

‫ض‬ ‫ض‬ ‫�ح��‬ ‫ے�‬ ‫ض‬ ‫ش‬ ‫���‬

‫ضش‬ ‫�ط��ھ�‬ ‫ے‬

‫ش‬ ‫�ع��ھ�‬ ‫ے‬

‫ضش‬ ‫�ع��ھ�‬ ‫ے‬

‫ت‬

‫ٹ‬

‫ش‬

‫ض‬ ‫ض�ک��ح�‬ ‫ے‬

‫ت‬ ‫ض�ک��ح�‬ ‫ے‬

‫��‬ ‫ضک��ح�‬ ‫ے‬

‫�ح��‬

‫�ح��‬

‫�ح��‬

‫گ‬ ‫��‬ ‫ضک��ح�‬ ‫ے‬ ‫ض‬ ‫�ح��‬

‫ض‬

‫ض‬

‫�‬ ‫ضک��ح�‬ ‫ے‬

‫�‬ ‫ضم��ح�‬ ‫ے‬

‫ض‬ ‫ض���ح�‬ ‫ے‬

‫�س��‬

‫�شس��‬

‫�س��‬

‫ہ�س��ح�‬ ‫ے‬ ‫ض‬ ‫ض‬ ‫�س��ح�‬ ‫ے‬ ‫چ‬

‫ض��ھ��ح�‬ ‫ے‬ ‫����‬

‫�ض‬ ‫�‬ ‫‪Sample‬ضس��ح�‬ ‫‪ of‬ضس��ح�‬ ‫ے‬ ‫ے‬

‫‪­author.‬‬ ‫ت���ح�‬ ‫ے‬ ‫ض‬ ‫ض‬ ‫����ح�‬ ‫ے‬ ‫چ‬

‫ض���ح�‬ ‫ے‬ ‫چ‬ ‫���‬

‫ش‬ ‫ش‬ ‫ت���ھ�‬ ‫��ھ��ھ�‬ ‫ے‬ ‫ے‬ ‫ض‬ ‫ ‪Figure 5.21‬‬ ‫ض����ح�‬ ‫ض����ح�‬ ‫ے‬ ‫ے‬ ‫���ح�‬ ‫ے‬ ‫چچ‬ ‫ض‬ ‫�‬ ‫��ح�‬ ‫ے‬ ‫چ‬

‫ت‬ ‫���ح�‬ ‫ے‬ ‫چ‬ ‫ض‬ ‫�ک��ح�‬ ‫ے‬ ‫چ‬

‫ح�‬ ‫ح�‬ ‫ح�‬ ‫ح�‬ ‫ح�‬ ‫ح�‬ ‫ح�‬ ‫ح�‬ ‫ضح�‬ ‫ح�‬ ‫ح�‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫چچ‬ ‫چ‬ ‫چ‬ ‫چ‬ ‫‪ it was‬چ‬ ‫چ‬ ‫چ‬ ‫چ‬ ‫‪rapidity‬چ‬ ‫چ‬ ‫‪the setup of the‬‬ ‫‪film pattern‬‬ ‫‪production,‬‬ ‫‪Jamil‬‬ ‫‪who,‬‬ ‫‪“with‬‬ ‫‪great‬‬ ‫چ‬ ‫گ‬ ‫ت‬ ‫ش‬ ‫ٹ‬ ‫ت‬ ‫ض‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫ھ‬ ‫س‬ ‫م‬ ‫ح‬ ‫ک‬ ‫ک‬ ‫ک‬ ‫ک‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫�‬ ‫‪and‬‬ ‫‪ingenuity”,‬‬ ‫‪made‬‬ ‫‪ with‬ضح�‬ ‫‪ no less‬ح�‬ ‫ح�‬ ‫ح�‬ ‫ح�‬ ‫ض ح�‬ ‫‪the original‬تح�‬ ‫ح�‬ ‫‪ drawings‬ہ ح�‬ ‫ح�‬ ‫ے‪ and‬ح�‬ ‫ے‪ film‬ح�‬ ‫‪patterns.85‬ح�‬ ‫ح�‬ ‫�ح�‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫چ‬ ‫چ‬ ‫چ‬ ‫چ‬ ‫چ‬ ‫چ‬ ‫چ‬ ‫چ‬ ‫چ‬ ‫چ‬ ‫گ‬ ‫ض‬ ‫ت‬ ‫ض‬ ‫‪ of the‬ض‬ ‫ض‬ ‫‪Here‬‬ ‫‪the‬‬ ‫‪implications‬‬ ‫‪approach‬‬ ‫‪described‬‬ ‫‪above,‬‬ ‫‪which‬‬ ‫ش‬ ‫‪ combined‬ض�‬ ‫�‬ ‫‪� a‬‬ ‫��‬ ‫��‬ ‫ک��ح�‬ ‫ک��ح�‬ ‫�ک��ح�‬ ‫�ک��ح�‬ ‫���ح�‬ ‫���ح�‬ ‫����ح�‬ ‫����ح�‬ ‫�س��ح�‬ ‫�س��ح�‬ ‫�س��ح�‬ ‫�س��ح�‬ ‫ح��ح�‬ ‫ح��ح�‬ ‫ح��ح�‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫ے‬ ‫چ‬ ‫‪language lexicon with the style’s morphology, attained their full bearing on the‬‬ ‫‪process. As linguistic, rather than morphological principles guided the design,‬‬ ‫‪any change of configuration within a letter group – i.e. the addition or subtrac‬‬‫‪tion of any one letter – necessitated its multiplication. In most typographic‬‬

‫‪Ibid.‬‬

‫ ‪85‬‬

Beginnings of Digital Arabic Type

‫ن‬

349

� �� �‫ےک‬, consisting of two letters representations the four letter groups ‫ے‬ ‫ے �ہ‬ ‫ے� ���س‬ each, required five individual characters – one for each first letter, and one for ‫ن‬ the second that could be combined with all others: ‫ے‬ �� ‍‫ک‍ ‍� �� ‍س ‍ہ‬. In Jamil and Saiyid’s approach these four letter groups needed only four characters; however, if one was to combine any of the five individual letters with any other letter, ‫� ن‬ ‫�� �ن‬ �‫ےک‬. Therefore not five, but eight characters were required: ‫ے ک���ی �ی���سی��ہی‬ ‫ے �ہ‬ ‫ے� ���س‬ a steep increase of characters became inevitable, and it was estimated that a staggering 20,000 characters were required in the production of the font.86 Once the artwork was finished, Jamil sent the drawings and patterns to Monotype’s UK office in “batches of about 1000, together with the codes that would control their use.”87 Then the film patterns were digitised in the customary fashion on a rotary scanner, “each character taking about two minutes.”88 During this period Monotype’s digitisation procedure was based on the assumption that “no significant modification other than a bitmap-editing process on screen” was required when adapting photocomposition founts to digital devices.89 Given the large number of characters, and the opinion that “there is in fact only a minimum of design modification required when digitizing a typeface, especially for higher resolutions”, it can be assumed that little or no editing of the artwork was effected.90 This thesis is supported by the relatively quick implementation, as “within seven months the complete font was produced and assembled, the software completed, and the job up and running.”91 The font’s vast character set was accessed through the computer’s character selection logic that chose the ligature which corresponded to the keyed sequence. It was thus possible to use a keyboard comprising only of the basic alphabet, in principle employing the same system as had first been developed by Linotype in 1967 (see page 187 and following). In the announcement of the new typeface – heralded as the “greatest exotic language triumph for ‘Monotype’” – the system’s debt to computerisation was acknowledged:

86

In the published text corpus only about 18,000 letter combinations are included.

87

Sheppard, “Urdu Nastaliq, 1981,” 31. This seems to imply that also the logic on which the software would be based was developed in Pakistan.

88

Ibid.

89

David Saunders, “Two Decades of Change 1965–1986,” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 33.

90

Ibid.

91

Sheppard, “Urdu Nastaliq, 1981,” 31.

350

Chapter 5

The solution to printing Urdu Nastaliq was not possible until new technology in high-speed phototypesetting [sic] was combined with large capacity computer storage for calligraphic characters.92 In October 1981 The Daily Jang featured a full-page announcement of the new composition system (Figure 5.22). It showcased the new typeface in conjunction with the new device’s capabilities: a number of type sizes, white type on black background, and, crucially, a normalised, typographic rendering of the preferred writing style of Urdu: Nastaʿlīq. For some time the newspaper employed the typographically composed text in parallel to columns that were still handwritten (Figure 5.23). This juxtaposition of the two techniques of text composition is telling for the specificities of Urdu typography. The cultural norm that defined the appearance of text was so pronounced that no typographical technology found acceptance until it could reasonably match

Figure 5.22 Announcement of Monotype Nastaliq, The Daily Jang, October 1981, reduced to 35% linear. DTGC, folder Nastaliq 1980s 1 of 3. 92

The Monotype Corporation Ltd, “Greatest Exotic Language Triumph for ‘Monotype’,” The Monotype Recorder no. 3, New Series (October 1981): 1.

Beginnings of Digital Arabic Type

351

Figure 5.23 Handwritten (left) and typeset (right) column, enlarged to 200% linear. Note the handwritten words inserted in the typeset column (bottom right), indicative of missing characters in Monotype’s font. DTGC, folder Nastaliq 1980s 1 of 3.

e­ xpectations. So for Urdu all technologies from foundry type to hot-metal composition, and the various instantiations of photocomposition were eschewed, to arrive at this unusual confrontation of manual with digital text composition. Here, the advantages of the typeset over the manually composed text are manifest in its superior clarity, alignment and consistency. The typeface’s achievement in becoming the first widely accepted typographic rendering of Nastaʿlīq was notable. It established the feasibility of typesetting a ‘hanging’ writing style and furthered Monotype’s prestige in Urdu typography, first established with the introduction of Monotype Series 507. Yet, the system’s very principles, although allowing an authentic reproduction of Nastaʿlīq, also proved to be irksome. The approach had no flexibility and was bound by the scope of the lexicon initially assembled by Jamil and Saiyid. Thus, any omissions of letter combinations caused a breakdown of the font and the newspaper’s scribes had to manually insert corrections. This systemic shortcoming required the continued additions of characters to the font;93 any new expressions, foreign loan words and changes of spelling, regular phenomena in any natural language, risked continued font failures if a sequence of letters had not been previously included; for the same reasons the font could not

93

When launched it was rumoured that the font did not contain the letter groups necessary to spell Pakistan. It was suggested that the newspaper’s scribes therefore had to manually fill the gaps in a grotesque contradiction of the desired effect of the new system. Ross, interview.

352

Chapter 5

be used for languages other than Urdu and remained a limited commercial proposition. The typeface, which was later called Noori Nastaliq, had been conceived for the Lasercomp and Monotype did not adapt it for new devices and evolving technologies, an apparent testimony to the inflexibility of the original system. However, Jamil subsequently licensed the typeface for inclusion in the page layout software InPage, first released in 1994, contributing to its continued presence in digital Urdu typography.94 The typeface’s history remains incomplete, and lacks substantial primary sources documenting its inception and evolution. Attempts by the author to locate material pertaining to its creation in the archives of Monotype Ltd were fruitless, and explicit enquiries with its staff yielded no further results.95 Linotype Ruq’ah, the Exploration of a Writing Style The Ruqʿä writing style has its origins in the Ottoman court, its invention dating from the nineteenth century.96 According to Gunnar Vilhjálmsson, the two calligraphers Mümtaz Efendi (1810–1872) and Mustafa İzzet Efendi (1801–1876) are commonly credited for the canonisation of the style (Figure 5.24).97 It is unrelated to the much older Riqāʿ style, and misattributions and confusion of the two names are frequent.98 Developed as a running hand in the court’s chancelleries, Ruqʿä is a relatively simple and plain style, intended for the rapid production of texts.99 As such, and contrary to most other styles, Ruqʿä 94

At the time of writing the current version of the software no longer includes the typeface. Concept Software Private Ltd, “InPage Urdu,” n.d., http://inpage.com.pk/ (accessed 16 May 2017).

95

A planned interview with former Monotype employees involved in the project was cancelled at short notice and the author was informed that Monotype would not discuss the history of the typeface beyond the already published articles.

96

Hassan Massoudy makes the implausible suggestion that the style was invented precisely in 1863. Hassan Massoudy and Isabelle Massoudy, Abcdaire de la calligraphie arabe (Paris: Flammarion, 2003), 101.

97

Gunnar Vilhjálmsson, “Ruq’ah – the new style”, The Recorder, no. 1, 2014, 107–116.

98

See, for example, Yasin Hamid Safadi, Islamic Calligraphy (London: Thames & Hudson, 1978), 20. Confusions of the two styles may have arisen from the imprecise use of transliteration in which the distinctiveness of the two Arabic names has been lost in its Latin َ ‫ُ ْق‬ ‫َق‬ form. Ruqʿä is spelt ‫ ر���ع��ة‬, whereas Riqāʿ is spelt ‫ر��ا‬. ‫ِع‬ Vilhjálmsson notes that compared to other writing styles Ruqʿä saved the scribes in the

99

Beginnings of Digital Arabic Type

353

Figure 5.24 Ruqʿä sample by the Ottoman calligrapher Mustafa İzzet Efendi, actual size. From ‫‫‬Nāǧī Zayin ad-Dīn, a‫ ‫‬l-Ḫaṭṭ al-ʿarabiyy‬,‬(Baghdad: Publications of the Iraq Academy, 1968)‬.

was not used for representative purposes or the calligraphic arts. In literature on Islamic calligraphy descriptions of Ruqʿä are habitually omitted, making it, despite its proliferation throughout large parts of the Middle East, a poorly documented style.100 It is marked by the strong angle words have in relation to the baseline, and the frequent stacking and nesting of letter groups. Its proportions are compact, and commonly Ruqʿä is written in simple, thick and short strokes, lending it a particularly graphic quality. These characteristics make it a frequent choice of sign painters throughout the Middle East up to the current day (Figure 5.25 overleaf); and before page make-up became purely typographical, Ruqʿä was the preferred hand for large newspaper headlines.

Ottoman chancelleries on average two seconds per word, reducing the time to pen a 500 word document by 16 minutes. “Ruq’ah – the new style”. 100

Little research and documentation about the style may be explained by its late inception and its unglamorous role as an everyday hand.

354

Chapter 5

Figure 5.25 Hand-lettered street sign, Damascus, Syria, 2008. The first line is written in the Nasḫ style, the second and third in Ruqʿä. Photograph by the author.

In the late nineteenth century renderings of the style were attempted in foundry type, notably by the Ottoman printer Ohannis Mühendisoğlu and at the Bulāq press, and in the mid-1950s the Lebanese Fonderie Typographique Orientale cut a Ruqʿä -inspired metal fount (Figure 5.26). However, marred by the constraints of foundry type, use of these designs remained limited. The principal characteristics of Ruqʿä, the angle of the joining stroke and the stacking of letter groups, prohibited a fully successful reproduction with type: the rigid physical properties of foundry type did not allow for the diagonal joins and made the flexible positioning and spacing of letter groups impossible. Mechanical hot-metal composition, in many ways more restrictive than foundry type, did not alleviate these difficulties. Despite the known constraints, in 1969 al-Ahram’s make-up editor Hišām Baḥarī endeavoured to develop a Ruqʿä heading typeface for the Ludlow Typograph caster.101 Apparently intended to replace the then current hand-lettered headlines, the characters were designed in a way that attempted to bypass the limitations of hot-metal composition (Figure 5.27). Instead of the typical 101

The Ludlow Typograph was a machine used for the casting of lines of type in large sizes. Its simple mechanism was semi-mechanic, thus requiring the manual composition and re-distribution of matrices for every line. Newspapers frequently employed Ludlow machines alongside their Linotypes or Intertypes for the composition of headlines. Douglas McMurtrie, The Ludlow System of Slugline Composition (Chicago: Ludlow Typograph Company, 1927), http://archive.org/details/McMurtrieLudlowSlugline (accessed 16 May 2017).

Beginnings of Digital Arabic Type

355

Figure 5.26 Roqua‘i, corps 36, specimen (detail), reduced to 65% linear. From Fonderie Typographique Orientale (Beirut: 1956), DTGC, Box Arabic General.

Figure 5.27 Detail of cover page, al-Ahram, 18 March 1966, reduced to 35% linear. Courtesy of Internationales Zeitungsmuseum, Aachen, Germany.

diagonal joins, this design – probably drawn by an in-house lettering artist – resorted to a generic flat joining stroke; the styling was generally based on Ruqʿä features, yet the manner in which individual characters joined appeared sorely compromised (Figure 5.28 overleaf). Moreover, the use of half-forms for a number of letters and the inability to kern resulted in impaired proportions of some characters (see, for example, the excessively narrow medial and initial ‫ ك‬kāf). Linotype produced the matrices at al-Ahram’s request for a range of sizes (24, 30, 48, 60 and 72 pt); by November 1970 the fount was tested by the newspaper using the Ludlow caster, but as Gabeyan reported, the results did not live up to the expectations:

356

Chapter 5

Figure 5.28 Character drawings for hot-metal Ruqʿä fount, 1969, reduced to 50% linear. DTGC, folder Ruq’ah.

They are not happy at all but nevertheless they are not blaming us because it was Bahari’s idea, and they are not blaming Bahari, because they do not want to blame a dead person [Baḥarī had died in a car accident on New Year’s Day 1970]. As a result they are not using our heading matrices which cost them a small fortune.102 It appears that the newspaper’s management made the decision to develop the type based on the character drawings, and underestimated the profound influence the constraints of hot-metal composition. As Gabeyan remarked, “the individual characters have nothing wrong, but when you start joining them, then the result is, especially in some combinations, unsatisfactory”.103 In 1972 the fount was revised by al-Ahram and 56 additional characters, mainly ligatures between the most unfavourable letter combinations, were ordered (Figure 5.29). The additions were manufactured by Linotype, but no use of the 102

Hrant Gabeyan to Walter Tracy, Letter, “Al Ahram - Ruq’ah Heading Matrices,” 26 November 1970, WT correspondence, folder 16, DTGC.

103

Hrant Gabeyan to Walter Tracy, Letter, “Al Ahram - Ruq’ah Heading Matrices,” 22 December 1970, WT correspondence, folder 16, DTGC.

Beginnings of Digital Arabic Type

357

Figure 5.29 Drawings of additional characters ordered by al-Ahram (detail), 1972, actual size. DTGC, folder Ruq’ah.

typeface could be found. With the beginning of photocomposition in the 1970s, this compromised design soon became obsolete as a relic of the hot-metal era. By 1975, and probably without relation to the earlier development, renewed interest in Ruqʿä appears to have been articulated by Holloway during his brief employment with Linotype. A letter from Tracy, written in reaction to Holloway’s “scheme for the composition of scripts in which the words are at an angle to the horizontal”, documents his initiative. In his reply Tracy noted: Derek Kyte [Linotype-Paul’s Technical Director] has asked me to say that he has studied your scheme […] He admires the way you have recognised the possible value in combining computer programs with the facilities of certain kinds of phototypesetting machines to overcome the inherent difficulties of particular forms of typography. However, it seems that your scheme would apply to only two scripts. The Nastaliq is only used substantially in Pakistan, a country which is, and will be for some time to come, unable to invest in typesetting machinery of any sort. This leaves the Ruqah script, which is certainly important as a secondary script in the Arab world, but it is thought unlikely that enough printers will agree to buy an additional program (since they must take one or both of the current programs) to justify the time and effort needed to create it […] His conclusion is that in this particular case he does not think the company should take up the scheme.104 In this case Holloway’s initiative was trumped by economic considerations, but by the 1980s a different climate provided a new opportunity for a Ruqʿä development. Throughout the second half of the 1970s the Middle East became an increasingly important market for Linotype. In 1976 a substantial number of installations made it a commercially interesting region, and by 1980 Linotype’s large market share made Arabic type a profitable aspect of the 104

Walter Tracy to Timothy Holloway, Letter, 4 November 1975, WT correspondence, folder 18J-2 Trad Arabic VIP & 505 1974-1977, DTGC.

358

Chapter 5

company’s business.105 At this time, the company’s management recognised the growing importance of non-Latin type developments and assigned a dedicated product manager to this part of its operations. Dr Michael Fellows, who had joined the company in 1976 as a computer developer, was appointed to this influential role in 1983.106 Against this background the prospect of developing a Ruqʿä typeface resurfaced and was reconsidered in light of the changed circumstances. According to Fellows, now Ruqʿä was seen as a development “that would be nice to do and we [Linotype] could afford to do it.”107 The new climate at Linotype thus allowed for experimental type developments without requiring a specific commercial prospect, in what must be called a profound change of approach. By September 1983 Holloway, now working as an independent designer, had provided Linotype with material of his research into Ruqʿä. He was then briefed about the requirements of the two devices on which the new design was to be implemented: the Linotron 202 and the CRTronic 200.108 When Holloway began his work on the new Ruqʿä design, the capabilities of the 202 were seen as liberating from previous constraints.109 The Linotron 202 had the ability of specifying kerning between letter pairs in smaller increments, taking advantage of its finer unit system. Whereas the upgraded, post-1976, V-I-P relied on a 54-unit drawing area, with 31/2-units reserved for kerning, starting from the Linotron 606 drawings for digital typesetters could take advantage of 9 discreet units assigned for kerning (Figure 5.30 and Figure 5.31). A related key advantage of the 202 was the increased flexibility of the accent placement. Where the V-I-P could only specify a generic ‘high’ or ‘low’ position for a vowel, the 202 could finally define precise coordinates which reflected the height of the base character. Thus, positions were specified with finer increments and depended on typestyle, greatly improving the appearance of vocalised text

105

According to Fiona Ross, Linotype held an estimated 95% of the Arabic-script newspaper market and 80% of the commercial market. Ross, “Non-Latin Type Design at Linotype”.

106

The non-Latin product manager was of equal rank as the managers for typesetters or front-ends which was “a very clear recognition that non-Latins were becoming very important to us.” Michael Fellows, interview by the author, 9 August 2013.

107

Ibid.

108

Sarah Morley to Timothy Holloway, Letter, 23 September 1983, Folder Ruq’ah 1980s, DTGC. Sarah Morley was a design graduate from the University of Reading, UK, and worked for several years as a member of Linotype Ltd’s Typographic Development department. For the Linotype Ruq’ah development she frequently liaised with Holloway.

109

Fellows and Ross, interview.

Beginnings of Digital Arabic Type

359

Figure 5.30 Left: Diagrams showing the exposure fields of the V-I-P and the Linotron 606, from Linotype Font Center, “Instruction for the Preparation of Artwork for Digital Systems,” 1981, 8, reduced to 30% linear. DTGC, folder Arabic CRTronic. Figure 5.31

Right: “Drawing guidelines for Arabic typefaces,” February 1983, reduced to 60& linear. DTGC, folder 202 typeface lists.

­setting.110 Also character set limitations were alleviated as one 202 font held 256 characters, more than two V-I-P fount strips.111 Yet, other constraints still imposed themselves. Besides the recurring issue of a limited character set, the dimensions of the machine’s exposure window in combination with the employed unit system were defining aspects of any character drawing. Every font had its own drawing mask, “showing the baseline, the kerning area and the size of the typesetting window along with vertical lines spaced according to the relative unit-width system employed.”112 These vertical dimensions of the typesetting area remained a limiting aspect for Arabic typefaces, typically marked by tall ascenders and deep descenders (Figure 5.32 overleaf). Drawing guidelines for new Arabic designs used in 1983 specified the following: 110

Ibid. The same functionality was already offered earlier with the KITAB software on the Digiset, see page 327 and following.

111

Linotype’s R&D department had also developed a process called ‘font merging’ by which two distinct fonts were presented to the user as one single font. Ross, interview.

112

Fiona Ross, “Selected Exhibits from the Non-Latin Type Collection, University of Reading,” in Non-Latin scripts, from metal to digital type (London: St Bride Library, 2012), 36.

360

Chapter 5

Figure 5.32 Left: Drawing of Ruqʿä extenders in relation to the exposure window of the Linotron 202, September 1983, reduced to 40% linear. DTGC, folder Ruq’ah 1980s. Figure 5.33 Right: Character drawings for Linotype Ruq’ah, n.d., reduced to 20% linear. DTGC, folder Ruq’ah 1980s.

Beginnings of Digital Arabic Type

361

The standard Latin mask has 54 units space above and 18 units space below the baseline. The baseline for a particular Arabic typeface may be moved upwards to accommodate long descenders provided that it is moved in values of whole units (1/54’ths), and that the overall height of any character does not exceed 72 units. It is not advisable to move the baseline more than 9 units upwards as the constraints of the 202 typesetting area (window) would mean that the face could not then be set up to the maximum size of 72pt.113 This constraint was relevant also in relation to a new aspect of the 202’s software. It contained a command called ‘Base line Jump’ which allowed specifying a vertical shift of a character.114 But as the exposure window of the device was of a limited height, any such base line jump had to be defined taking the point size and the remaining vertical space into account. As the software manual explained, “for example, at 72 point any amount of base line jump potentially moves the character over the edge of the CRT”, thus preventing its exposure.115 ∵ The other device for which the Ruqʿä typeface was developed was the CRTronic. Although some specifications for Arabic on the CRTronic had not yet been defined when the Ruq’ah development began, other characteristics were established: the font could have ligatures, and up to 200 character pairs for kerning or letterspacing could be defined.116 Holloway was provided with technical information of both machines, but then appears to have worked largely independently on the artwork (Figure 5.33). According to Ross the typeface was inspired by Writing Arabic by Terrence Mitchell, but it is not clear how much it influenced Holloway in his design.117 113

Linotype-Paul Ltd, “Drawing Guidelines for Arabic Typefaces,” 3 February 1983, Locker C1, DTGC.

114

B. S. Cash, “Linotron 202 Software” (Linotype-Paul, Phototypesetters Development Department, Cheltenham, October 1980), Locker C1, DTGC.

115

Ibid., 28.

116

Morley to Holloway, 23 September 1983.

117

Fellows and Ross, interview. T. F. Mitchell, Writing Arabic, A Practical Introduction to Ruq’ah Script, (London: Oxford University Press, 1953).

362

Chapter 5

Figure 5.34 Detail of a dot-matrix printer proof of Linotype Ruq’ah, 12 pt, reduced to 70% linear. DTGC, folder Ruq’ah 1980s.

Documentation of the project is sparse, and continues only one year later, in August 1984.118 Here, Morley asked the designer to review the kerning classifications [in original], as well as to provide accent positions for all characters.119 Around the same time proofs of the already digitised characters, printed in Germany at the Linotype Font Center on a dot-matrix printer in 12 pt, were sent to Holloway for his review (Figure 5.34). The characters are displayed individually between capital H characters, and thus provide no conclusive clues about the design’s success at reproducing the Ruqʿä style. A text proof from February 1985, output in 11 pt on a Linotron 202, provides better insight into the type’s appearance when composed (Figure 5.35), and the compromises a generic baseline imposed on the design are noticeable at once. The overall ­impression 118

According to Ross, much documentation was unfortunately lost in later years when the records of Linotype Ltd were transferred to the German seat of Linotype GmbH in Bad Homburg. Email message to the author, “Chapter 5,” 27 August 2013.

119

Sarah Morley to Timothy Holloway, Letter, 2 August 1984, Folder Ruq’ah 1980s, DTGC.

Beginnings of Digital Arabic Type

363

Figure 5.35 Detail of Linotron 202 proof of Linotype Ruq’ah, 11/15 pt, enlarged to 200% linear. DTGC, folder Ruq’ah 1980s.

of the text is much flatter than is typically the case in Ruqʿä, and the lack of the distinctive stacking and nesting of characters give a pronounced horizontality to lines. Medial characters were conceived with generic right-hand connections to ensure their compatibility with any preceding character. Although this was standard practice in typefaces inspired by the Nasḫ style, with Ruqʿä this approach caused considerable changes of letterforms. The compromises of shape that are necessitated by one generic connection along a single horizontal alignment are particularly noticeable, and detrimental to the overall impression of the text. For example the medial form of ḥāʾ, commonly a ‫ح‬ compact sequence of two strokes in diagonally opposed directions, was now replaced by the letter’s initial form, combined with a flat connecting stroke to the right (Figure 5.36). Although the individual letterforms followed – where possible – the Ruqʿä style, in text the loss of its typical morphology is clearly visible. By April 1985 the typeface had been tested on a 202 installation in Damascus and a report was conveyed to Linotype. The report, signed by George Henry, an employee of Linotype’s Middle East Liaison Office (MELO), identified a number of shortcomings: among them were excessive word spaces, the relatively small size of the face, clashes of vowels with base characters, the angle of the

364

Chapter 5

Figure 5.36 (a) Left: combined character drawings, from right to left: initial ‫ �ه‬hāʾ medial ‫ح‬ ḥāʾ and final mīm, reduced to 40% linear. Ibid. (b) Right: hand lettering of same ‫م‬ letter sequence illustrating common Ruqʿä morphology with the same nominal baseline, actual size. Arrangement and sketch by the author.

Figure 5.37 Comparison of typical Ruqʿä morphology (left) with Linotype Ruq’ah typeface (right), reduced to 40% linear. From George Henry, “Report on Ruq’ah Typeface,” 4 (detail).

Beginnings of Digital Arabic Type

365

Kashida character and general discrepancies between the typeface and typical Ruqʿä morphology (Figure 5.37).120 As the manager of the R & D section, and aided by comments from Holloway, Ross reviewed the observations by Henry. From the corresponding notes it emerges that most of the issues were related to technical constraints of the typesetting machine. The relative size of the face appeared small because it had to be chosen in accord with the exposure window’s height: the font’s dimensions were determined by allowing the tallest character, a ligature of ‫ ك‬kāf and  mīm, to fit inside the window at 72 pt (see Figure 5.32 above). In relation to the ‫م‬ Kashida enquiry, Ross noted that “in Ruq’ah calligraphy, justification is normally achieved without stretching any characters”, and that the Kashida character contained in the font was provided and “designed to be inserted before the following characters, only when the preceding character takes a diacritical sign, which would clash with the backward kerning stroke” ­(Figure 5.38 overleaf).121 It was not to be used for justification, and its shape was prescribed by the uniform alignment (Figure 5.39 overleaf). In conclusion, Ross’ notes about the stylistic discrepancies are instructive for the development of the typeface: The main explanation for the differences shown in the report is that we started with a very limited character set and no special software for varying alignments and shapes of the same character. […] Some of the suggested logotypes [ligatures] would entail the addition of a considerable number of characters, some of which do not appear in our Naskh faces. Also additional software is required to select the appropriate form in certain sequences. The latter has already been requested. We would like M.E.L.O., however, to specify those logotypes or series of logotypes – which should be added to the fonts.122 In addition to this review, the software controlling the placement of vowels was corrected to prevent further clashes. Whether, and to what extent additional characters were added to the typeface could not be established, but by 120

George Henry, “Report on Ruq’ah Typeface,” n.d., Folder Ruq’ah 1980s, DTGC. The complaint about a Kashida character in a Ruqʿä typeface is peculiar. In this style elongations between letters are not permitted because the medial ‫ ��س‬sīn can be written as a straight stroke, raising the potential of confusion if elongations were to be employed.

121

Fiona Ross, “Ruqah. Notes Following George Henry’s ‘Report on the Ruq’ah Typeface’,” n.d., Folder Ruq’ah 1980s, DTGC.

122

Ibid., 3.

366

Chapter 5

Figure 5.38 Above: Sketch explaining the use of the Kashida character in the Linotype Ruq’ah font, Holloway, “Comments on report by George Henry” (detail), actual size. DTGC, folder Ruq’ah 1980s. Figure 5.39 Below: Sketch explaining design characteristics of the Linotype Ruq’ah typeface, Holloway, “Comments on report by George Henry” (detail), reduced to 50% linear. DTGC, folder Ruq’ah 1980s.

October 1985 the typeface was released to manufacturing for the Linotron 202.123 Its development for the CRTronic proceeded in parallel, and the typeface was first exhibited at the DRUPA 1986 (Figure 5.40).124 According to Ross, the typeface was not a commercial success, an assessment that is underlined by the fact that Linotype Ruq’ah was not adapted to 123

Fiona Ross to Reinhard Hiltmann and Jürgen Eger, Memorandum, “Typeface Development Acceptance,” 15 October 1985, Folder Ruq’ah 1980s, DTGC.

124

Due to continued changes in the company structure from 1 June 1985 Linotype-Paul Ltd became Linotype Limited (henceforth Linotype Ltd). In July 1986 an internal memorandum notified Linotype Ltd staff that the Ruq’ah font was available on the CRTronic, providing support for Arabic, Kurdish and Persian. Ros Coates to Fiona Ross, Memorandum, “CRTronic Ruq’ah,” 16 July 1986, Folder Ruq’ah 1980s, DTGC.

Beginnings of Digital Arabic Type

367

s­ ubsequent typesetting technologies.125 Despite its design for then state-of-theart devices, the demands of the morphology of Ruqʿä proved too complex for the typographic concepts employed. The principal problem lay in the static baseline alignment which prohibited diagonal connections, and thus substantially altered the morphology, and consequently the appearance of the style. But although the resulting font could not be considered a success, its development provided Holloway and Linotype with valuable experience in the design of a ‘hanging’ script that could be put to use in another project. Whereas Tracy had counselled against the development of Holloway’s studies of diagonally joining scripts because Pakistan did not appear to be developed enough as a market, Ross recognised that the availability of this style on typesetting machines “opened up a whole country”, making it a valuable development.126

Figure 5.40 Page from “Layinūtāīb tuqaddim ǧihāz al-Layinūtrūnīk 500,” promotional text introducing the new typeface (detail), composed in Ruq’ah, n.d., actual size. DTGC, folder Ruq’ah 1980s. 125

Fellows and Ross, interview.

126

Ibid.

368

Chapter 5

Linotype Sheeraz, a New Approach to Arabic Type Against the background of the Ruqʿä project, Linotype Ltd embarked on a new and more ambitious design. Nastaʿlīq, the other major ‘hanging’ writing style still in use today, had previously proved to be an insurmountable technical challenge for the Linotype group. Its diagonal joins and the particular morphology of the style had been difficult to reproduce with previous technologies. Monotype’s Nastaʿlīq on the Lasercomp circumvented the typographical complexity by means of a large character set, yet failed eventually because of its inefficiency. This design flaw was recognised by Linotype’s R & D section soon after the publication of the Monotype font (see page 346). A report dated 25 May, 1982, proposed a scheme for typesetting Urdu in the Nastaʿlīq script “through the positioning and joining of individual letterforms”.127 In apparent reference to Monotype’s system, it noted: The advantage in this approach to the typesetting of Nasta’liq lies in its flexibility and versatility: individual digitised phonemes [sic] may be joined in every possible combination, enabling all words to be spelt with equal ease; the typeset output is not restricted by being able to access only particular letter combinations and words. All characters could be accessed from a conventional Arabic keyboard.128 The system’s principle was based on the provision of characters with two alternative joining strokes. A ‘thick’ and a ‘thin’ join should be provided for every character, resulting in up to nine distinct characters per letter. The proposal’s feasibility was tested by means of simple draft drawings which were based on the Monotype font (Figure 5.41). Although the report found the results satisfactory from a design point of view, it recommended further testing in consultation with a calligrapher. Based on these design principles, it was estimated that an entire font would comprise less than 1,000 characters – a substantial reduction in comparison to Monotype’s scheme. Yet, no actual development based on this proposal is documented, and Linotype’s explorations of Nastaʿlīq only yielded tangible results in early 1986.

127

Sarah Morley, Proposed Scheme for Typesetting Nasta’liq, 25 May 1982, 1, Folder Nastaliq 1980s 1 of 3, DTGC. According to Ross, this report, although signed by Morley, would have been authored in conjunction with Fellows and Ross, and was based on discussions with Holloway. Ross, email message to the author, “Re: Nastaliq Beginnings,” 27 April 2013.

128

Morley, Proposed Scheme for Typesetting Nasta’liq, 1.

Beginnings of Digital Arabic Type

Figure 5.41

369

(a) “Feasibility tests based on Monotype artwork,” (b) “The combining of individual characters to form a word,” Linotype’s proposed scheme for typesetting Nasta’liq, 1982, reduced to 40% linear. DTGC, folder Nastaliq 1980s 1 of 3.

According to Ross, the impetus to develop a Nastaʿlīq typeface came from the company’s agents in Pakistan, namely its representative Feroze K. Khambatta, and the conviction that a better system than the one proposed by Monotype would be possible with the available technology.129 A key element in the ensuing development was the “support of a dedicated non-Latin software team, initially headed by Dr Mike Fellows”, an asset few, if any other type manufacturers had at the time.130 The design of the font’s behaviour and the design of the 129

Ross, interview.

130

Ross, “Selected Exhibits from the Non-Latin Type Collection, University of Reading,” 37.

370

Chapter 5

Figure 5.42 Photographic trial setting of Holloway’s Nastaʿlīq artwork, March 1986, reduced to 30% linear. DTGC, folder Sheeraz Nastaliq.

Figure 5.43 Selection of manuscript samples from Pakistan, March 1986, reduced to 30% linear. DTGC, folder Sheeraz Nastaliq.

Beginnings of Digital Arabic Type

371

typeface’s characters were interrelated to an unprecedented degree, and thus the development began with the establishment of a new composition scheme. Research and analysis of Nastaʿlīq led the team – principally formed by Fellows, Ross and Holloway – to re-evaluate established concepts of Arabic typography. A draft document written by Holloway in March 1986 introduced numerous novel and revolutionary concepts that were to be implemented in the new font.131 In contrast to Monotype’s design, no ligatures were to be included, and the principal design approach sought a further subdivision of characters, rather than their combination in letter groups. Thus, diacritical signs were to be provided as separate characters and positioned during composition according to programmed coordinates; this principle substantially reduced the number of characters that had to be included in the font and provided increased flexibility for other design aspects. The selection of characters relied on the basic Arabic selection routines, augmented by additional algorithms which analysed the letter sequences for the context-specific substitution of characters. In parallel to the conceptual work, Holloway, engaged to provide the artwork for the new typeface, began drafting the character drawings. By March 1986 a photographically reduced proof was made from large letter drawings in order to canvas opinion from potential clients (Figure 5.42). The proof was brought by Peter Watkins of the MELO to Pakistan. There, Watkins it “to as many people as possible”, and found that “the basic reaction is very positive”. Watkins further reported that the angle of Holloway’s design seemed too pronounced, and that the preference in Pakistan was for a more horizontal look. He illustrated this with seven samples in which local scribes copied the same text according to their own preference and skills (Figure 5.43). In reference to local differences in style, Watkins noted: Interestingly the people in Sind (Karachi) prefer a flatter Nastaliq to those in Punjab (Lahore) and this is again apparent in the copies of which the first four were done in Karachi. The Monotype face is more acceptable in Karachi than it is in Lahore for this reason. Other criticisms of the design were final characters projecting too deep under the baseline, its loose inter-word spacing, and the height of ascenders, which were considered too short. In addition to the written samples, Watkins gathered printed specimens, which included a sample of the Monotype face,

131

Timothy Holloway, Nasta’liq Project: Outline of Proposed System, 13 March 1983, 1, Folder Sheeraz / Nastaliq, DTGC.

372

Chapter 5

and he noted that it was equally criticised for its loose spacing. Overall, however, Watkins concluded with a positive assessment: If these changes can be made i.e. flatten out the whole thing by reducing the slant and ensuring that the lowest characters i.e. finals are only just below the imaginary baseline, then we have a script which we can sell as there is a very good demand for it.132 In the design of the Nastaʿlīq typeface Holloway adopted a working method tailored to the writing style. Rather than drawing typographic character outlines, Holloway initially wrote the letterforms with customised broad nib pens. According to Ross, Holloway felt that like a pianist, he had to practise every day so that his hand remained attuned, and he thus wrote and drew incessantly.133 Holloway would then choose his preferred renderings of letterforms, cut them out and paste them up in groups (Figure 5.44). These incidentally decorative collections then served as a starting point for discussions with Ross, who artdirected the further development with considerations of letter frequency, usability and suitability to the design concept.134 A draft of the font’s software specification, written in July 1986, illustrates the concomitant advance of typeface and logic design. At this stage Holloway commented on the document that had been authored by Ish Parekh, one of the programmers involved in the project. This close collaboration was necessary because the font’s behaviour prescribed the characters required, and vice

Figure 5.44 Nastaʿlīq letter drawings by Holloway, n.d., reduced to 60% linear. DTGC, ­folder Nastaliq 1980s 2 of 3, drawings. 132

Peter Watkins to Fiona Ross, Memorandum, “Nastaliq,” 21 March 1986, Folder Sheeraz / Nastaliq, DTGC.

133

Ross, interview.

134

Ibid.

Beginnings of Digital Arabic Type

373

Figure 5.45 Page from Sheeraz Nastaliq software specification draft (detail), July 1986, ­reduced to 50% linear. DTGC, folder Sheeraz Nastaliq.

versa the characters defined how the logic had to be written. For example two variant forms of a medial ‫ ك‬kāf were programmed in such a way that if multiple ‘base jumps’ caused a long word to extend vertically beyond a certain measure, the shorter form of the letter was used (Figure 5.45). By December a new and more detailed document outlined how the Nastaʿlīq font could be implemented on the CRTonic. In addition to the above characteristics, the specification articulated additional concepts that were invented for the correct processing of Nastaʿlīq. Two of these concepts merit special attention. Because diacritical signs were provided as separate characters, they could be treated independently of the base character that they modified. This flexibility of position could thus be used to prevent collisions between dots and other letter parts. Every character that was narrower than the diacritical signs it could carry was therefore given a specific attribute, called the ‘knockon value’. If such a character was then preceded by another character whose diacritical signs would cause a clash, the diacritics of the preceding character

374

Chapter 5

were displaced horizontally (Figure 5.46). As defined in the specification, “this knock-on effect will ripple on through the word if the next character carries a clashing diacritical”.135 Furthermore, for characters whose elements projected beyond their assigned width, so called ‘zones of influence’ were defined. An algorithm then calculated potential clashes between such protruding elements and marks, and re-positioned the latter to avoid collisions (Figure 5.46). Further algorithms maintained even character spacing after the various positioning routines were completed. As all these functions were interrelated, the font’s logic attained a considerable degree of complexity, and its development became one of the most time-consuming aspects of the project.136 But whereas the software had to be more advanced than the logic for any previous Arabic typeface, the elegance of approach was manifest in the small number of characters required: a draft document describing the principles of character selection and positioning logic stated that in total only around 250 digitisations were required for the entire font.137 Close collaboration between designer, art director, and staff of the R&D department ensued throughout 1986. Holloway played a key role, for in addition to the artwork, he provided feedback on software specifications, desirable font behaviour and assembled documents detailing the contextual character selection principles, contributing decisively to the implementation. Because of the font logic’s complexity, prolonged testing was required to overcome any behaviour issues; in particular the positioning of marks and the diagonal joining of characters – controlled by base jumps – were prone to collisions or incorrect shaping (Figure 5.47 & Figure 5.48). Initially the font was planned for the CRTronic only, but the logic required for its correct shaping could be implemented as part of the general Arabic software, therefore enabling easy switching between the Nastaʿlīq and any other Arabic fonts.138 A milestone chart drafted in February 1987 illustrated the distribution of responsibilities over the final months of the project ­(Figure 5.49). After Holloway delivered the final drawings, the Type Design Studio of Linotype Ltd cut the friskets that were then delivered to Linotype GmbH for font 135

Paul Elsdon, Nastaliq, 8 December 1986, 2–3, Folder Sheeraz / Nastaliq, DTGC.

136

Ross recalls that her Research Assistant, Ros Coates, was responsible for inputting most of the data, and that “the amount of time spent on encodings, contextual look-ups, mark/dot positioning, knock-on values, etc. and testing was enormous, possibly outweighing the time spent on artwork.” Ross to Nemeth, “Chapter 5.”

137

Timothy Holloway, Character Selection, 8 August 1986, 9, Folder Sheeraz / Nastaliq, DTGC.

138

Paul Elsdon, Nastaliq, 25 February 1987, 1, Folder Sheeraz / Nastaliq, DTGC.

Beginnings of Digital Arabic Type

375

Figure 5.46 Left to right: illustrations of ‘knock-on value’ and ‘zone of influence’ concepts, patent drawings (details), reduced to 50% linear. Holloway, Timothy, Michael Fellows, and Fiona Ross. Printing. UK Patent 2208556A.

Figure 5.47 Explanatory note specifying font behaviour (detail), Holloway to Ish Parekh, F­ ebruary 1987. DTGC, folder Sheeraz Nastaliq.

Figure 5.48 Proof of the Sheeraz Nastaliq font, marked up for errors (detail), note that the two highlighted characters do not join; September 1988, actual size. DTGC, folder Nastaliq 1980s 3 of 3.

production which started in April.139 In parallel to the digitisation done in Germany – the typeface required three distinct CRTronic fonts – Linotype Ltd 139

Ros Coates to Jürgen Eger, Letter, “Font Specification for CRTronic Nastaliq,” 23 April 1987, Folder Nastaliq 1980s 3 of 3, DTGC.

376

Chapter 5

c­ ontinued to work on the software development. The company representatives in Pakistan were responsible for samples and promotional material. By August the inventors of the system, Holloway, Fellows and Ross, applied on behalf of Linotype for a patent protecting the novel typesetting concepts in the Nastaʿlīq font.140 Around the same time the daily newspaper ­Nawa-I-Waqt was identified as a first potential customer, and by September the order appears to have been secured as the newspaper was chosen as the site of the font’s beta testing.141 The official launch of Nastaʿlīq was scheduled for 1 November.142 Preparations for the release included naming of the typeface and the drafting of promotional material. A text written by Khambatta in Pakistan named the typeface as Awami Nastaliq and explicitly underlined the differences between the new Linotype font and Monotype’s competing system: With the advent of computers and phototypesetting one manufacturer managed to reach the stage of setting Nastaliq by the tedious and troublesome method of digitising combinations of joined character as one ligature and using computer software to do the needful. This led to numerous problems, the main one being the inclusion of a hard disk system to store the more than 50,000 ligatures existing and ever increasing.143 The official launch had to be postponed, and in mid-November the typeface was eventually called Sheeraz after the daughter of Linotype’s agent in Pakistan, Feroze Khambatta.144 Upon its release, the font could be used on the ­CRTronic 360 and the Linotronic 300.145 The latter, released in 1984, was 140

Holloway, Timothy, Michael Fellows, and Fiona Ross. Printing. UK Patent 2208556A, filed 12 August 1987, and issued 5 April 1989.

141

A. E. Genower to Derek Palmer, Letter, “Nastalique,” 21 August 1987, Folder Nastaliq 1980s 3 of 3, DTGC. The client was interested in a substantial order involving 3 CRTronic machines, keyboards and proofing devices, worth around £  100,000. This order was subject to proof that the typeface was a convincing rendering of the Nastaʿlīq style.

142

Ish Parekh to Michael Fellows, Letter, “Nastalique Meeting 17th September,” 22 September 1987, Folder Nastaliq 1980s 1 of 3, DTGC.

143

Feroze Khambatta, “Urdu Nastaliq,” n.d., Folder Nastaliq 1980s 3 of 3, DTGC. The number of characters stated here appears to be an exaggeration, probably motivated by marketing purposes.

144

Martin Boothman to Dr. V. Stückradt, Letter, “Typeface Trademarks,” 13 November 1987, Folder Nastaliq 1980s 3 of 3, DTGC.

145

Ish Parekh, Nastaliq for Series 300, Sales Release Bulletin, 23 November 1987, 1, Folder

Beginnings of Digital Arabic Type

377

Figure 5.49 “Nastaliq milestone chart,” February 1987. DTGC, folder Sheeraz Nastaliq.

a ­raster scan imaging device working in a similar manner to Monotype’s Lasercomp. It also employed a laser for full page-makeup, and had a resolution of 2,540 lines per inch.146 The similarity of the two machines suggested their comparison, and from Linotype’s perspective it appeared favourable: When compared with Monotype’s system (Amicus [the software] with Lasercomp) our Series 300 system offers a number of significant benefits: Lower cost and more flexible solution Superior language scheme allows output of any Urdu Nastaliq word (Monotype revert to Naskh for some words) Plus we offer Mixing of Urdu Nastaliq and Latin Point size range 1–186 points.147 However, considerable shortcomings were also discovered during tests in Pakistan. Khambatta reported that the advertised output speed of the ­CRTronic Nastaliq 1980s 3 of 3, DTGC. 146

Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 65.

147

Parekh, Urdu Nastaliq for Series 300, 1.

378

Chapter 5

at 10 news lines per minute was disappointing and cause for embarrassment with prospective newspaper customers.148 The agent had been promoting the device as being capable of speeds between 40 and 50 lines per minute, but the output measured during the launch was only 7.79 lines, a figure he considered “ridiculous” for the intended market.149 Khambatta thus enquired whether speed could be increased, as the company’s strategy was based on the significant price difference between the CRTronic and the Lasercomp. Offering the faster Linotronic 300 risked failure of the project because small newspapers could not afford the machine which was priced similarly to Monotype’s system. Beta testing at Nawa-I-Waqt appears to have revealed additional issues, and corrections had to be implemented before the planned public release in January 1988. A further 60 characters were digitised and added to the fonts, and by March Linotype Ltd advised their German colleagues “that this project should still take top priority over everything else”, suggesting continued testing and improvements (Figure 5.50).150 Indeed, it appears that the font’s complexity required review and corrections for months to come, as by August 1988 renewed changes to the font’s behaviour are documented (Figure 5.51). But despite these hurdles, and demonstrating the successful implementation of improvements, by July 1989 a total of ten printers in Pakistan – including the Government Printing Press of Sindh in Karachi and the Nawa-I-Waqt newspaper in Rawalpindi – had functioning Linotype installations with the new Sheeraz font (Figure 5.52 overleaf).151 The approach put forward in the development of Sheeraz was revolutionary, and an official specimen aptly described it as “one of the pioneering designs for digital technology”.152 Here, for the first time an analysis of the Arabic script that fully emancipated its typographic representation from the concepts established in Latin type-making provided the flexibility for an uncompromised rendering of Nastaʿlīq. The division of Arabic letterforms into small, repetitive elements that were assembled dynamically had been contemplated before, notably by Plooij and MacKay, but its implementation as a commercially ­viable product had not been successful. The accomplishment of Sheeraz was 148

Feroze Khambatta to A. E. Genower, Letter, “Nastaliq - CRTronic 360,” 1 December 1987, Folder Nastaliq 1980s 3 of 3, DTGC.

149

Ibid.

150

Gillian Robertson to Jürgen Eger, Letter, “CRTronic Sheeraz,” 28 March 1988, Folder Nastaliq 1980s 3 of 3, DTGC.

151

Ronnie Dubash to Ros Coates, Letter, 31 July 1989, Folder Nastaliq 1980s 3 of 3, DTGC.

152

“Sheeraz” (Linotype-Hell AG, n.d.), Folder Nastaliq 1980s 1 of 3, DTGC.

Beginnings of Digital Arabic Type

379

Figure 5.50 Proof of Sheeraz Nastaliq vowel positioning in various contexts, March 1988, reduced to 75% linear. DTGC, folder Nastaliq 1980s 3 of 3.

Figure 5.51

Illustration accompanying suggestions for font behaviour changes, Holloway, ­August 1988, reduced to 80% linear. DTGC, folder Nastaliq 1980s 3 of 3.

380

Chapter 5

Figure 5.52 (a) Clipping of Nawa-I-Waqt, column composed in Linotype Sheeraz, July 1988, actual size, (b) Detail, enlarged to 200% linear. DTGC, folder Nastaliq 1980s 3 of 3.

Beginnings of Digital Arabic Type

381

Figure 5.53 Detail of “Sheeraz” from LinoTypeCollection: Linotype Laser Fonts/CRT Fonts (Eschborn bei Frankfurt: Linotype-Hell AG, 1992), 368. Actual size. A footnote below the type showing warned potential customers that “to set Urdu/Nastaliq Typefaces special programs and layouts are required”, indicating the imminent obsolescence of the font.

382

Chapter 5

testimony to Linotype Ltd’s pioneering work in non-Latin type-making during in the 1980s, and can be described as the outcome of sustained efforts to establish best practice in this still novel field. Crucial to the success were the pairing of substantial and sound research, built on significant experience, with exceptional artistic skills and ingenious software programming. In particular the latter, often overlooked in more visually-centred accounts, was decisive for the design, and with hindsight appears to have been years ahead of its time. Yet, despite its radically new and largely successful concept, Sheeraz remained a short-lived product with a restricted influence on the trade (Fig� ure 5.53 on page 381).153 Its limited use was not, however, a reflection of the typeface’s quality, but a direct result of changes in technology. Whilst the font was conceptually pioneering, the equipment for which Sheeraz had been conceived was soon overtaken by radical technological change, and became obsolete within years of the font’s inception. The 1980s brought about the latest of the cataclysmic changes type-making experienced during the twentieth century, with arguably the most far reaching effects on the trade. In the early 1980s two Americans developed PostScript, a so-called page description language. It became the definitive expression of the gradual progression towards digital type-making.

153

According to Fellows and Ross the speed problems on the CRTronic could not be overcome. Email messages by Fellows and Ross to the author, “Re: Nastaliq Tests,” 3 May 2013.

Chapter 6

Looking Forward: PostScript and Beyond The End of Machine-dependence The emergence of laser imagesetters in the 1980s raised a new problem of page layout. Generally speaking, before the Lasercomp typesetting machines only set type in galleys. Galleys of type were then made-up with images and elements like rules and borders at the page imposition process. In letterpress printing three-dimensional objects – galleys, wood blocks and furniture – were arranged to form a page; in photocomposition these elements were replaced by film and paper, cut to size and pasted-up to define the configuration of the page. But with the shift to fully digital composition of entire pages, new means of controlling the layout had to be devised. As has been pointed out, lack thereof contributed to the limited success of early imagesetters: In most respects, the Lasercomp was ahead of its time insofar as the full benefits of image generation by raster scan can only be realised when fully made-up pages (including graphics) are presented for output; a prerequisite not easily fulfilled even in the industry of 1983.1 This requirement led to the development of page specification languages, computer readable code which defined “all the graphic elements, both textual and non-textual, that might appear on a page.”2 In lockstep with the emergence of such languages came the development of the Raster Image Processor, or RIP, a crucial element for the new technology. The RIP interpreted the specification and converted the intended appearance of a page to instructions for a given output device. Thus, it produced an instantiation which amounted to the closest possible match of the specification within the limitations of the device. Different output devices could therefore output various instantiations

1 Wallis, “Dates with Phototypesetting,” in Typomania, 57. 2 Southall, Printer’s Type in the Twentieth Century, 164. Southall emphasises the term ‘specification’, as distinct from ‘description’, as the latter would entail a perfect rendition of a page by a given output device. Because every device, irrespective of the fidelity of the emulation of the page specification, introduces variance between specification and actual instantiation – e.g. the printed page – Southall considers the term description as misleading.

384

Chapter 6

Figure 6.1

Comparison of high-resolution phototypesetter output (top) and 300 dpi resolution laser printer output (below). From Wilson-Davies, et al, Desktop ­Publishing.

from a single page specification. The resulting implications are explained by Southall: This is the notion of device independence in imagesetting: the specification of a page to be produced does not have to be adapted to take account of the special characteristics of any particular output device. Thus, for example, corrections to a print job can be made from a marked-up proof produced on a low-resolution laser printer, and the job itself subsequently output on a high-resolution imagesetter, with the reasonable certainty that there will be no unintended changes in the makeup of its pages between proofing and output.3 Although not the first commercial product that successfully implemented these principles, the PostScript page specification language became dominant in the trade.4 Its development began in December 1982 when Charles Geschke and John Warnock, two former employees of Xerox PARC (Palo Alto Research Center), founded the new company Adobe Systems Inc.5 Based on previous work at Xerox, Geschke and Warnock set out to invent a device-independent page specification language. According to Southall, the principles of PostScript were built around a “simple imaging model and a single formalism for 3 Ibid., 165. 4 The first such system was the SuperSetter by the Camex Corporation. Ibid. 5 Pamela Pfiffner, Inside the Publishing Revolution: The Adobe Story (Adobe Press, 2002), 23.

Looking Forward: PostScript and Beyond

385

d­ escribing the outlines of objects of all kinds on the page.”6 Outlines of objects were specified as straight lines and Bézier curves, providing the advantage of easy and lossless scaling.7 A crucial difference of PostScript to earlier page specification languages was that text characters were treated in the same way as any other graphic objects with the benefit that “no special mechanisms were needed in the language to handle character shapes.”8 In early 1983 Apple, then working on its Macintosh personal computer, approached Adobe in order to license PostScript for its upcoming desktop laser printer. Against a substantial investment and advances on royalties, in December 1983 Adobe agreed to develop PostScript for the 300 dpi printer Apple was developing (Figure 6.1).9 Adobe officially launched PostScript in spring 1984 and consequently sought to interest leading typesetting manufacturers in licensing their product. In what has been described as “a leap of faith”, the Linotype group of companies was the first manufacturer to agree, licensing its Times and Helvetica typefaces and committing to develop the first PostScript enabled high-resolution imagesetters.10 The resulting versions of the Linotron 101 and Linotronic 300 were announced in early 1985 and became typical machines of the new technological era.11 Combining the quality of conventional typesetting equipment with the ease of personal computers, it became possible “to drive a 2540 dpi Linotronic 300 imagesetter directly from a standard Macintosh or PC”, where the Linotronic acted “simply as another output device attached to the machine”.12 (Figure 6.2 overleaf) The Linotype group was instrumental in the early acceptance and implementation of PostScript as “the first vendor seriously to recognise the DTP phenomenon [before it was thus

6

Southall, Printer’s Type in the Twentieth Century, 165.

7

Bézier curves, named after the French car designer Pierre Bézier, are mathematical descriptions of curves that are widely used in computer graphics and product design.

8

Southall, Printer’s Type in the Twentieth Century, 166.

9

Previously Adobe aimed PostScript primarily at high-resolution imagesetters. Pfiffner, Inside the Publishing Revolution, 34–36.

10

Adobe first approached Compugraphic, but no deal was concluded. Pfiffner, Inside the Publishing Revolution, 40.

11

John W. Seybold, “An Imagesetter for the Rest of us: Apple’s LaserWriter,” The Seybold Report on Publishing Systems 14, no. 9 (January 28, 1985): 3.

12

Kirty Wilson-Davies, Joseph St John Bate, and Michael Barnard, Desktop Publishing, 3rd ed., Publisher’s Guide Series (London: Blueprint, 1989), 14.

386

Chapter 6

Figure 6.2

Schematic illustration of a setup combining a low-resolution laser proofing printer and high-resolution output device. From Wilson-Davies, et al, Desktop Publishing.

called] as something new, and to develop specific hardware and software links from their traditional equipment in to the new systems.”13 The industry-changing role of PostScript attained even wider public recognition with its implementation in the Apple Laserwriter printer, ­released in January 1985.14 This small and relatively inexpensive device was aimed at the office market, and formed an important element in Apple’s strategy to make inroads into this segment hitherto dominated by IBM.15 Its maximum output 13

Ibid., 158.

14

Jim Bartimo and Michael McCarthy, “Is Apple’s Laserwriter On Target? Delays in Apple’s Mac Office Put Burden of Success on Laser Printer,” InfoWorld, 7, no. 6, 11 February 1985, 15.

15

The Laserwriter was described as “a cross between a Macintosh computer and a Canon PC copier”, heavily relying on the established photocopier technology. It was initial-

Looking Forward: PostScript and Beyond

387

speed of eight pages per minute at a resolution of 300 dpi and compatibility with the ground-breaking Apple Macintosh computer and IBM PCs through the Appletalk network was seen as a competitive proposition for large offices; its far reaching impact for typographical design, it appears, was not anticipated at the time. According to Southall, the system which had been developed by Adobe and Apple with its “combination of computer, page specification language and printer […] was uniquely powerful in terms of its simple interface, its graphics capabilities and its ability to act as a front end to high-resolution imagesetters”.16 Moreover, the Laserwriter provided output of typeset matter on plain paper instead of bromide or film, as had been the case with dedicated typesetting machines. At the same time as Apple’s printer, a so-called “third-party software package” to work with the Apple Macintosh computer was announced by the Aldus company in Seattle: Aldus PageMaker was the first page-layout software designed to work with PostScript, enabling users “to compose a page as if it were being pasted up for printing. When the page layout is satisfactory, it can be sent to the Laserwriter.”17 Here, the notion of WYSIWYG (What You See Is What You Get) display was implemented, contributing to the appeal of the system.18 Hitherto on-screen proofing and editing functions were limited, and commonly the appearance of the final typeset page differed substantially from what the device’s display showed. WYSIWYG profoundly changed the composition process, as it allowed the direct and interactive manipulation of the page configuration. The novelty of the concept, and the impact its implementation must have had on the trade can be gathered from Seybold’s conclusion of a review of existing devices and methods. Writing as late as 1984, Seybold imagined an ideal scenario which, incidentally, became a reality the following year: […] the approach which appears to represent the best of all possible worlds would be to be able to present the text, after composition, in hyphenated and justified form, with true or simulated type fonts in true sizes, and to be able to make corrections in real time and see that the ly priced at $6,995, which, although dramatically cheaper than high-end typesetting equipment, remained expensive in comparison to existing laser printers by Hewlett Packard ($3,500) or the Digital Equipment Corp. ($4,500). Ibid., 18. 16

Southall, Printer’s Type in the Twentieth Century, 171.

17

Bartimo and McCarthy, “Is Apple’s Laserwriter On Target? Delays in Apple’s Mac Office Put Burden of Success on Laser Printer,” 18.

18

Although it was not the earliest WYSIWYG system, the success of the Macintosh and the Laserwriter contributed decisively to the spread of the concept.

388

Figure 6.3

Chapter 6

Article announcing ITC’s publication of Arabic typefaces by the Boutros family, p. 1, reduced to 50% linear. From “ITC Arabic? Yes.”, in U&lc, Upper and lower case, the international journal of typographics, 15, no. 3, August 1988, 8–11.

Looking Forward: PostScript and Beyond

389

corrections are properly incorporated and that they satisfactorily resolve any problems either of format or contents.19 PageMaker thus became the last element in a complete composition system that fitted on to a desktop; a characteristic that was adopted as a marketing slogan by Aldus’ president Paul Brainerd when coining the term Desktop Publishing.20 The technological upheaval these new developments brought about were quickly recognised for their potential to dramatically change the printing trade, leading to the notion of the Desktop Publishing Revolution. Writing in 1988, Wallis noted that “fundamental to the concept [of desktop publishing] is the postulation that the preparation of typographic documents can be undertaken by non-specialist personnel in offices, factories, and commercial enterprises generally.”21 People without prior training or knowledge of typography thus had access to the tools to design documents themselves, a democratisation of the process that was widely praised. Yet, it also implied an immediate erosion of quality standards. The availability of equipment that required little training to produce a typographical document implicitly suggested a redundancy of the trade’s expertise. A large number of publications from this era which intended to provide guidance to the amateur user of type are testimony to the sudden, technology-induced changes of document design.22 Type-making, previously in principle the domain of typesetting equipment manufacturers, also changed profoundly. Adobe’s font format initially had two versions, PostScript Type 1 and Type 3. When it was introduced in 1984 the details of the Type 1 format were not published, “and significant parts of the information in Type 1 fonts were encrypted to protect them from copying or modification”.23 Although Type 3 was public, its inferior quality in low resolution output effectively meant that type-makers had to license the Type 1 format. Linotype’s early licensing arrangement with Adobe, established before 19

Seybold, The World of Digital Typesetting, 221.

20

Robin Kinross, “Übergangsmomente: Schrifttypen von 1968 bis 1997,” in postscript. Zur Form von Schrift heute (Ostfildern-Ruit: Hatje Cantz, 2004), 19.

21

Lawrence W. Wallis, “Forty Years On, a Personal Retrospective and Prospective of Typesetting Technology,” in Typomania, 85.

22

See the following, by no means exhaustive examples: James Felici and Ted Nace, Desktop Publishing Skills : a Primer for Typesetting with Computers and Laser Printers. (Reading, Massachusetts: Addison-Wesley, 1987); Wilson-Davies, Bate, and Barnard, Desktop Publishing; Paul Luna, Understanding Type for Desktop Publishing. (New York: Blueprint, 1992); Pam Bennett, Practical Desktop Publishing. (London: McGraw-Hill, 1991).

23

Southall, Printer’s Type in the Twentieth Century, 169.

390

Chapter 6

the industry-wide recognition of PostScript’s influence, gave it an important head-start over other type-makers in the adoption of their library to the new format.24 Adobe recognised the importance of type for the spread of its technology, and acquired new designs for its type library. From 1984 a licensing arrangement with the International Typeface Corporation (ITC) – the first dedicated typeface publisher without previous history in machine or type-making – provided additional designs;25 and Adobe founded its own type development programme under the direction of Sumner Stone.26 In 1987 the Linotype group of companies became the first major foundry to license PostScript and corresponding production tools from Adobe and began the in-house conversion of its library to the new format. Adobe’s business model was based on these license fees, and with PostScript’s widespread success, foundries and equipment manufacturers became increasingly reliant on Adobe’s technology – a reliance that eventually turned to resentment against the company. Moreover, PostScript-fonts were no longer device-dependent but could be used on any PostScript enabled device, providing unprecedented flexibility to the user and, simultaneously, causing erosion of business for type foundries and typesetting machinery manufacturers alike. As PostScript became the de facto standard of the printing trade through Adobe’s licensing policy, type foundries had to adopt their offer or face obsolescence. Long-standing Adobe employee David Lemon is thus cited: PostScript was like the barbarians at the gate […] It was the end of the exclusive type market. They had to join in or get out of business.27 Adobe’s position increasingly had monopolistic tendencies, and eventually it caused a backlash from competitors Apple and Microsoft through the 24

Cynthia Batty, the manager of Type Design and Development at Compugraphic (by then owned by Agfa-Gevaert), recalls the urgency her department, initially barred from access to the Type 1 technology, felt in obtaining the license from Adobe. Cynthia Batty, email message to ATypI member discussion list, “Re: [ATypI] History,” 17 February 2012.

25

From 1988 ITC also began to publish Arabic typefaces by the Boutros family, which had previously been published by Letraset (Figure 6.3 on page 388). See “ITC Arabic? Yes.”, in U&lc, Upper and lower case, the international journal of typographics, 15, no. 3, August 1988, 8–11.

26

The Laserwriter’s successor, the Laserwriter Plus, introduced in 1986, now held 35 fonts, including typefaces from Linotype and ITC. At this point the Adobe Type Library included more than 100 fonts. Pfiffner, Inside the Publishing Revolution, 59.

27

Cited by ibid., 63.

Looking Forward: PostScript and Beyond

391

d­ evelopment of the TrueType font format. Conversely, the emergence of PostScript and the undermining of the typographical establishment also brought a democratisation of type-making. In 1986 the Texan company Altsys released Fontographer for the Macintosh platform, providing the first mainstream Bézier curve font editor.28 Although initially limited to Type 3 fonts, it opened a field hitherto closed to individuals with a relatively modest budget. At “a fraction of the investment which has traditionally been needed for the manufacturing of type fonts, either in metal, for photocomposition, or for the creation of digital masters”, PostScript fonts could now be made on a desktop computer with a single affordable software package.29 The nominal levelling of the playing field for type-making came in 1990. After the competing TrueType format had been announced in autumn 1989, Adobe was under pressure in a market it had previously dominated. It was consequently compelled to publish the full Type 1 specification.30 Updated versions of Fontographer and other emerging type design software like Letraset’s FontStudio now provided in principle the same tools as Adobe to large companies and individuals alike. Although widely praised for breaking the former dominance of established type manufacturers and providing unprecedented ease for typeface design and production, it was soon recognised that tools alone were no substitute for training, research and experience. Through use of these programs a large number of new typefaces are appearing which are instantly accessible, although many are of poor typographic quality. The ‘professional filters’ have largely been removed. The typographer at the typesetting house who once would have excised errors of judgement is now by-passed as technology allows a complex piece of typography, typeset by the graphic designer on his Mac, perhaps using his own newly-created typeface, to go straight to the output device.31 From the introduction of the Apple Macintosh in 1984, individuals had begun creating typefaces on personal computers, yet they remained limited by the technology. The typefaces then made by Western avant-garde designers, as exemplified by Neville Brody, Zuzana Licko, and Rudy VanderLans, were 28

Kinross, “Übergangsmomente: Schrifttypen von 1968 bis 1997,” 20.

29

Michael Daines, “Some Aspects of the Effects of Technology on Type Design,” in Computers and Typography (Oxford: Intellect Books, 1993), 77.

30

Pfiffner, Inside the Publishing Revolution, 94.

31

Daines, “Some Aspects of the Effects of Technology on Type Design,” 84.

392

Chapter 6

d­ efined by the Type 3 format and the Laserwriter.32 Here, technological limitations were embraced to create a radically new aesthetic with designs that formally reflected the constraints of low resolution printing. Whilst succeeding in turning limitations to their own advantage, these typefaces did so too obviously for their use in text typography. Until 1990 type-making with the most advanced character specification technology still remained the exclusive domain of foundries who had licensed the Type 1 format and the accompanying tools Adobe provided – a domain in which the Linotype group of companies played a pioneering role as the first major type business to adopt PostScript. Linotype’s Arabic PostScript Programme The advance of Linotype’s digital type programme had already begun well before PostScript. In an assessment of the market’s best digitised libraries Seybold reported in 1979 that the company had “embarked on an ambitious program to expand its font collection and (eventually) to digitize the entire library”.33 At this stage it included 1,200 typefaces, adding 240 newly digitised faces per year, a rate at which “Mergenthaler will be able to claim the largest library of digitized type faces by the end of this year”.34 Seybold’s prediction that “every font which the vast majority of people will typeset will have to be typeset in digital form” and “that the successful manufacturers of digital output devices will have to provide extensive libraries of digitized type” proved remarkably accurate.35 This assessment equally applied to Arabic type, a market which had become an important element of the Linotype group’s business. By 1979 D. Stempel AG, the main manufacturer of Arabic fonts for the Linotype group, listed 15 different Arabic typefaces which were expressed in 116 different fonts (Figure 6.4).36 Although some had already been digitised for the Linotron 606, none was 32

Kinross, “Übergangsmomente: Schrifttypen von 1968 bis 1997,” 17–18.

33

Seybold, “Digitized Type: What Is It? What Does It Mean for Typesetting and Word Processing?,” 14.

34

Ibid., 15. Seybold adds that “from what we can see, we doubt that any of the other manufacturers will ever catch up.”

35

Ibid., 17.

36

“Stempel Information, Photocomp Type Faces. Arabic Typefaces” (D. Stempel AG, May 1979), Folder 202 Typeface Lists, DTGC. Fonts were listed according to the machine on which they could be used (e.g. Linocomp, V-I-P, Linotron 303), and font layout, the latter reflecting different language requirements and weights.

Looking Forward: PostScript and Beyond

393

Figure 6.4

“Stempel Arabic typefaces,” 1979, p. 1, reduced to 50% linear. DTGC, folder 202 typeface lists.

Figure 6.5

Diagrammatic illustration of the outline description employed in the Laserfont format. From Otmar Hoefer, Schriftqualität auf dem PC (Bad Homburg: Linotype Library, 1988).

available in outline format, the then apparent future of character descriptions. With the introduction of the Linotron 202 and the CRTronic, a digitisation programme was initiated to convert the existing Arabic typeface library into outlines. By January 1983 eight typefaces in 14 fonts were available for the 202, and by June also seven typefaces in 12 fonts on the CRTronic.37 But with continued technological changes, the library had to undergo renewed adaptations. In 1985 the Linotype group introduced the Laserfont format which was used on the Linotronic 300 imagesetter. It provided arcs as a substitute for the straight line segments hitherto used in the specification of curves, dramatically improving the quality of character reproductions (Figure 6.5). However, because it was released in the same year as PostScript, this proprietary format 37

“202 Traditional Arabic Typefaces,” January 1983, Folder 202 Typeface Lists, DTGC; and “Have I Got the Right CRTronic Arabic Fonts?,” June 1983, Folder Arabic CRTronic, DTGC. Differences in font layout for language support are not counted in these numbers.

394

Chapter 6

was quickly rendered obsolete by the success of Adobe’s product. With the Linotype group’s licensing of PostScript in 1987, a renewed library conversion to the novel format was initiated. At this point Linotype Ltd committed to adapt all existing non-Latin typefaces to PostScript, and in recognition of the scope of the task the so-called special projects department was set up.38 It was managed by Dr Michael Fellows, who recalls its significance thus: […] for the first time there was common management of software, fonts and marketing in recognition of the specific challenges of Non-Latin system development and ensured close collaboration between software developers and the type department.39 For the Arabic typefaces this meant that Linotype Ltd not only converted existing designs and font data, but investigated the new technology’s potential to improve their designs. One example of such improvements was implemented in the approach to Arabic vocalisation. Hitherto the vowels of Arabic typefaces shared the same design between different fonts because of the physical character set limitations of photocomposition founts. But when Linotype consulted Adobe on the conversion of Lotus to PostScript, it was found that the new format offered an “opportunity to revise the diacritical signs with considerable ease”, and new drawings were made (Figure 6.6).40 During the adaptation of the Arabic typefaces to the new format Timothy Holloway was engaged to consult on the process. In a detailed and thorough review of the existing designs, numerous changes were implemented. Individual character shapes were revised and improved to overcome compromises inherited from earlier technological constraints. In particular the increased flexibility of fitting, in PostScript the em had 1000 units instead of 54, contributed to better balanced proportions and spacing.41 Kerning facilities were less constrained, enabling character shapes 38

Fellows emphasised that he reported directly to the managing director of the company, Martin Boothman, in recognition of the importance of non-Latin type. Michael Fellows, interview by the author, 9 August 2013.

39

Michael Fellows email message to the author, “Re: PS,” 30 August 2013.

40

Fiona Ross to Dikram Panossian, Letter, 20 October 1987, Folder 202 Typeface Lists, DTGC.

41

This was a considerable enhancement over earlier digital character representations with strokes. In 1979 Bigelow noted that “if you want to make a straight translation from a phototypesetter font to a digital font, the digital raster must be roughly four times as fine as the analog unit system to give comparable quality of spacing”, i.e. 200 strokes to the em were necessary for artwork based on the 54 unit system. Charles Bigelow and

Looking Forward: PostScript and Beyond

395

Figure 6.6 Design review of floating marks, Timothy Holloway to Lesley Sewell, June 1985, reduced to 25% linear. DTGC, locker C1 folder Arabic Diacriticals.

Figure 6.7

Review process of the character for final ʿayn, Yakout Bold, May 1988, reduced ‫ع‬ to 25% linear. DTGC, folder Arabic PostScript developments.

396

Chapter 6

to project beyond their body, and improved spacing rhythm in text (Figure 6.7 on page 395). A further characteristic improvement of Linotype’s Arabic PostScript fonts were refinements in the positioning of vowel marks in all typefaces (Figure 6.8).

Figure 6.8

Accent test Jalal light, February 1995, actual size. DTGC, locker C1 folder Softdressing proofs.

According to Ross, the above approach was not unusual as Linotype’s general policy took format conversions as an occasion for re-assessments and improvements of existing designs to take full advantage of the evolution of typesetting technology.42 Arabic typefaces and non-Latin fonts in general were the main beneficiaries of this approach, as they tended to be more affected by technical limitations than Latin script designs. Yet, the new format also imposed its own limitations, among them a fixed 256 character limit. The conversion therefore required a corresponding character set definition. As a number of pre-PostScript fonts were larger than could be accommodated, a selection of the most important characters had to be made (Figure 6.9). Whilst PostScript fonts were initially exclusive to the Macintosh platform, the subsequent rise of Microsoft’s Windows operating system with its own character sets required further additions of IBM PC compatible fonts. Because of the character set limitations, multiple fonts of the same typeface were made to cater to different language requirements or font layout changes: many of Linotype’s Arabic typefaces were provided in two versions featuring either Arabic or Latin figures, while others contained Kurdish or Persian characters. Moreover, fonts with programmed behaviour like Linotype Ruq’ah depended Seybold, “Technology and the Aesthetics of Type. Maintaining the ‘Tradition’ in the Age of Electronics,” 14. 42

Ross, interview.

Looking Forward: PostScript and Beyond

Figure 6.9

397

Apple Macintosh Arabic font layout, January 1988, reduced to 50% linear. DTGC, folder 202 typeface lists.

on dedicated software environments. Yet the device-independent PostScript format encouraged cross-platform developments, and soon ­rendered specialised software solutions for typographical composition untenable. Here, the liberation from the proprietary equipment entailed a loss of tailor-made ­solutions.

398

Chapter 6

The common denominators of the PostScript format and the Macintosh platform prescribed the new standards which, in case of the more complex Arabic designs, proved to be regressive, rather than progressive. For example Sheeraz, lacking the dedicated software responsible for its genuine Nastaʿlīq rendering, could no longer be used with the new standard equipment. Notes made in the planning of Linotype’s Arabic library conversion make clear the urgency associated with this latest technological change, which was implemented in parallel to new typeface developments such as Sheeraz. At the end of November 1987 it was planned that three Arabic PostScript fonts be released by the coming April, a rate of almost one typeface per month.43 This speedy transition was further complicated by the novelty of the process: proprietary Adobe tools required for the PostScript development were not finished yet, making Linotype contemplate the use of the commercially available Fontographer as an interim solution.44 A first result of the conversion programme was the launch of the Arabic Series 1000 system in 1988 – a setup specifically conceived for desktop publishing, which was then still considered a distinct branch of the trade.45 It provided dedicated software packages for use with an Apple Macintosh computer that was paired with the Laserwriter and Linotype’s high-resolution imagesetters.46 By 1990 an in-house report by Linotype listed 16 released Arabic PostScript fonts, with a further eight in beta testing and eight in a demo state.47 A comparative chart from the report illustrates the relative complexity of the non-Latin font development process (Figure 6.10). Where character descriptions were not yet in outline form, as was the case with designs derived from the Linotron 606 masters, new outline drawings had to be made. Crucially, it involved a range of processes including software development, an aspect then largely 43

Ross points out that the design work was carried out by four members of staff, two working on Arabic fonts and two on Indian fonts, with external support by Holloway. When Linotype began in-house PostScript font development two additional staff were employed for non-Latin design work. Ross, email message to the author, “Re: Remaining Elements,” 30 August 2013.

44

Fiona Ross, handwritten note, 30 November 1983, Folder 202 Typeface Lists, DTGC.

45

“Arabic System Oasis” (Linotype-Hell AG, 1993), 3, DTGC.

46

A similar setup was already released in 1985 under the almost identical name Linotype Series 100. Wallis, A Concise Chronology of Typesetting Developments 1886–1986, 67.

47

Linotype Ltd, “Non-Latin PostScript Solutions,” 1990, 6, Folder Arabic PostScript development, DTGC. During the same period Indian, Hebrew and Thai typefaces were also produced in the new format. The Arabic typefaces released by 1990 were: Yakout, Al Harf Al Jadid, Lotus, Qadi and Mofid Mahdi.

Looking Forward: PostScript and Beyond

399

Figure 6.10 Comparison of font production processes, Linotype Ltd, 1990, reduced to 40% linear. DTGC, folder Arabic PostScript developments.

Figure 6.11 Oryx Page Make-up. From cover of promotional publication (detail), Linotype-Hell, 1995, reduced to 35% linear.

absent from Latin font development. The report estimated that between 1988 and 1990 Linotype staff edited 64.4 characters per person per month. Based on this rate it was planned to produce 32 new non-Latin fonts, of which 22 were Arabic, entailing some 4074 characters in 1991.48 48

Ibid., 7.

400

Chapter 6

With the commercially successful Microsoft operating system Windows 3.1, released in 1992, Linotype expanded its Arabic desktop programme. The company introduced the exclusive Oryx software, which, according to a promotional publication was the “first dual language page make-up software for the Arabic market”.49 Oryx formed part of the Arabic System Oasis, which built on the earlier Series 1000, combining various workstations into a complete publishing system (Figure 6.11 on page 399). The Oryx software ran under Windows and provided many of the typical features of page-layout software from this period. In addition it benefitted from Linotype’s Arabic type library, a considerable asset given the small number of Arabic PostScript fonts available at the time.50 Upheaval and Convergence, on the Road to Unicode By the early 1990s the initial dominance of the Apple Macintosh as the primary workstation for desktop publishing applications gradually began to fade.51 Adobe already released a PC compatible version of its Illustrator software in 1989, and when Microsoft launched the third edition of Windows in 1990 it became the first successful and widely used operating system from the company.52 In a series of events, often referred to as the ‘Font Wars’, the exclusive position of the PostScript format was challenged. Apple had become increasingly reluctant to keep paying the substantial royalties to Adobe for its PostScript license, and dependence on the format limited Apple’s margin of manoeuvre.53 With PostScript’s success, “it rapidly became obvious to the major system software creators (Apple, Microsoft, and later IBM) that it was important to have scalable font technology supported at the level of the operating system itself. But neither Apple nor Microsoft wanted to have a key piece of their system

49

“Oryx Page Make-Up” (Linotype-Hell AG, n.d.), Folder Arabic PostScript development, DTGC.

50

By 1994 Linotype had released the following typefaces in various configurations for the PostScript format: Yakout, Qadi, Lotus, Al Harf Al Jadid, Ahmed, Mofid Mahdi, Mitra, Kufi, Nazanin, Jalal, Mariam, Amer and Badr. Typographic Development, “PostScript Macintosh & IBM Font Releases” (Linotype-Hell AG, 15 February 1994), Folder Arabic PostScript development, DTGC.

51

The mystifying reverence for the Macintosh within the design community prevailed for much longer.

52

Pfiffner, Inside the Publishing Revolution, 124.

53

The $ 2,739,000 the company paid in 1988 amounted to 33 % of Adobe’s revenues. Ibid., 90.

Looking Forward: PostScript and Beyond

Figure 6.12

401

Illustration and explanatory text about the ATM 3.9 as shown in the WinSoft PageMaker 5.5 ME manual.

software technology controlled by an outside company.”54 The two companies thus collaborated to develop the TrueType format, incorporating a number of improvements over PostScript.55 TrueType was conceived in a way that allowed font files to contain more information for device-dependent rendering. Where PostScript required separate screen and printer fonts, with generally poor display on-screen, TrueType fonts provided all information within a single file. Moreover, the language allowed for dramatically better control over the rendering of type forms in lowresolution environments, making TrueType fonts the de facto standard for onscreen display.56 For Adobe the new format endangered substantial aspects of its business. In an attempt to counter this threat, in 1989 the Adobe Type Manager (ATM) was developed (Figure 6.12). It provided improved on-screen display of Post�Script fonts and the ability to print to non-PostScript printers.57 Because of the ATM’s timely release and TrueType’s belated market-entry, PostScript fonts remained widely in use.58 Indeed, the competitive climate between the companies maintained its antagonistic tone only for a few years. From 1991 Apple 54

Thomas W. Phinney, “TrueType & PostScript Type 1: What’s the Difference?,” truetype-typography.com, 1 October 1997, http://www.truetype-typography.com/articles/ttvst1.htm (accessed 16 May 2017).

55

TrueImage, the printer language Microsoft developed in this context never became successful.

56

For an overview of the font formats see for example Jakob Gonczarowski, “Industry Standard Outline Font Formats,” in Visual and Technical Aspects of Type (Cambridge University Press, 1993), 110–25.

57

Pfiffner, Inside the Publishing Revolution, 96–97.

58

TrueType was officially introduced in Apple’s System 7 in 1991 and the following year in Microsoft’s Windows 3.1.

402

Chapter 6

also included Adobe’s type rasteriser in System 7, and the early 1990s saw a gradual convergence of the parties. It culminated in 1996 in the cross-platform format OpenType which had been jointly developed by Adobe and Microsoft.59 In a parallel development, efforts were made to establish a universal encoding standard for computing.60 The simultaneous emergence of different computer platforms and widely differing linguistic requirements showed that the de facto ASCII standard was inadequate for the rapid evolution of personal computing.61 The 256 character limitation of extended ASCII was insufficient for most languages of the world and reflected the Anglo-centric perspective of early information technology (Table 6.1). With the spread of computers and increasing internationalisation the need for a universal, interchangeable document encoding became steadily more important, and was recognised by the leading IT companies. In the late 1980s Apple and Xerox engineers began discussions aimed at conceiving a universal code for the representation of all characters of the world’s living languages in a process that led to the development of Unicode. The first draft document mentioning the name Unicode was published in 1988 by one of the originators of the concept, Joseph Becker of Xerox, summarising the project’s aims and approach: Unicode is intended to address the need for a workable, reliable world text encoding. […] In the Unicode system, a simple unambiguous fixedlength character encoding is integrated in to a coherent overall architecture of text processing. The design aims to be flexible enough to support many disparate (vendor-specific) implementations of text processing software.62

59

Pfiffner, Inside the Publishing Revolution, 65. OpenType was largely based on Microsoft’s TrueType Open format, an extension to the original TrueType format that provided glyph substitution and positioning functionality. Published in 1994, TrueType Open was designed to compete with Apple’s TrueType GX font technology.

60

Already in April 1988 Apple decided to incorporate Unicode in the emerging TrueType font format. “Chronology of Unicode Version 1.0,” History of Unicode, n.d., http://www. unicode.org/history/versionone.html (accessed 16 May 2017).

61

ASCII is the acronym of American Standard Code for Information Interchange, a character encoding scheme based on the English language and limited to 128 characters.

62

Joseph D. Becker, “Unicode 88” (The Unicode Consortium, 1988), 1, http://www.unicode. org/history/unicode88.pdf (accessed 16 May 2017).

Looking Forward: PostScript and Beyond

Glyph(s)

Unicode Character(s)

A

U+0041 Latin capital letter A

fi f i

U+0066 & U+0069 Latin small letter f followed by Latin small letter i

a a a

U+0061 Latin small letter a

‫�ه �ه‍ �‍�ه‍ ‍�ه‬

U+0647 Arabic letter Heh (positional forms)

Table 6.1

403

The Unicode character-glyph model. Adapted from The Unicode ­Consortium, The Unicode Standard.

A Unicode working group, representing major US software companies, was formed to develop the new encoding scheme with the goal “to remedy serious problems common to most multilingual computer programs”.63 From 1988 to 1991 the Unicode working group grew continuously and developed and promoted the emerging system. By 1991 the first volume of the standard was published, followed by volume 2 in June 1992, containing over 28,000 character definitions. A fundamental tenet of Unicode distinguishes between ‘character’ and ‘glyph’, the former denoting discrete, semantic units of text, the latter their visual representations. In principle Unicode only encodes characters, leaving the consideration of glyphs squarely outside its scope. Thus, the long-held direct relationship between the form and the meaning of a letter was deliberately severed, as illustrated in Table 6.1. The Unicode standard specifies the distinction as follows: Unicode characters represent primarily, but not exclusively, the letters, punctuation, and other signs that comprise natural language text and scientific and technical documentation. Characters reside only in the machine, as strings in memory or on disk, in the backing store. The Unicode standard deals only with character codes. In contrast to characters,

63

The Unicode Consortium, The Unicode Standard. Worldwide Character Encoding, vol. 1, 1st ed. (Reading, Massachusetts: Addison-Wesley, 1991), 1.

404

Chapter 6

glyphs appear on the screen or paper as particular representations of one or more backing store characters. A repertoire of glyphs comprises a font.64 The character-glyph model was decisive for the aspired goal of standardised, multilingual document exchange. Encodings with a direct, one-to-one relationship between semantic units and their graphical instantiation implied a specification of appearance whenever text was encoded. Thus a text containing ligatures of a given font could not be rendered correctly in a different font lacking the specified ligatures. The distinction between characters and glyphs as two discrete layers avoids the problem: a change of font does not alter the underlying code, resulting in a reversion to the individual characters the ligatures were initially composed of. This model has particularly far-reaching consequences for scripts with more complex morphologies than the Latin script, so-called ‘complex scripts’ like Arabic:65 When applications have needed to provide more complicated text processing for complex scripts or sophisticated typography, they have generally made use of proprietary engines to shape text based on custom character sets, or have obliged users to resort to font switching to access variant glyph forms. The idiosyncratic nature of these solutions frequently results in text that cannot be exchanged outside of particular systems and applications.66 While generally welcomed by the computing industry, and widely adopted as the dominant international encoding standard, Unicode has also been 64

Ibid., 1 : 12.

65

Hudson defines them thus: “The term ‘complex script’ refers to any writing system that requires some degree of character reordering and/or glyph processing to display, print or edit. In other words, scripts for which Unicode logical order and nominal glyph rendering of codepoints do not result in acceptable text. Such scripts, examples of which are Arabic and the numerous Indic scripts descended from the Brahmi writing system, are generally identifiable by their morphographic characteristics: the changing of the shape or position of glyphs as determined by their relationship to each other. It should be noted that such processing is not optional, but is essential to correctly rendering text in these scripts. Additional glyph processing to render appropriately sophisticated typography may be desirable beyond the minimum required to make the text readable.” John Hudson, “Microsoft Typography - Windows Glyph Processing: An OpenType Primer,” November 2000, http://www.microsoft.com/typography/Glyph%20Processing/intro.mspx (accessed 16 May 2017).

66

Ibid.

Looking Forward: PostScript and Beyond

405

criticised.67 Although its explicit goal is a global encoding scheme, its initial definition was a 16-bit system, effectively limiting font size to 65,536 characters. The concept argued that this was achievable by including only “modernuse” characters, giving “higher priority to ensuring utility for the future than to preserving past antiquities.”68 Based on the assumption that the characters in all newspapers and magazines printed in 1988 was “undoubtedly far below 214 = 16,384”, the 65,536 characters appeared to provide a comfortable margin.69 Yet, this number fell short of providing sufficient code points for the inclusion of all characters of East Asian Scripts, a shortcoming which elicited considerable controversy and criticism of the approach taken in the definition and later implementation of Unicode.70 Later versions of the Unicode Standard provided means of addressing these shortcomings, providing more than one million character codes. In relation to Arabic, criticism focused on encoding principles. Particularly the assignment of code points to a large number of ligatures contradicts the fundamental character-glyph model, making it a typical example of the oft-criticised inconsistencies of the Standard. Indeed, in its first version Unicode did not encode the allographic variants of Arabic letters, but the merging with other encoding standards and support for legacy encodings introduced this compromise in later versions.71 Other questionable Arabic script encodings have been raised by Milo on various occasions.72 Despite its shortcomings, Unicode (at the time of writing in version 9) became the pre-eminent encoding standard of current computing. In 2001, ten years after its introduction, a business pundit noted: Although Unicode might not be a perfect solution to the challenge of handling all the world’s characters, it has moved us a long way toward being able to create systems that can deal with a wide range of languages.

67

See for example John C. Dvorak, “Kiss Your ASCII Goodbye,” PC Magazine 11, no. 15 (September 15, 1992): 93.

68

Becker, “Unicode 88,” 5.

69

Ibid., 4–5.

70

See for example Steven J. Searle, “Unicode Revisited,” n.d., http://tronweb.super-nova. co.jp/unicoderevisited.html (accessed 16 May 2017).

71

Charles Bigelow and Kris Holmes, “The Design of a Unicode Font,” in Electronic publishing, vol. 6 (Nottingham: University of Nottingham, 1993), 289.

72

See for example Thomas Milo, “Some Comments on the Arabic Block in Unicode,” 1999, http://www.unicode.org/L2/L2001/01329-milo-com.pdf (accessed 16 May 2017).

406

Chapter 6

During its evolution, the original design goals have had to evolve, and in some ways degrade, to meet real-life challenges.73 Microsoft’s Windows NT, released in 1993, was the first mainstream operating system to implement Unicode, but adherence to the standard only spread gradually. The Apple platform merely supported Unicode with the release of its OS X system in 2001, and some mainstream design applications were even later to implement the standard.74 Indubitably, among the most important facilitators of widespread compliance to Unicode was the beginning of the Internet age from the mid-1990s, necessitating seamless multilingual document exchange across platforms. For type-making, OpenType, introduced in 1996, became the most widely used format: Unicode compliant, cross-platform compatible, and with the ability to provide a degree of programmed behaviour within the font file, it has provided increased flexibility and variety for users. Desktop Publishing Tools for Arabic In 1992 – before the publication of OpenType – Robin Kinross recognised where digital type-making offered the greatest potential, irrespective of format: TrueType, too, offers infinitely flexible and intelligent typefaces, which in the case of self-inserting ligatures may be good, and in the case of selfdeciding swash letters may be pretty daft. Or not daft at all, when one remembers that ‘non-Latin’ is the larger part of the world, and that ideographic and otherwise non-alphabetic typography – in which swashes are communication rather than decoration – has been the poor relation, marginalised in every development in typography since Gutenberg. It is here, outside Western typography, that the digital revolution will have its greatest effects.75 73

Suzanne Topping, “The Secret Life of Unicode: A Peek at Unicode’s Soft Underbelly,” 1 May 2001, http://tetrud.com/Billys/Lima/The%20Secret%20Life%20of%20Unicode.pdf (accessed 16 May 2017).

74

Apple’s TrueType GX was also Unicode compliant, but did not succeed as a commercial product. However, significant parts of TrueType GX are still used in the Apple Advanced Typography software, Apple’s proprietary ‘smart’ font technology. The QuarkXPress page layout software only implemented Unicode support in version 7, released in 2006.

75

Robin Kinross, “The digital wave”, Eye, vol. 2, no. 7 (1992): 39.

Looking Forward: PostScript and Beyond

407

In Arabic type-making the effects that Kinross forecast in 1992 were already being felt. As the established type businesses were preoccupied coming to terms with the latest technological changes and the adjustments that they required (with Latin type-making at the centre of attention), new developers began to investigate and pioneer multilingual typography for the desktop publishing sphere. Whilst the new generation of designers operating in the Latin script world was largely concerned with artistic experimentation using the novel digital tools, developers for Arabic and multilingual typography sought to provide products which enabled basic non-Latin script functionality on personal computers. However, there appears to have been remarkably little knowledge transfer from typographical businesses to the emerging technology companies, as numerous cases of the proverbial reinvention of the wheel occurred during this period. Existing achievements of earlier technologies such as character selection and substitution routines, justification principles, keyboard layouts and vocalisation techniques were at best reinvented, and often copied from previous technologies. In 1992 an academic user of Arabic software noted in his review of existing Arabic software: the labor of teaching DOS, Unix, TeX and the like to use Arabic scripts, to go right-to-left and to justify on the right margin, is now behind us. A major operating system is ready to go in any language we wish.76 Apart from the factually incorrect observation about language support, this perspective illustrates the gap between older professional typesetting techniques and early desktop publishing software: the functionality of the newer, in theory more advanced technologies, was lagging years, if not decades, behind their nominal precursors. Arguably, the often hailed Desktop Publishing Revolution therefore also meant, and especially so in the case of non-Latin typography, that the first decades of development served only to achieve capabilities which had been common place in the composing rooms half a century before. A prominent exponent of early Arabic computing software is the British company Diwan Software Ltd, founded in 1985. From 1986 it distributed alNashir al-Maktabi, an Arabic-enabled page layout software that was based on Ready, Set, Go! for the Macintosh, released by Manhattan Graphics in 1985.77 As the first Arabic desktop publishing application, and despite its limited 76

A. Kevin Reinhart, “Arabic on the Macintosh: Overview and Review,” Middle East Studies Association Bulletin 26, no. 2 (December 1992): 162.

77

Ready, Set, Go! was later bought by Esselte Letraset, before being acquired by Diwan.

408

Chapter 6

f­ unctionality, it found considerable use and established Diwan in the field. The company also collaborated with Apple on the development of its Arabic OS, first published in 1987 for System 3 running on the Macintosh Plus.78 To this end Diwan provided four typefaces: Nadeem, Geeza, Kufi and Qahira, yet in its design approach the company made no pretence to originality. According to Boutros et al., when Nadeem was created “the main design principle was to match the popular Yaqut [sic] font from Linotype”, and the nominally distinct typeface Geeza was merely “a larger [sic] version of Nadeem for interface purposes”.79 Although Diwan was a key Arabic software developer in the late 1980s and early 1990s, its position was eroded as competitors entered the market. One such company was WinSoft from Grenoble, France.80 Its reputation was initially built on the WinText application, launched in 1987, which became the standard Arabic word processor on the Macintosh platform.81 WinSoft later developed an Arabic version of PageMaker which succeeded al-Nashir al-Maktabi as the most advanced Arabic desktop publishing software. After Aldus was acquired by Adobe in 1994, WinSoft closely collaborated with Adobe in the development of so-called Middle Eastern versions of its most successful design applications such as Illustrator and Photoshop.82 In consequence, WinSoft attained a central role for Arabic script software, a position it held until the 2010s (see also page 432). Despite the technological break of the digital revolution, many of these endeavours remained based on assumptions which had defined type-making and typography for centuries. The adaptation of software for Arabic typography barely went beyond providing right-to-left composition, basic character selection routines that had been in existence since the late 1960s, and later simple vocalisation techniques, but did not explore fundamentally new ways of making and setting Arabic type. The perpetuation of existing techniques for type-making and typesetting was also furthered by individuals who worked for multiple companies or set 78

The company was initially called Diwan Science and Information Technology Ltd.

79

Boutros et al., Talking About Arabic, 50. The qualification that this was done “with important differences”, namely some modifications for low-resolution laser printing, does not, as Boutros et al. suggest, remedy or justify the illegitimate copying of existing ­designs.

80

The company was first called Unisoft.

81

Boutros et al., Talking About Arabic, 54.

82

Adobe had acquired Aldus in 1994.

Looking Forward: PostScript and Beyond

409

out to establish their own businesses. A case in point is Glyph Systems, founded in 1988. It was formed by Gabriel Habis and Steven Reef, two former Compugraphic employees who had been involved in the development of the EditWriter 7800 (see page 335 and following). Glyph Systems developed an Arabic version of the popular Ventura Publisher software for Windows.83 In the adaptation of Ventura Publisher, Glyph Systems bundled Compugraphic’s Arabic typefaces with its product. Whilst this was a questionable choice in itself (see also page 462), Glyph Systems also carried-over known shortcomings of Compugraphic’s phototypesetter. According to Boutros et al, “Arabic-enabled Ventura continued to experience the same diacritical placement problem, and size difference in text scaling, as the EditWriter 7800” – a startling deficiency given that the physical constraints of the phototypesetter’s exposure window no longer applied.84 The legacy of Compugraphic did not end here, as Glyph Systems divested its Arabic Ventura Publisher to Layout Ltd, a company founded in Cyprus by Gabriel Habis and Joseph Abi Raad in 1989.85 Layout Ltd assumed a key role from 1992 when it started collaborating with a US software company that had gained a significant share of the desktop publishing market. Denver based Quark Software Inc was founded in 1981 and had released its QuarkXPress page layout application in 1987. From version 3, published in 1990, QuarkXPress increasingly dominated the field of desktop publishing software, eventually becoming the market leader by the end of the 1990s. Since 1989 QuarkXPress features the application programming interface XTensions which allows third-party developers to create plugins for added functionality. In 1992 Layout Ltd took advantage of this opportunity and built ArabicXT, enabling Arabic composition with Quark­XPress. Whether any residue from the EditWriter still influenced its development is not known, yet the approach and techniques employed in this product were problematic by themselves. Although Unicode had been published the year before, neither QuarkXPress nor ArabicXT supported it. Indeed, rather than following any established standards, ArabicXT ignored the existing Arabic codepages for Windows and Macintosh, and required instead dedicated fonts that were encoded using the Mac Roman codepage. In what amounted to a hack, ArabicXT fonts 83

Boutros et al., Talking About Arabic, 44. Ventura Publisher was first released in 1986 by Ventura Software, an affiliate company to Xerox. In 1993 the Canadian company Corel acquired the rights to the application and continued its publication as Corel Ventura ­until 2002.

84

Boutros et al., Talking About Arabic, 44. One may surmise that the developers used the codebase of the EditWriter’s software in the development of the Arabic Ventura Publisher.

85

The company later moved to Beirut, Lebanon, from where it still operates today.

410

Chapter 6

thus looked like standard Latin fonts to the OS, but featured Arabic outlines with the character codes for ‘a’, ‘b’, ‘c’, etc. Ironically ArabicXT thus brought back the one-to-one correspondence between semantic value and its representation in type the same year as Unicode’s definite arrival to computing. Text composed with ArabicXT was therefore unintelligible for any other text processor, making the exchange of documents between different users and platforms virtually impossible. Nonetheless, due to the market penetration of QuarkXPress and the lack of convincing alternatives for multilingual and Arabic composition, ArabicXT became widely used throughout the Middle East. Given the questionable approach in the architecture of the software, it comes as no surprise that ArabicXT also became infamous for the poor quality of its fonts and their questionable origins with frequent accusations of design piracy. Against this background, another contributor that historically emerged from within the era of the Desktop Publishing Revolution stands out as exemplifying a fundamentally different approach. The Dutch company DecoType pursued an uncompromising search for novel concepts and principles of Arabic type-making that ignored legacy practice. In the process DecoType pioneered what may arguably be called the most advanced typographical technology in the world. DecoType The origins of DecoType lie outside the domain of typography and can be traced back to Thomas Milo’s background and his interest in comparative linguistics. In the 1970s he studied Slavic languages at the University of Amsterdam and Turkic languages at the University of Leiden, leading him to Ottoman and Arabic language studies, but he did not pursue a career within the confines of academia. Instead Milo worked in a variety of professions – amongst them as an electrician in the construction industry, a city bus driver, a long distance coach driver, and as a truck driver in the Middle East – whilst continually furthering his language skills.86 Further to these activities, and as a result of his Arabic proficiency, Milo was accepted to serve as an intelligence officer and army interpreter in the Dutch contingent of the United Nations Interim Force in Lebanon (UNIFIL) in 1980–81 and 1983.

86

Because of his fluency in Turkish Milo was hired by the Dutch truckage company Rynart as its Middle East transport manager. In this role Milo was stationed in Dammam, coordinating Rynart’s activities in Saudi Arabia. Milo, interview.

Looking Forward: PostScript and Beyond

411

The two other partners that form DecoType are Mirjam Somers, who holds a degree in fine arts from the Gerrit Rietveld Academie in Amsterdam, and practised as a carpenter and architectural designer, and Peter Somers with a background in aeronautical engineering and a career with the Dutch aircraft manufacturer Fokker.87 Milo’s and Somers’ first experiments with Arabic type predate the founding of DecoType and included the creation of a lettering system that was designed to teach the morphological rules of the script (Figure 6.13). It consisted of ply�wood modules that were cut to form a joining pattern, each module having an Arabic letterform painted on its surface. Every module was fitted with a connector, hidden from sight, which only allowed combinations with modules that corresponded to the morphological rules of the script (Figure 6.14). Although based on simple means and entirely handmade, the underlying thinking and craftsmanship of this creation from 1980/81 already demonstrates some fundamental elements of the approach of DecoType.88

Figure 6.13

Left: Wood lettering type case, open with sample text composed on the inside of the lid. Photograph by the author.

Figure 6.14

Right: Side-view of modules, showing the connectors. Photograph by the ­author.

87

Milo and Mirjam Somers are partners in work and in life, Peter is Mirjam’s brother. In the following Somers refers to Mirjam Somers, unless Peter is explicitly named. Beyond this core team, DecoType established a network of experts who contribute to its projects and products, amongst them Stan Jesmanowicz, Glenn Adams, Ruud van Tol, Robert Bringhurst and Mohamed Zakariya.

88

See also Stefania Cantú & Paolo Daniele Corda, La scrittura araba e il progetto Decotype: dai manoscritti alla calligrafia informatica, (Mergozzo: Sedizioni) 2013.

412

Chapter 6

Throughout the following years Milo and Somers studied the Arabic script, its calligraphy, and history. In what amounts to a scientific approach to typemaking, Milo thoroughly analysed Ottoman calligraphic practice to find patterns and rules governing the Nasḫ style (Figure 6.15). In 1983 Milo was pre� sented by a friend with a petition that the printer Ohannis Mühendisoğlu had addressed to the Ottoman Sultan ʿAbd al-ʿAzīz in 1865. It announced the making of a Nasḫ metal fount that was based on the calligraphy of Mustafa İzzet Efendi, according to Milo a design that became a model for most later Ottoman Nasḫ typefaces. Milo acknowledges the profound influence that his historical research and Mühendisoğlu’s work have had on his approach to Arabic type-making: In the quest to understand the aesthetic ideal for Arabic typography this petition provided an essential clue. From it followed logically that new technology had to be developed, as none of the existing methods were designed to meet the traditional calligraphic norms.89 Whilst Milo and Somers began their research into Arabic typography with the Nasḫ style, from late 1983 they shifted their attention towards the Ruqʿä style. Following Milo’s tours in the Lebanon, the Dutch military command commissioned him to write and produce an Arabic language manual that could be provided to the troops during their mission in Lebanon. Consequently Milo considered it as an operational necessity for the printed Arabic to correspond to the image of the script as seen in the streets of the country. However, this turned out to be impossible with existing typographical technology.90 The choice to concentrate on Ruqʿä was also strongly influenced by Milo’s Arabic studies, where he first encountered T. F. Mitchell’s Writing Arabic and developed his own Ruqʿä hand as part of the University’s curriculum. The structural and far-reaching analysis of the style’s morphology in Writing Arabic eventually formed the point of departure for DecoType’s own systematic exploration of Arabic writing styles.91 Beginning with the language manual for the Dutch army, Milo was repeatedly confronted with Arabic typographical technology and typeface design that did not match his expectations, an experience which set him on a course to develop his own tools. 89

Thomas Milo, “A 19th century petition to the Sultan of the Ottoman Empire: How a chance discovery inspired the development of Arabic technology,” unpublished, 2005.

90

Thomas Milo, interview by the author, December 14, 2016.

91

No comparable analysis of any other Arabic writing style has been published to this date.

Looking Forward: PostScript and Beyond

Figure 6.15

413

Detail of a photocopy of an Ottoman manuscript used by Milo in his analysis of morphological rules in the Nasḫ style, reduced to 50% linear. Courtesy of Thomas Milo.

Through the evolution of ideas and concepts first articulated with the wooden lettering set, Milo’s and Somers’ goal then became the development of an Arabic typewriter that was able to reproduce authentic, uncompromised Arabic. At this point they were joined by Peter Somers, whose engineering and programming skills filled a key competence gap. The team thus began with the adaptation of a Brother daisy wheel printer to allow for the composition of the Ruqʿä style. Inspired by Mitchell’s charts of Ruqʿä morphology, Milo and Somers established the character selection logic and conceived of an algorithm for Ruqʿä style shaping with the available 70 characters of the typewriter.92 Peter Somers then turned this data into a programme written in a BASIC language which could control the typewriter through its computer interface. The programme resided on a Zilog Z80 computer and was able to move the typewriter’s scroll in minute steps along x and y axes, providing the basis to reproduce the diagonal joining property of the script. 92

Although 100 characters were available on the wheel, 30 were reserved for punctuation and figures.

414

Figure 6.16

Chapter 6

Close-up photograph of brass matrix and proof casting in resin. Courtesy of DecoType.

As a proof of concept, the team commissioned brass matrices of the allographs required in the shaping of the letter mīm in its four basic positions ‫م‬ (Figure 6.16). From these matrices, and through a reduction process with a pantograph, the required type ‘petals’ were engraved in metal, and could then be mounted to the typewriter’s bars (Figure 6.17). By summer 1985 the first proofs established that the approach taken was working, confirming that it was possible to output Arabic text on a daisy wheel printer with the diagonal joining behaviour typically found in the Ruqʿä style (Figure 6.18). During this period the first affordable laser printers were released (the Hewlett-Packard LaserJet in 1984 and the Apple Laserwriter in 1985), and Peter Somers recognised that it indicated the direction that office printing was going to take, leading to the abandoning of their prototype. Nonetheless, it had laid the foundations for the technology that was at the heart of most subsequent DecoType products: a typesetting computer programme that was conceived for the characteristics of the Arabic script without recourse to concepts originating from Latin foundry type. This programme, initially referred to as the Arabic Calligraphic Controller, became known as the Arabic Calligraphic Engine (ACE). Based on this key asset DecoType was founded in 1985. At this stage, however, no commercial product was yet on offer by the nascent company. During the following years, and whilst continuing ­development on

Looking Forward: PostScript and Beyond

Figure 6.17

415

Close-up photograph of custom type petals made for the Brother daisy wheel printer. Courtesy of DecoType.

Figure 6.18

Detail of a proof from the Brother daisy wheel printer, 1985, actual size. ­Courtesy of DecoType.

DecoType’s technology, Milo pursued other, largely related lines of work.93 Whilst DecoType acted as a re-seller for multilingual-computing products, Milo privately taught courses in Arabic and Turkish, and worked as a typographer and typesetter specialised in Arabic script and multilingual publications. Numerous commissions for various Dutch cultural and governmental organisations not only provided a steady stream of income, but a test-bed for DecoType’s designs and technology (Figure 6.19 overleaf). In this line of work Milo used a range of advanced typewriters – frequently modifying them to achieve added functionality – as well as some of the first multilingual desktop publishing applications, amongst them Multi-lingual Scholar (MLS), Ventura Publisher and Aldus PageMaker.94 93

In the second half of the 1980s Somers was looking after their daughter Sarah, born in 1985, and therefore less involved in Milo’s activities that went beyond DecoType’s own developments.

94

Thomas Milo, email message to the author, “Re: Tasmeem”, 21 October 2016. MLS was

416

Chapter 6

Figure 6.19

Two publications by the municipality of Amsterdam from the late 1980s, reduced to 35% linear. Titles set in DecoType’s earliest Ruqʿä style typeface. Courtesy of DecoType.

DecoType’s pioneer spirit and unapologetic do-it-yourself attitude is exemplified in a commission for the Brill publishing house, an early and ongoing user of DecoType products. The English-Persian Dictionary of Legal and Commercial Terms was issued for the Iran-United States Claims Tribunal in The Hague.95 The texts were provided as two separate digital files in English and Persian encodings, and had to be combined and set in a two-column layout. Lacking a graphical user interface or software capable of the task, Milo and Peter Somers thus developed a dedicated set of routines which read, combined and output the fully made-up pages (Figure 6.20). To achieve this, page layout, line composition and Arabic type composition routines had to be developed, demonstrating the approach that is at the heart of DecoType’s work: wherever existing technology is unsatisfactory or does not yet exist, it is tailor-made for the requirements of Arabic and multilingual typography.96 made by GammaProductions from Santa Monica, CA. 95

K. Fahim, J. Habibion & F. Vittor, English-Persian Dictionary of Legal and Commercial Terms, (Leiden: Brill), 1989.

96

According to Milo, this production is acknowledged as the first fully electronically typeset bidirectional dictionary ever made. Thomas Milo, email message to the author, “Re: Tasmeem”, 24 October 2016. The typeface used in this publication was the only element

Looking Forward: PostScript and Beyond

Figure 6.20

417

Spread from K. Fahim, J. Habibion & F. Vittor, English-Persian ­Dictionary of Legal and Commercial Terms, (Leiden: Brill), 1989, pp. xvii–1, reduced to 40% linear.

From 1985 DecoType abandoned the development of a typewriter and shifted to fully digital type-making. Through its contacts with the PR office of the city of Amsterdam, and some serendipitous personal relations, DecoType had early access to equipment that was out of reach for most individuals. Thus, the drawings for its Ruqʿä style typeface could be digitised with a scanner in the Fokker laboratory that was directed by Peter Somers, and the type could be proofed on laser printers of the PR department of the municipality of Amsterdam (Figure 6.21 overleaf). Furthermore, the US company Hewlett-Packard (HP) became an early supporter of DecoType through its external developer not designed by DecoType. The publication required a Nasḫ style design and because Milo did not have a suitable typeface available he used a modified version of the bitmap font that was bundled with the Multilingual-scholar software package. This font, in turn, appears to have been a clone of Monotype’s Series 549, which in turn was based on the type used at the Egyptian Government Press.

418

Chapter 6

programme and provided hardware at various occasions.97 By 1986 DecoType’s first Ruqʿä style typeface was operational. Whilst its typeforms were still stored as bitmaps, the underlying ACE effectively made it the first ever smart-font created (Figure 6.22). During this time DecoType also sought to find potential partners amongst the established type businesses and emerging software companies. But whilst the former showed little appreciation for the novel approach to type-making, perhaps because existing techniques appeared sufficient or were too entrenched to be revisited, most technology companies either could not take advantage of DecoType’s software or did not see a business case for its development.98 Eventually, in January 1989, an exhibition on Islamic Calligraphy at the Zamana Gallery, London, provided the opportunity for the first public showing of DecoType’s technology. For this occasion DecoType used its Ruqʿä typeface in conjunction with a basic page composition software to typeset all the captions for the calligraphic art on show (Figure 6.23). Moreover, and with the help of HP, DecoType was able to install a PC with a LaserJet printer at the venue and provided live demonstrations of its typesetting system. Figure 6.21 Thomas Milo with his daughter Sarah in front of DecoType’s first Apple Macintosh displaying digitisations of the Ruqʿä typeface on its screen, 1985. Courtesy of Thomas Milo.

Figure 6.22

Right: Proof of DecoType’s Ruqʿä style bitmap font, 1986, reduced to 50% linear. Courtesy of DecoType.

97

Milo, interview.

98

With hindsight it appears to have been an experience similar to Edward Plooij’s a few decades earlier, see page 274 and following.

Looking Forward: PostScript and Beyond

Figure 6.23

419

A caption from the calligraphy exhibition at the Zamana Gallery London, actual size. Courtesy of DecoType.

A leaflet, designed and distributed at the occasion of the exhibition, introduced the software and presented its approach to Arabic typesetting (Fig� ure 6.24 overleaf). The text with the somewhat oxymoronic title “Typesetting Islamic Calligraphy” is structured under five headings: The Arabic Alphabet, Richness and Variety, Printed Arabic, A New Stroke and Riq’ah Script.99 It offers a concise summary of DecoType’s approach to the problems of Arabic typesetting, and provides insights into Milo’s thinking at an early stage of DecoType’s evolution. The text is explicit in its postulations, a hallmark of Milo’s writing, and does not shy away from some generalisations and some categorical judgements: In designing Arabic typefaces it is assumed possible to straighten out those loops and ligatures, the very ones prescribed by calligraphy seen as part of the wider Islamic culture. Arabic lettering is forced into a straightjacket.100 99

The conjunction of typesetting and calligraphy as terms and concepts in the title of this publication highlights a problematic aspect of DecoType. Because of the proximity of DecoType’s products to the calligraphic arts and traditions, in some quarters the company had difficulties to be recognised as a type technology developer, and could not fully shed related associations of conservatism and traditionalism. Later marketing material eschews mentioning the influence of the calligraphic tradition.

100

DecoType, “Typesetting Islamic Calligraphy,” Amsterdam: DecoType, 1989. The sometimes polarising tone and dismissal of other approaches to Arabic type-making are typical for texts published by Milo and may have contributed to the slow acceptance of DecoType within the wider field.

420

Figure 6.24

Chapter 6

“Typesetting Islamic Calligraphy,” promotional leaflet, (Amsterdam: DecoType), 1989, reduced to 50% linear, recto above, verso below. Courtesy of DecoType.

Looking Forward: PostScript and Beyond

421

Here, Milo presented DecoType not so much as designers of type, than designers of technology, and consequently its product was not a typeface, but what was then called the Arabic Calligraphic Printer (ACP) – a software package comprising a page layout application, bitmap fonts in six sizes, and the underlying ACE, all embodied in a standalone Microsoft DOS application. The typeface followed closely the Ruqʿä style and was a ground breaking achievement as such – the style had never before been reproduced successfully in type for machine composition – but here it was primarily presented as a vehicle for the technology.101 Contrasting DecoType’s approach to conventional Arabic type-making, the text is emphatic that the company had built a platform, rather than a single typeface: “The Arabic Calligraphic Printer is a computer program designed to typeset Islamic scripts without distorting them.”102 Irrespective of its marketing, the product DecoType had developed was exceptional. Within the stark limitations of early PC software, which barely allowed for the rendering of strongly simplified Arabic, the ACE reproduced the full morphological repertoire of the Ruqʿä style. The only typographic rendering of comparable complexity and fidelity to one of the ‘hanging’ styles was Sheeraz, developed at around the same time by Linotype Ltd (see page 368 and following). Whilst any knowledge of each other’s designs is improb�able – DecoType had not published any typefaces yet and Linotype Sheeraz was a specialised high-end product for the Pakistani newspaper market – their concepts were similar.103 Due to the division of lettershapes into smaller, recurring elements like pen-strokes and dots, and their dynamic composition with a computer programme, the font could be very efficient: with only 70 type forms the Ruqʿä style was faithfully reproduced (Figure 6.25 and Figure 6.26). Moreover, stylistic variations of letter combinations could be selected by the user, providing a degree of flexibility and sophistication that was years ahead of mainstream desktop publishing tools. Indeed, DecoType’s product was revolutionary, a fact that may not have been stressed enough by its makers: “The 101

The most successful attempts to create a typographic rendering of the Ruqʿä style were done with foundry type for manual composition. However, it remains questionable whether any one of them achieved an equally complete reproduction of the style.

102

DecoType, “Typesetting Islamic Calligraphy.”

103

Edward Plooij’s concepts for Arabic type-making is another example of a comparable approach. But since Plooij’s work remained entirely conceptual it is doubtful that his ideas had any bearing on either Sheeraz or the Arabic Calligraphic Engine. By contrast, Pierre MacKay’s work was highly inspirational for Milo who acknowledges that MacKay’s use of computers for Arabic type-making was a source of influence in his own approach. Milo, interview.

422

Chapter 6

Arabic Calligraphic Printer in its initial version uses riq’ah script. For anyone familiar with the Arabic script the novelty of this will be obvious.”104 Yet, the outcome of the presentation at the Zamana Gallery remained below expectations as, according to Milo, “nobody noticed the technical revolution.”105 Only subsequent showings of DecoType’s work at the first and the second International Conference on Arabic-English Bilingual Computing, in Cambridge, UK, in 1989 and 1990, brought the intended peer attention. From this point a number of cooperations with companies working on various aspects of Arabic computing and word processing emerged.106

Figure 6.25

Selection of type drawings made in the development of DecoType’s first Ruqʿä font, reduced to 50% linear. Courtesy of DecoType.

104

Ibid. Note that Milo used the Turkish transliteration of the style’s name, a nod to its Ottoman origins.

105

Milo, email message to the author, “Re: Tasmeem”, 21 October 2016.

106

DecoType contributed conceptual work to the Nafitha Arabisation of MS-DOS that was developed in Bahrain and cooperated with the Alis Technologies Corporation in Canada, a specialist of multilingual information management. In the late 1990s DecoType also collaborated with Tradigital. Close integration of the proprietary technologies of the two companies did not, as was initially planned, materialise, limiting the scope of collaboration. Ibid.

Looking Forward: PostScript and Beyond

Figure 6.26

423

Sample setting of DecoType’s earliest Ruqʿä font. Courtesy of DecoType.

Of these joint ventures, the cooperation with the Microsoft Corporation (MS) eventually yielded some of the most prominent and lasting results.107 It began when George Hanna, the program manager for Middle Eastern developments who was working at the time on the Arabisation of MS software, contacted DecoType with the intention to integrate their ACE into Arabic MS-DOS or an early version of the Windows operating system. Despite considerable interest by Microsoft, all attempts to achieve this failed, as at the time neither environment could be adapted to host the ACE: according to Milo the responsible developers at MS conceded that their font and line layout technology were too archaic for the task.108 A few years later this first contact resulted, however, in a viable cooperation. In March 1994 Microsoft contracted DecoType for the development of five conventional Arabic fonts that adhered – as far as possible – to its typical design quality. DecoType therefore began to develop a set of fonts in the True­Type format which could be combined to imitate some of the functions that the ACE provided. Based on a design that he had commenced in 1989, Milo developed a Nasḫ typeface as a set of 31 complementary fonts. These included two base fonts, each with three spacing and three Harakaat (vocalisation) variants, four 107

In 1992 DecoType licensed its type technology for the first time. The Dutch company Screencheck, a specialist in identity card printing, purchased a license to use ACE and the first Ruqʿä typeface for the creation of Arabic script ID card. Screencheck, founded in 1991, is still active in this field today and has established a dedicated Middle Eastern company.

108

Milo, interview.

424

Chapter 6

Figure 6.27 Contents of the Microsoft Office 6 retail package for the Arabic market, showing the enclosed promotional poster by DecoType that advertised its DecoType Naskh Professional fonts. Photograph by the author.

Figure 6.28 Samples of typographical options provided through the DTP Naskh font set. Fonts from top to bottom: Variants, Swashes, Harakaat, Phrases. Adapted from the typeface’s manual “DTP Naskh, The DecoType Professional Font Series,” (Amsterdam: DecoType), 1995, reduced to 50% linear.

supplement fonts with letter variants, extensions, swashes, and kashida variants, each in three vocalisation variants, as well as one dedicated font containing ligature glyphs of common pious phrases (Figure 6.28). Microsoft licensed a subset of four fonts which were complemented by a single Ṯuluṯ style font, and bundled the two typefaces as DecoType Naskh and DecoType Thuluth

Looking Forward: PostScript and Beyond

425

with the Arabic and Persian versions Microsoft Word 6.109 As part of the agreement, DecoType could add promotional material to the MS Word package and used this opportunity for the advertisement of one of its own products (Fig� ure 6.27). Contrary to Microsoft, DecoType made the complete set of 31 fonts available for licensing and advertised them with the MS Word package. Sold as DecoType Professional Naskh (DTP Naskh), the typeface amounted to an upgrade of the fonts bundled with the Microsoft product.110 In conjunction with DTP Naskh, a bilingual Arabic and English specimen and manual of 40 pages was published, explaining in detail how the various fonts could be used and combined for best effect (Figure 6.29 overleaf).111 Its introduction made explicit reference to the type’s model of Mustafa İzzet Efendi’s calligraphy, however, this pedigree appears to have served mainly promotional purposes, rather than an accurate description.112 As the typeface could not follow DecoType’s archigraphemic design approach – it had to be conceived within the template of the existing Traditional Arabic font synopsis – its concepts are more akin to foundry type than DecoType’s designs for ACE. Whereas DecoType always strove to sideline typographic legacy in favour of technological advances, here it perpetuated established techniques of typemaking. Whilst Milo was acutely aware of this aspect of the project and its apparent contradictions – with hindsight he called it “one big practical joke” – it allowed him to demonstrate that even with inferior technology DecoType was able to produce outstanding Arabic type.113 109

DecoType Naskh consisted of four fonts: DecoType Naskh, DecoType Naskh Extensions, DecoType Naskh Special, and DecoType Naskh Variants.

110

The double-sided poster was printed by the Drukkerij Rob Stolk B.V.. Prior to opening a print shop Robert (Rob) Stolk (23 January 1946 – 31 March 2001) was an influential anarchist activist and co-founder of the magazine PROVO. Stolk helped Milo and Somers in the early phase of their research about the Arabic script by providing photographic enlargements of manuscripts. Milo, interview.

111

DecoType, “DTP Naskh: The DecoType Professional font series for Microsoft Windows,” Amsterdam: DecoType, 1995.

112

This promotional text was unequivocal about the company’s aspirations: “DecoType, Designers of Computer-aided Typography are the specialists par excellence in the field of Arabic script. A novel combination of linguistic expertise, historical research and computer technology forms the basis of DecoType Arabic fonts designs. What makes DecoType fonts stand out is their unique faithfulness to the great Middle Eastern calligraphic tradition.” Ibid., 6.

113

Milo, email message to the author, “Re: Tasmeem”. The popularity of these fonts remains considerable to this day, as use of the typefaces, as well illegitimate clones of them, are frequently found throughout the ­Arabic script world.

426

Chapter 6

Figure 6.29 Cover page of “DTP Naskh, The DecoType Professional Font Series,” (Amsterdam: DecoType), 1995, reduced to 50% linear.

Figure 6.30

Sample setting of DTP Naskh, demonstrating some of the functionality provided with the 31 font set.

Indeed, and somewhat ironically, DTP Naskh achieved historical significance within the constraints of conventional font schemes as it re-introduced elements of Arabic type that had been lost with mechanisation: DTP Naskh was probably the first instance of a Nasḫ style typeface that was made with a technology other than foundry type which offered a large repertoire of stylistic

Looking Forward: PostScript and Beyond

427

variants, swash letterforms and elongations. The typeface not only provided a rich palette of typographical choices – amongst them the wide form of ‫ك‬ kāf, a novelty in digital type – it was better drawn than many contemporary fonts, tightly spaced and offered a faithful rendering of classical Nasḫ, a feat that could not be found anywhere else at the time (Figure 6.30). Whereas the release of DecoType’s first commercially available typefaces necessitated years of research and development, subsequent products followed at a faster pace. Less than one year after commissioning the first set of fonts Microsoft – this time through the company’s Middle East Product Development department – re-initiated another collaboration. Microsoft had released its Object Linking and Embedding (OLE) technology already five years before, but only in 1995 its potential for DecoType’s ACE was recognised. OLE allowed editing applications to link to and embed objects that were generated with other programmes, in effect extending the host application’s functionality. This meant that typography created with the ACE could now be inserted into documents generated with Microsoft Word, then the market leader in word processing applications. In June 1995 an agreement was reached between Microsoft and DecoType for the development of a so-called OLE server for the rendering of Arabic typography within MS Word. DecoType based the artwork for this font on their original Ruqʿä design, yet redesigned it in an outline format. A substantial portion of the development consisted in integrating DecoType’s ACE in the OLE environment, as well as designing a corresponding user interface – tasks which necessitated close collaboration with Microsoft.114 The resulting product, the DecoType Authentic Ruq’ah Server, was provided as a software package that installed DecoType’s application on the system. The Ruq’ah Server was strongly shaped by the nature and limitations of the OLE environment. Rather than being displayed as a font a user could select from the font menu, content created with the Ruq’ah Server had to be inserted into a MS Word document as an editable object (Figure 6.31 overleaf). The ap�plication’s command opened a dialogue which allowed the user to type Arabic text with a standard system font, and then have it rendered with DecoType’s Ruq’ah application (Figure 6.32 overleaf). Four basic parameters could be con�figured before inserting the object into the document: type size, spacing, and type and background colour. Due to this workflow and the limited typographical choices it offered, the application was essentially restricted to the setting of short titles or headlines; and as continuous text was beyond its capabilities, the potential for the product’s success was impeded. Nonetheless its achievement was significant as the 114

DecoType was further assisted by programmer Stanislaw Jesmanowicz.

428

Chapter 6

Figure 6.31 Interface of the DecoType Authentic Ruq’ah Server in use in the Wordpad application on a Microsoft Windows 10 operating system.

Figure 6.32

Sample setting of the DecoType Authentic Ruq’ah Server.

DecoType Authentic Ruq’ah Server introduced dynamic fonts to the sphere of desktop publishing and word processing, something which had hitherto been accomplished only once before with high-end proprietary typesetting equipment. Here, for the first time a fundamentally different approach to Arabic type-making, leaving behind principles originating from Gutenberg, became readily accessible for the broadest possible user base – a revolutionary development that, however, went largely unnoticed. Further and more visible projects emerged soon after, but were based on DecoType’s conventional fonts. In 1995 WinSoft approached DecoType expressing its interest in the DecoType Professional Naskh font set. WinSoft was developing the Arabic language version of Aldus PageMaker and sought to enhance its typographical repertoire. The two companies thus proceeded to conceive a graphical user

Looking Forward: PostScript and Beyond

429

Figure 6.33 Window of WinSoft PageMaker 5.5 ME showing the DecoType Setter’s controls. From the software manual.

interface that presented the 31-font set and all its options as a single typeface with various typographic features, a tool that became known as the DecoType Setter (Figure 6.33).115 Through four sliding controls the user could configure the amount and distribution of alternative letterforms, swash glyphs and elongations, as well as the spacing of the line. In a second step an individual glyph could be selected and through a keyboard shortcut all alternative shapes of that glyph were displayed, letting the user choose and fine-tune the composition (Figure 6.34 overleaf).116 This project marked the beginning of a collaboration with WinSoft that was going to continue for a number of years. Moreover, in 1998 Apple licensed the same five fonts from DecoType that Microsoft had bundled with its Office product; but whereas DecoType Naskh and DecoType Thuluth were only distributed with Microsoft’s Middle Eastern Office products, Apple included them as part of its operating system, leading to an even greater availability of these typefaces. In the meantime DecoType pursued other projects, amongst them the development of another OLE server for Microsoft which started in 1996. Following the DecoType Authentic Ruq’ah Server, MS now commissioned a Nasḫ style font based on the same technology. According to Milo, this “signalled a new phase in our project to reinvent Arabic typesetting from scratch, ignoring all legacy practice and designs while going back to the best manuscript sources.”117 The resulting DecoType Authentic Naskh Server incorporated experiences gained in the design of the preceding Nasḫ style typefaces, but was built with the architecture of ACE. It was thus dynamically composed from stroke elements and dots and was DecoType’s most advanced and accomplished Nasḫ 115

Thomas Milo, “Balancing Arabic and Latin typography,” Book 2.0, 1, no. 2 (2011): 241.

116

Note that this functionality was introduced for OpenType fonts in Adobe InDesign (the theoretically more advanced successor of PageMaker) in 2016, more than 20 years later – testimony to the non-linear advancement of technology.

117

Milo, “Balancing Arabic and Latin typography,” 248.

430

Chapter 6

Figure 6.34 Explanatory text and illustrations of the controls of the DecoType Setter. From the software manual.

design at the time (Figure 6.35). Although its interface was slightly modified and provided controls for line spacing and kerning, the structural deficiencies of the OLE architecture for typography remained unchanged (Figure 6.36). The DecoType Authentic Naskh Server was bundled with the Middle Eastern editions of Microsoft Office 97, and was integrated with the Arabic Microsoft WordArt functionality. Developed within the OLE environment, using proprietary codepage encodings and being platform-bound, the Authentic Naskh Server was the last DecoType product from the desktop publishing era. But although its inherent limitations and renewed technological change made this product soon obsolete, it provided the basis for the culmination of DecoType’s type-making concepts in the Tasmeem application.118 118

While the OLE concept was designed to work in Microsoft’s Office applications, coinci-

Looking Forward: PostScript and Beyond

Figure 6.35

431

Sample setting of the DecoType Authentic Naskh Server.

Figure 6.36 Interface of the DecoType Authentic Naskh Server in use in the Wordpad application on a Microsoft Windows 10 operating system.

Jointly developed with WinSoft from 2001 and published in 2006 as a plug-in for Adobe InDesign, the successor of PageMaker, Tasmeem is the most comprehensive implementation of DecoType’s technology yet. Built with the ACE architecture, and drawing from usability and interface concepts that were developed for the OLE servers and the DecoType Setter, Tasmeem provides an intuitive, fully integrated interface for the currently leading page layout application; moreover, Tasmeem is Unicode-compliant and thus bridges DecoType’s technology with conventional typographical tools and standards.119 It features dentally it was also supported by PageMaker. In-house DecoType could therefore combine its OLE typefaces with PageMaker and use it for some typographical works. Milo, interview. 119

Note that Thomas Milo was involved with ISO standardisation and contributed to the Unicode Standard already from 1988. In 1989 he consulted Adobe on its Cyrillic encod-

432

Chapter 6

a small range of typefaces, notably new DecoType designs in the Nastaʿlīq and Ruqʿä styles, and provides controls for spacing and mark placement that allow significantly greater precision and flexibility than achievable with conventional type formats. Furthermore, DecoType’s Tasmeem typefaces provide the rich palette of stylistic variations and elongations known from earlier DecoType designs, which are easily accessible through a so-called ‘Word Shaping’ selection interface (Figure 6.37). Yet, critically, this functionality does not remain limited to individual words. Through another interface called ‘Text Shaping’ detailed settings for the distribution of specific glyph variations, elongations and swashes can be controlled and saved as styles for later use (Figure 6.38). Such defini�tions can then be applied to entire sentences, paragraphs or indeed a complete text. Owing to DecoType’s archigraphemic design, substantial language support and extended typographical functionality do not result in large file sizes (Figure 6.39 overleaf). ACE fonts are therefore more compact than conventional Arabic fonts, which may perform poorly because of their architecture. For example, Naskh Regular, the most elaborate font bundled with Tasmeem, requires approximately 550 glyphs for full Arabic Unicode coverage and its vast range of letter variations; by contrast, Microsoft Arabic Typesetting, a font which aspires to perform similarly but is designed within the OpenType architecture, relies on around 2,500 Arabic script glyphs. The Tasmeem ­‘Position Tuner’ gives the user the ability to re-position marks, letters or entire lettergroups through drag-and-drop of the mouse pointer – another function that is possible because of DecoType’s underlying ACE technology (Figure 6.40). Due to its unique ca�pabilities, Tasmeem became a benchmark for Arabic typesetting technology, and influenced Arabic type-making practice; notably so in a number of recent OpenType fonts which borrowed heavily from its concepts. Yet, despite the technology’s pre-eminence, the commercial success of Tasmeem remains shy of its apparent potential. Made for InDesign, Adobe’s page layout application, Tasmeem is unavailable for users of other popular design applications such as Illustrator and Photoshop, or indeed the entire sphere of office software. Furthermore, WinSoft pursued an opaque and inconsistent marketing and pricing strategy, contributing to slow and limited sales, difficulties that were aggravated through technological developments; WinSoft’s business model relied heavily on providing the basic Arabic and Hebrew functionality and interfaces for Adobe’s popular software range. However, increasing ing and DecoType provided the Arabic glyphs shown in the Unicode Standard, a contribution which earned the company the status of Associate Member. The relatively late adherence of a DecoType product to the Unicode standard was only due to the technological limitations of the vehicles for its earlier products: TrueType, OLE and PageMaker.

Looking Forward: PostScript and Beyond

Figure 6.37

433

Word Shaping interface of Tasmeem 6 in Adobe InDesign CS 6 illustrating the plethora of variant renderings that a long word yields when using the DecoType Naskh Regular typeface.

Figure 6.38 Control of letter variations with the Text Shaping interface using the DecoType Nastaliq typeface. Note how specific letter variations can be deactivated or the frequency of their occurence in text defined through sliding controls.

internationalisation and a search for new markets led Adobe to provide much of this functionality built into its products, effectively making WinSoft’s localisation services redundant. As no Middle Eastern version of Adobe software

434

Chapter 6

Figure 6.39

Illustration of the break-down of lettershapes into smaller, recurring elements, one of the principles that lies at the heart of the ACE. Set with DecoType Nastaleeq Press, the company’s latest typeface that was conceived in line with the Nastaʿlīq style that is preferred in subcontinental use.

Figure 6.40

The Position Tuner tool’s interface allows the user to alter the position of individual marks, diacritical signs or entire letter groups. Modifications can be made numerically or through drag-and-drop with the mouse pointer.

was required anymore, WinSoft lost a substantial part of its customer base, and consequently Tasmeem’s potential market share became severely curtailed. Irrespective of the outlook for Tasmeem, the contribution of DecoType to Arabic type-making has been remarkable and its influence is here to stay. The company’s pathway has been unusual as its approach to type-making fundamentally challenged principles followed by the trade since the fifteenth century. The fact that DecoType, a three-person company run from a small apartment in Amsterdam, succeeded in turning this radically new way of designing Arabic type into viable, mainstream products without the backing of any major established type or software company is exceptional, and proof of the quality of its work. Indeed, DecoType exemplifies how the Desktop Publishing Revolution, and more generally the digital era, opened up the trade to newcomers and outsiders, and how small companies and individuals were now able to play a formative role with minimal capital and infrastructure.

Chapter 7

Complexities and Simplifications: Factors in the Evolution of Arabic Type-Making Progress through Technology In 1916 two authorities on technical matters of typography, Lucien Alphonse Legros and John Cameron Grant published their seminal work Typographical Printing-Surfaces: the Technology and Mechanism of Their Production.1 Discussing a range of subjects, Legros and Grant provided a thorough account of the trade at the beginning of the twentieth century. Similar to the present book, Legros and Grant emphasised the defining role of machinery on the appearance of typographical printing surfaces. The scope of their study included amongst numerous other subjects also the typographical representation of the Arabic script, thus discussed in their chapter on legibility: In arabic [sic], legibility is greatly dependent on the dots and their combinations […] This character [script], if it were freed from the complication of the small vowel signs and were made less inclined, would be one of the most legible scripts. Amongst oriental peoples, especially Mohammedans [sic], this face is so largely used that its improvement in the direction of increased legibility, easier composition, and diminished kerning, would enable it to take a very much higher rank than at present. Excessive kerning necessitates the use of soft metal, which gives a poor printing-surface, and, by the yielding of the type, decreased legibility. The influence of modern mechanical methods in diminishing this evil is a great aid to the obtaining of clean, clear printing.2 While outlining the recognised difficulties arising from the reproduction of the Arabic script in type, Legros and Grant also evoked the potential benefits mechanical composition could bring to the peoples of the Arabic script world. Here, a defining theme for the evolution of Arabic type emerged, in which existing technology is seen as prescribing the direction of progress.

1

Legros and Grant, Typographical Printing-Surfaces.

2

Ibid., 184.

436

Chapter 7

The typographical issues Legros and Grant identified – namely small diacritical signs and vowels, the joining of letters, and the need for kerning – were problematic in view of the constraints of the Linotype, the only typesetting machine capable of composing Arabic at their time. Legros and Grant took the characteristics of the technology for granted, and considered their prescriptive influence on type unavoidable, and possibly a necessary evil. Indeed, trust in the technology led to the belief that its limitations may work as a corrective to the ‘evils’ that were identified in the Arabic script. The successful development of an Arabic linecaster appeared as a feat which demonstrated the unlimited possibilities of technology: The arabic [sic] Linotype machine composes over one hundred and fifty sorts [sic] from the keyboard, in addition to the other sorts which may be inserted by hand in the line of matrices when necessary. [...] The form of character is perforce more upright than would be the case were kerning permissible, but the result attained testifies to the skill of those workers at the subject, who have overcome the really great and exceedingly complicated typographical difficulties involved in adapting the arabic [sic] character to the restrictions of the composing machine.3 Legros and Grant had no doubt that the Arabic script had to be adapted to the restrictions of the composing machine, rather than a composing machine be developed which could reproduce the Arabic script without undue compromises. Simplification, as it would later be called, seemed to be the route Arabic would have to take in typographical composition. Legros and Grant, however, were not alone in their conviction that progress shaped by technology was desirable. Judging from the account of Salloum Mokarzel’s contribution to the development of the first Arabic Linotype, there was no reason to have any regret or concern about the compromises which had to be sought in adapting Arabic to the restrictions of the machine.4 After all, the Linotype stood for progress, education, literacy – indeed, in the case of al-Hoda it stood for the free Press and emancipation from repressive Ottoman rule. In the early twentieth century most of the Arabic script world only began to experience the revolutionary power of large-scale printing and publishing, and its potential was recognised and embraced. Then the Linotype was the most powerful tool in the chain of 3

Ibid., 542.

4

See Mokarzel, Al-Hoda, 1898–1968.

Complexities and Simplifications

437

processes that contributed to the further acceleration of print production, and consequently to a lowering of its costs and an increase of its reach. It comes as no surprise then that aesthetic considerations were side-lined in favour of the agency for change that mechanical composition promised at the time. So it appears that when Arabic was adapted to mechanical composition, imperfections of typography were acceptable as long as the content could be conveyed efficiently. In the absence of solidly established conventions of what constituted quality in Arabic typography – widespread and high volume printing were still novelties in the Arabic script world, not to speak of a developed typographical culture – mechanical composition attained particular significance. The gradual evolution of Latin script typography over 450 years has no equivalent there, as newspapers composed on a Linotype were amongst the earliest widely-read forms of Arabic typography. Indeed, for the first 30 years of mechanical Arabic typography exclusively linecasters defined its image, for Monotype’s earliest Arabic fount was only released in 1939.5 The Monotype provided a more refined mechanism, lending itself better to the higher complexity of Arabic script morphology than the Linotype; yet its use remained limited to the domain of quality book typography. For newspapers Linotype and its competitor Intertype dominated the market. Indeed, speed and economy of composition remained decisive elements in the further evolution of the form of Arabic type. In the mid-1940s, when Monotype demonstrated that Arabic hot-metal composition could attain a quality comparable to foundry type with substantial productivity increases, the ill-informed reform project of the Academy of the Arabic Language still sought to fundamentally change the Arabic script. Here, it appears that the ideological conviction in progress through technology was indeed so strong that even the complete abolishment of the Arabic script was considered as an option. But the prolonged project brought no results, and eventually all proposals were rejected.6 By the 1960s, when the project was abolished, the Academy had perhaps recognised the misconceptions on which 5

From the early 1930s Intertype also provided Arabic founts on its linecasters. Since Monotype’s début in Arabic type-making coincided with the outbreak of the Second World War, the distribution and use of Series 507 probably remained marginal before the late 1940s. Even then the typeface was not popular in the Middle East, and Monotype only gained a significant market share in Arabic countries with the introduction of Series 549 in 1947.

6

As has been demonstrated, the rejection of the proposals did not stop some of its inventors from pursuing their ideas further, sometimes using the notion of script reform to advance personal agendas.

438

Chapter 7

Figure 7.1

Left: “We write a different Arabic,” advertisement in an-Nahar (detail), 1973, ­reduced to 35% linear. DTGC, Box Simplified Arabic.

Figure 7.2

Right: Cutting from an-Nahar (detail), 19 April 1977, enlarged to 200% linear. DTGC, Box Simplified Arabic.

the notion of script reform was based. At this time other solutions, both conceptual and technical, allowed for practicable, efficient Arabic composition, the primary goal the reform project had aimed at. The development of Mrowa-Linotype Simplified Arabic stands out as the most influential design in which practical needs dictated the form of printed Arabic. The flourishing newspaper market of the late 1950s and 1960s spurred further increases in composition speed and the reduction of costs; societal progress, embodied in an active Press, was thus connected to technological advance. But whereas earlier typographical representations of Arabic were principally based on a rationalisation of manuscript forms and the reduction to their minimal allographic repertoire, with Simplified Arabic the script was

Complexities and Simplifications

439

deliberately altered to conform to the characteristics of a single-magazine l­ inecaster. Subsequent widespread adoption of simplified typefaces by newspapers throughout the Arabic script world contributed decisively to their acceptance among the reading public. In addition to Intertype’s and Monotype’s designs, a lesser known example is the Arabic 90 type by the Lebanese newspaper an-Nahar.7 It was introduced in 1973 and the newspaper promoted it with a pragmatic rationale, noting that “standard type forms wasted space” and were “often bulky and difficult to read”.8 According to Hrant Gabeyan, the founts were made by Intertype from designs by an Englishman, a deliberate and revealing decision: Walid Tueni [of an-Nahar] says that he did not want to give this job to an Arab calligrapher because he knew that it was going to be extremely difficult, and almost impossible to convince an Arab calligrapher to produce a typeface where the shape and legibility come before the beauty and tradition.9 (Figure 7.1 & Figure 7.2) Gabeyan described Tueni repeatedly as an exceptionally committed proponent of the simplification of Arabic. This commitment to simplification, it appears, was not shared by craftspeople from the region, as Tueni had to resort to the help of a foreigner to pursue his design ideas for the Arabic 90 type. Nonetheless, supported by its practical advantages, simplified type eventually became the de facto standard of newspaper typography. The success of MrowaLinotype Simplified Arabic also exemplifies another trend in the evolution of Arabic type in the twentieth century: international collaboration. Collaboration between East and West The origin of Arabic typesetting machinery in the Western hemisphere resulted in a Latin script bias that is apparent in the principles of type-making and 7

An-Nahār (the day).

8

“We Write a Different Arabic”, Advertisement (An-Nahar, 1973), Box Simplified Arabic, DTGC. The fact that an advertisement for a new Arabic typeface is written in English may be interpreted as a Lebanese peculiarity. By implication, it appears that the advertisement was aimed at an English-reading audience, rather than the actual readers of the new Arabic typeface.

9

Hrant Gabeyan to Walter Tracy, Letter, “‘Arabic 90’ — Intertype,” 15 June 1973, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC.

440

Figure 7.3

Chapter 7

Page composed in the hot-metal version of Monotype Series 549, reduced to 70% linear. From Vincent Monteil, Anthologie bilingue de la ­littérature arabe contemporaine, (Beirut: Imprimerie Catholique, 1961), 125.

Complexities and Simplifications

441

composition. Its negative influences on Arabic type have been demonstrated and juxtaposed with the positive effects that technological progress entailed. Yet besides the strictly technical influence of Latin-centric machinery on Arabic type, also on a different level the design and manufacturing processes reflected the Western origin of typesetting technologies. The companies in the vanguard of technological innovation initially lacked cultural and linguistic competence, and therefore had to seek the skills and knowledge for successful Arabic type-making elsewhere. As a consequence, international collaboration became a central theme of Arabic type-making in the twentieth century. As shown in chapter two, the impetus for the first Arabic Linotype came from the Lebanese migrant Press in the United States, and the collaborative effort between Salloum Mokarzel and Mergenthaler Linotype introduced the kind of joint venture that became a recurring feature of the evolution of Arabic type. Mergenthaler’s technological know-how and Mokarzel’s script expertise were indispensable elements for the successful implementation of Arabic on the Linotype, and subsequent examples continued the trend of collaboration that they had pioneered. Roughly 40 years later Mrowa-Linotype Simplified Arabic epitomised – in both name and significance – the potential of cooperation between Middle Eastern client and Western manufacturer. Around the same time as Simplified Arabic’s début, Monotype pushed the existing technology in a different direction. Series 549, and the subsequent Series 589, were attempts at providing the best possible typographic rendering of Arabic with the available technology. Whereas the latter was a remarkable engineering achievement, Series 549 continued Monotype’s tradition of adapting high-quality movable type designs with great fidelity and set a standard to emulate (Figure 7.3). Roughly 20 years after its release Gabeyan still considered it the benchmark of Arabic type-making: None of our existing Arabic typefaces, L&M [Linotype] or MLCo., [Mergenthaler] is good enough to compete with the Monotype design which at the present time is considered to be the best on the market. Therefore, a new Traditional Arabic typeface must be designed.10 The success of Series 549 derived from the interplay of three factors: quality and selection of the model, cultural and script competence for the fount’s conception, and technical and mechanical expertise for its implementation. (1) The original artwork was of irreproachable quality as the design was based on 10

Hrant Gabeyan to Richard J. Caesar, Letter, 2 February 1973, WT correspondence, folder 18H Simp Arabic on Linocomp VIP 505C, DTGC.

442

Chapter 7

renowned foundry type. Indeed, a typeface which had been deemed worthy for the composition of the Qurʾān would have satisfied the most discerning judge.11 Yet at that time, Monotype had neither the experience nor the subject-specific knowledge for an informed choice of artwork, and therefore relied on its representative in Egypt, Joseph Lindell. (2) Lindell’s close collaboration with the client Dār al-Maʿārif provided not only the model but also decisive elements for the development of Series 549. The representative’s influence contributed to the shaping of the fount synopsis, the keyboard arrangement and the actual design of character shapes; furthermore, for Series 589 fundamental technical principles of vocalisation were articulated and proposed by the client. (3) Whereas the contributions of Lindell and Dār al-Maʿārif were essential, the considerable technical know-how and ingenuity for the fount’s implementation lay with Monotype. At the Type Drawing Office the original artwork was unitised and large pattern drawings were made from the original fount – a process requiring considerable interpretative skills. The Monotype Works defined the matrix-case arrangement, conceived the keyboard layout and developed the mechanics for right-to-left composition. Indeed, the extraordinarily fine mechanical tolerances required for smooth joins, wide kerns and small type forms were as much part of the type’s design as the choice of the artwork. Inevitably, the technology also imposed constraints on the appearance of this type, for example in the form of a limited character set. Here, an important role fell to the Monotype Works which, building on the flexibility of the system, developed new mechanisms for improved Arabic composition. The significance of the collaborative development of Series 549 and 589 is further underlined when comparing them to later Arabic typefaces by the Corporation. As Monotype never built a dedicated team for non-Latin typemaking, designs made for less informed and exacting customers than Dār alMaʿārif, and produced without the guidance of Lindell, betrayed the Corporation’s lack of Arabic expertise. Monotype’s best and most successful Arabic typefaces were made for hot-metal composition, and resulted from close cooperation between specialists with cultural sensitivity, artistic skill, and technical excellence. ∵

11

Note that nothing suggests that Monotype and its consultants were aware of this specific use. The typeface’s status derived generally from its use for publications of the Egyptian Government Press.

Complexities and Simplifications

443

International collaboration in Arabic type developments continued over the course of the following decades, proving its value irrespective of changes in technology. As discussed in chapters two and three, Hrant Gabeyan played a key role in Linotype’s Arabic developments for most of the 1960s and 1970s, in effect acting as Arabic typographical consultant for the company. Gabeyan’s case also demonstrates that cultural background only partially plays a role in the field of typography. Gabeyan was raised in Egypt in an Armenian, Christian family, learning Arabic and French at school and Armenian at home. He never designed a single typeface, but his intimate knowledge of the requirements of the market, the competitor’s offers, and the interplay between technology and design made him an invaluable contributor to Linotype’s Arabic type programme. His mark was left on numerous inventions and designs, among them the first computer-aided Arabic composition system for al-Ahram – the precursor of all Arabic selection routines – the names of some of the most popular Arabic typefaces of the twentieth century, the design of Badr (previously named Osman), keyboard layouts, improved fount synopses and the first adaptations of Arabic typefaces to Linotype’s photocomposition machines, amongst others.12 By contrast, Gabeyan did not contribute to either Armenian or Latin script typography. As Linotype’s representative in Egypt, Gabeyan was the interlocutor of Walter Tracy, the company’s typographical adviser. From this work relation, documented in ample numbers of business letters in the Non-Latin Type Collection, DTGC, emerge diverse influences on the evolution of Arabic type design. Gabeyan repeatedly encouraged more developments and the improvement of existing offers, advocating new designs and the elevation of quality standards. As demonstrated, for example, in relation to the development of Osman, Gabeyan’s support of innovation was not always shared by Tracy (see page 243 and following). Whereas Gabeyan urged Linotype to commission a new design, Tracy sought to revamp old hot-metal founts for photocomposition, indicating a stance in which pragmatic and economic factors were at least as important as quality improvement. But although Tracy’s vision of Arabic type design was heavily influenced by business considerations – as also evident from his advocacy of Simplified Arabic and his rebuttal of Plooij – he valued Gabeyan’s views, and their collaboration proved remarkably fruitful. By 1975, toward the end of Gabeyan’s employment with Linotype, Tracy noted that “for opinions on Arabic type designs, I rely on Hrant Gabeyan, one of our representatives in

12

According to Ross the invention of the phonetic keyboard for Bengali was also directly influenced by the Arabic selection logic developed in 1967 for al-Ahram. Ross, interview.

444

Chapter 7

the Middle East”, demonstrating his influential position.13 Prior to his departure, Gabeyan recruited a successor who came to play a similarly important role in the following decade. Dikram Panossian worked at the Middle East Liaison Office (MELO) and provided the UK office with similar insights into the trade, language support, and assessments of market needs. Panossian’s interlocutors in the United Kingdom also changed when a generational shift happened at Linotype. Although Tracy’s retirement initially left a gap that could not be filled, once a successor had been found in Fiona Ross the importance of collaboration, both local and international, was quickly recognised and actively pursued. From commissions of Timothy Holloway as external designer, to Haghighi’s involvement in the adaptation of Nazanin to digital typesetting, and the cooperation with Adobe and other companies, a broad range of skills and expertise that could not be covered in-house was sought. In the late 1970s and 1980s a changed technological environment furthered the need for the cooperation between multiple contributors as the rise of computing elevated the role of software development in type-making to an unprecedented level. In addition to the cooperation between staff, external designers, company representatives, and regional offices such as the MELO, computer programmers now became a decisive element in the skill mix required for successful Arabic type developments. Moreover, it is worth bearing in mind that the specialisation of the diverse tasks did not imply that experts in one domain could afford ignorance of other factors. On the contrary, it appears that the most successful collaborative projects arose when all contributors had some knowledge of all elements that were involved. Lindell and Dār al-Maʿārif’s staff not only provided cultural expertise, but had notions of the technical requirements of the Monotype, and Mrowa’s know-how of printing and typesetting was decisive for the invention of Simplified Arabic. Similarly, Gabeyan’s insights into cultural appropriateness and his knowledge of technological constraints led him to successful proposals for the improvement of Arabic composition with three different companies: Linotype, Klett-Interpart and Compugraphic.

13

Walter Tracy to Michael Parker, Letter, “Albert Botros: Arabic Type Designs,” 12 May 1975, Box N9940, NMAH. In the event, Tracy advised Mergenthaler not to publish type designs that had been proposed to it by a certain Albert Botros. Any relation to Arabic type designer Mourad Boutros could not be established.

Complexities and Simplifications

445

Cultural Identity Expressed through Type The cultural identity of a typeface is an abstract concept, with little scope for objective, measurable definitions; its discussion risks resorting to common place reasoning and clichés. Although the formal properties of a typeface can be described, the association of certain characteristics with specific identities, or indeed cultures, is prone to generalisations, and the evocation of stereotypes. A famous example is found in Pierre-Simon Fournier’s comments – as conveyed by Harry Carter – about types that were made in the eighteenth century Netherlands. According to Carter, the rivalry between the French typefounder and his Dutch colleagues, notably Joan Michaël Fleischmann, led Fournier to the observation that the popularity of condensed, space-saving designs which he called the ‘goût hollandais’, demonstrated an alleged Dutch trait of seeking profit over all other concerns.14 The difficulties in describing national or cultural typographical preferences without resorting to stereotypes may account for the lack of such discussions in current literature on Arabic type. Broadly speaking, Arabic type is associated with Arabic language and culture, paying little consideration to other cultures and languages using the same script. In treatises on calligraphy, the umbrella term ‘Islamic calligraphy’ is frequently employed to underline the unifying faith of the culturally diverse Arabic script world; but whereas the traditional proximity of calligraphic practice to religious expression justifies this terminology, it would be inappropriate to refer to typography in the same way. Arabic type is used for the composition of secular and religious texts in numerous languages, and considering the breadth of cultures using the Arabic script it should come as no surprise that distinct typographical identities emerged. Yet the differences of such identities are rarely acknowledged or described, by implication suggesting congruence between Arabic culture and Arabic type. But as has been demonstrated throughout this study, influences from other, 14

“The descendants of the Elzeviers [sc. Dutch printers] are more shopkeepers than artists. They judge books by the profits they make. Not content with the superb round types of their forebears, they have procured a new variety more pleasing to their commercial mentality. The desire to economise upon the space which a well-made letter has a right to occupy has led them deliberately to acquire types of a cramped, starved look, so that they may get in more words to the line and more lines to the page. They are not troubled by their ugliness, provided they are profitable.” Quoted by Harry Carter, “Fournier on Typefounding,” in The Manuel Typographique of Pierre-Simon Fournier le jeune., ed. by James Mosley, vol. 3 (Darmstadt: Lehrdruckerei Technische Hochschule Darmstadt, 1995), 272.

446

Chapter 7

non-Arabic cultures were a recurring, formative factor in the evolution of Arabic type in the twentieth century. Monotype’s first Arabic typeface was commissioned and used by clients in India. As it was used for Urdu, an Indian language, it was modelled on foundry type that corresponded to local stylistic expectations. The resulting typeface, Monotype Series 517, was a curious hybrid, merging features of the Nastaʿlīq hand with the typical characteristics of Nasḫ type (Figure 7.4). As discussed above (see page 148 and following), the choice of Nasḫ as model for the typeface ran counter to the prevailing cultural preference for Nastaʿlīq, yet the inadequacy of typesetting technology left no choice if mechanical composition was to be achieved. The resulting typeface compromised the preferred appearance for an Urdu readership, yet provided an acceptable rendering of the language. By contrast, Mergenthaler’s synchronous attempt at devising a simplified Nastaʿlīq fount for its linecaster failed. Although the choice of the preferred style was an attempt at providing a more authentic reproduction of written Urdu, the radical compromises it imposed on the style were even less acceptable than a Nasḫ fount. Monotype Series 507 thus succeeded as the lowest common denominator of cultural identity, language suitability, and mechanical composition. The resulting appearance and associated identity were recognised beyond India, and when Monotype tried to sell the type in the Middle East it was rejected on the grounds that it did not look Arabic. Consequently, Monotype developed a new typeface, this time closely based on foundry type that was preferred by the Egyptian client. Here, the type of the Egyptian Government Press, which had also been used for the 1924 King Fuʾād Qurʾān, provided a culturally appropriate model. Indeed, the specific cultural and linguistic associations of the typeface were not exclusively Egyptian, but also recognised in India. When in 1938 the Hyderabad Government Press published a bilingual Qurʾān in Arabic and English, it used the same fount, suggesting that it was perceived as particularly suited for Arabic language texts (Figure 7.5).15 Monotype’s interpretation of the design, series 549, turned into one of the most successful typefaces for Arabic book typography, but did not achieve the same popularity for the composition of other languages. ∵

15

Undoubtedly, the religious importance and technical specificities of the text provided other reasons for the choice of typeface.

Complexities and Simplifications

Figure 7.4

Page composed with the hot-metal version of Monotype Series 507, reduced to 80% linear. From A. K. S. Lambton, Persian Vocabulary, Student’s ­Edition (Cambridge University Press, 1969).

447

448

Chapter 7

Of the languages which use the Arabic script, Persian developed one of the most distinct typographical identities. From the beginnings of typographical composition in the Persian language the importance that was assigned to a distinct appearance of typefaces became a recurring notion in the printing history of Iran. As Marzolph pointed out, the dominance of lithography as the principal printing process for Persian publications until late in the nineteenth century was probably influenced by reverence for the national writing style Nastaʿlīq.16 It defined Persian manuscript production, and was all but irreproducible in foundry type, while the standard Arabic Nasḫ hand could more easily be adapted to type. Typography, therefore, aligned more closely with Arabic than Persian aesthetics. It appears that publishers and printers in Iran deliberately sought alternative typographic forms to avoid the cultural connotations of existing founts. The precise origins of Persian typeface design remain unclear, but evidence of this study suggests that the style of imported foundry type may have influenced the early evolution of a national typographical identity.17 The specificities of this Persian-ness remained elusive for Linotype until the 1960s, when prolonged exchanges with the Ettelaʿat publishing house provided insights into the preferences of the Iranian client. Indeed, the process was so long and Linotype’s appreciation of the client’s requirements so underdeveloped that eventually Ettelaʿat resorted to commissioning an in-house artist for the design. In this role Hossein Abdollahzadeh Haghighi was perhaps the first Persian designer to develop a new, original typeface for mechanical composition. Haghighi’s resulting Ettelaʿat Simplified marked a milestone in Persian type design, as it demonstrated the stylistic treatment that was considered to reflect a decidedly Iranian typographical identity (Figure 7.6 overleaf). This design remained noticeably compromised by the simplification principles and was never adapted for subsequent technologies, yet Haghighi left another lasting mark on Persian type design. Nazanin, designed in 1976 for Linotype’s Linotron 303, has obvious similarities to his earlier type but seems to develop 16

Marzolph, “Early Printing History in Iran (1817 – ca. 1900),” 265.

17

The beginnings of the formation of a Persian typographical identity are likely to predate the introduction of mechanical typesetting, and are therefore outside the scope of the current study. Further research on this subject may contribute to clarify this specificity of Persian typographical history. When Ettelaʿat approached Linotype in the 1950s for the development of new, specifically Persian founts, models of German foundry type were presented to typify what Persian type ought to look like. According to Milo, other types commonly used in Iran can be traced to Russian origins. In either case, an ironic source of a Persian typographical identity. Milo, interview.

Complexities and Simplifications

Figure 7.5

449

Page from the 1938 Hyderabad Government Press Qurʾān; compare also to Fig�ure 2.72 on page 161, a page from the 1924 Egyptian King Fuʾād Qurʾān showing the same text, composed in the same foundry type. From Marmaduke Pickthall, The Meaning of The Glorious Qur’ân, vol. 1, 2 vols. ­(Hyderabad-Deccan: Government Central Press, 1938), 13, reduced to 75% linear.

450

Chapter 7

Figure 7.6

Ettelaʿat Simplified, composition of Haghighi’s full name from character d­ rawings, see Figure 4.13 on page 302, reduced to 50% linear.

Figure 7.7

Nazanin, proof (detail), composed on a Linotronic 300 in 18-pt, Linotype-Hell 1991, enlarged to 200% linear. DTGC, folder PostScript bromide proofs.

Figure 7.8

Mitra, proof (detail), composed on a Linotronic 300 in 18-pt, Linotype-Hell 1991, enlarged to 200% linear. DTGC, folder PostScript bromide proofs.

the aesthetic further, emancipating itself stylistically from the apparent constraints of hot-metal composition (Figure 7.7). With Nazanin’s subsequent popularity in Iran, Haghighi’s prominent influence on the identity of Persian type design was affirmed. A similar impetus for new typeface developments stood behind the design of Linotype’s Mitra. Commissioned in 1974 by the Kayhan newspaper in Iran, it was also conceived as a distinctly Persian typeface. Moreover, its brief reflected the specific needs of newspaper typography, in which economy, clarity and perceived size were crucial design factors. Stylistically, Mitra is a distinct typeface, and shares little with Nazanin; its angularity and sharp terminals even stand in stark contrast to the gentle roundness that marks Haghighi’s

Complexities and Simplifications

451

type (Figure 7.8). Nonetheless, Mitra also became one of the most popular typefaces in Iran, further contributing to the difficulty in defining a particular Persian identity in type. Judging by Nazanin and Mitra, as well as Monotype’s Mudir, another typeface of attested Persian character, a common trait is found in more synthetic and rationalised typeforms. Contrary to many typefaces that are perceived as Arabic, in these designs the traces of the writing tool are reduced (albeit least so in Mitra), with some typeforms being so removed from their manuscript origins that they could not be written at all (e.g. the mīm ‫م‬ characters of Nazanin and Mudir). Proportions of characters, such as the narrow descenders of ‫ ر‬rāʾ and ‫ و‬wāw, contribute to the specificity of these designs. Arguably, Persian typeface identity is thus defined by a distinctly typographical look, which is more removed from manuscript forms than was typically the case in Arabic designs.18 Whereas Nazanin and Mitra aspired to convey a Persian character, their designs were based on the established norms of typographical representations of the Arabic script and remained technologically conservative. By contrast, developments for Urdu, while driven by the same motivation of providing a culturally appropriate rendering, repeatedly pushed the limits of existing typographical techniques. Monotype’s and Linotype’s Nastaʿlīq typefaces of 1981 and 1987 both explored new approaches to Arabic composition, albeit in thoroughly different ways. The impetus to provide culturally sensitive designs thus resulted in the evolution of distinct aesthetics, as well technological progress, contributing lastingly to the advancement of Arabic type. By contrast, as the following section considers, not all innovations, irrespective of their merits, succeeded in contributing to Arabic type-making. Radical Innovation and its Limits Pragmatic considerations were decisive for the evolution of Arabic type in the machine age: speed of composition, economic gains, and efficiency of means all influenced design and manufacturing decisions; and by extension they imposed parameters defined by machinery and typographical principles that often originated from a Latin-centric perspective. Nonetheless, and in contrast to this dominant practice, there were also attempts to devise radically new 18

In the first years of this century this delineation can no longer be observed, as numerous contemporary Arabic typefaces are radically removed from any apparent influences of the writing tool. Current Persian designs frequently reflect, however, the influence of Nazanin.

452

Chapter 7

approaches to the problems of Arabic composition. Broadly speaking, a distinction between improvements and innovations may be useful to categorise different type developments. Whereas improvements resulted from an approach which took advantage of the evolution of technology, actual innovations emerged from the active search for new concepts which could advance the way in which Arabic type was conceived, creating new technology. One of the earliest proponents of such an innovative approach was the Dutchman Dr Edward Bernard Plooij. Plooij was an outsider to the trade and considered the issues of Arabic composition without assumptions that were based on existing practice, a position which allowed him to propose fundamentally new ideas to Arabic typemaking. Plooij analysed Arabic independently of the paradigms established in type, and with special regard to the morphology of the script. In his system pen strokes and parts of letters formed the basic elements of a fount, rather than characters which reproduced complete letterforms. Plooij argued that this approach was advantageous, for it could provide a more efficient rendering of the script in which its aesthetics no longer had to be compromised by technological limitations. But when Plooij first suggested his novel approach to Arabic type-making, the market was dominated by a small number of equipment manufacturers with rigid notions of design and manufacture, and monopolies of distribution. At this time an individual like Plooij depended on a collaboration with one of the major companies. Yet, all of Plooij’s attempts to interest and convince Monotype, Linotype, and other manufacturers failed. In the case of Linotype, although Tracy recognised the knowledge of the inventor and credited his aspirations, he considered his goals romantic and too far removed from economic concerns to succeed. For Tracy the existing methods and techniques for Arabic type-making were sufficient and their gradual improvement the only realistic way forward. Here, an innovative approach collided with a stance favouring gradual improvements, illustrating the systemic difficulties outsiders to the trade faced during this period. The financial, organisational and infrastructural requirements of type-making prohibited the experimental exploration of novel concepts, resulting in the comprehensive miscarriage of Plooij’s endeavours in Arabic type-making. Whereas Tracy may be seen as overly emphasising pragmatic considerations, sometimes with the result of stifling innovation, Plooij’s underdeveloped grasp of commercial and practical aspects of the trade undoubtedly contributed to the failure of his work. By the time that Pierre MacKay began exploring novel methods for Arabic typesetting, the dominance of the established equipment manufacturers had started to fade, and the emergence of computers for the first time opened up

Complexities and Simplifications

453

the prospect of machine-independent type. Similar to Plooij, the impetus for MacKay’s development of novel tools for Arabic typesetting was not commercial, but originated in personal motivations. But where Plooij required help from the profession, MacKay could rely on the infrastructure and equipment of the University of Washington. Demonstrating an unbiased perspective on the issues of Arabic typesetting, MacKay conceived of fundamentally new principles for the creation of type. The concept employed by the Hattat programme introduced – probably for the first time – the notion of pen strokes for the creation of typographic letterforms.19 But although the ideas and principles behind MacKay’s software packages were innovative and promised potential for the advance of Arabic type-making, their success was hampered by aesthetic shortcomings. Visually the typeset matter produced with Hattat and Katib fell short of conventional Arabic typefaces of the same period, irrespective of the revolutionary manner in which the characters were created. Radical innovation did not in itself constitute an improvement. What Robin Kinross attested in relation to Donald Knuth’s Metafont type design software, applies in equal measure to MacKay’s earlier work: The vision that Metafont offered was that all the elements of a font could be generated from the design of just one letter; and, as well, a family of variants of slope or stroke-width could also be produced. The letterforms of Metafont were generated from the strokes of an imaginary pen; but as others were soon to suggest, making of letterforms as outlines would be the way forward for making typefaces. Early published results of this project were visually crude and found to be illegitimate by the typographically initiated: as a mathematician and computer scientist, Knuth was an outsider and apparently without the ‘eye’ that follows from years of immersion in the typographic heritage. This conflict was the ancient one of the rationally correct against the aesthetically satisfying: letters that were exactly describable, as all computer-generated forms must be, did not look quite right, in comparison with those freely drawn by hand.20

19

Donald Knuth’s Metafont software, whose development began in 1977, systematically employed the same principle. MacKay’s article “Setting Arabic with a computer”, summarising the development of Hattat and Katib, was published in January 1977. It is not known whether Knuth was inspired by this work.

20

Kinross, Modern Typography, 165–66.

454

Chapter 7

For substantial and lasting advances beyond technological invention, the confluence of multiple elements was necessary, and required the combined expertise of different specialists. Whereas Plooij failed as an individual lockedout of a structurally closed and conservative industry, MacKay’s work remained within the confines inherent in typographical experimentation as pastime. Although limited by fundamentally different circumstances, in both cases the ideas and concepts initially articulated were not developed to a practicable state because of a lack of complementary competencies. By contrast, the examples of successful radical innovation emerged when specialists of different disciplines combined their expertise to develop novel typographical technologies. Of these successes, two stand out as exceptional: Linotype Sheeraz and DecoType’s Arabic Calligraphic Engine. Sheeraz was developed specifically for the Urdu newspaper market, and it was supported by favourable circumstances. At the same time that Linotype Ltd’s management recognised the commercial viability of non-Latin type, the team around Fiona Ross and Michael Fellows appreciated the cultural and artistic value of a new, original Nastaʿlīq design. Due to the scale and complexity of the task the project called for a team effort: Ross, then the Manager of Typographic Development, art-directed the new font, taking conceptual and organisational responsibility; the art work was delivered by Timothy Holloway, an experienced designer of Arabic typefaces with an explicit interest in the Nastaʿlīq style; and as the concept of the new typeface implied substantial software developments, Fellows formed an integral part of the team. The project started out with research into the main deficiencies of existing Nastaʿlīq fonts, and the design of novel concepts for Arabic type-making. Where Tracy had dismissed Plooij’s approach, the principles that were implemented in Sheeraz followed similar concepts as had been suggested by the Dutchman some thirty years earlier.21 Just as in Plooij’s and MacKay’s systems, the characters of the Sheeraz font were pen strokes and parts of letters, rather than entire letterforms. Indeed, the abolishment of the notion of a one-to-one correspondence between letterform and character provided the flexibility that was needed for an authentic rendering of Nastaʿlīq in type. Sheeraz became a successful typeface, embodying an ideal case of radical innovation. Here, various factors combined favourably, resulting in a technologically pioneering, culturally sensitive, and aesthetically accomplished design. Yet, within a few years of its development, renewed technological upheaval changed type-making again and Sheeraz was rendered obsolete. After 21

None of the team members appear to have known about Plooij’s concept and his approach of Linotype in the 1960 and 1970s.

Complexities and Simplifications

455

approximately five years of use, mainly by Pakistani newspapers and printers, a slight redesign of the typeface was made between 1993 and 1994. Typeforms were re-drawn with narrower proportions, kerning was tightened and ascending elements revised in order to achieve more compact setting for newspaper typography. Released as a PostScript font under the name Qalmi in 1995, its life-span was short. Type 1 PostScript fonts were limited to 256 characters and lacked in-built means to specify font-specific behaviour. Since Qalmi relied on specialised, dedicated software, it was not in line with the technological trend. Only one year after its release, and with the introduction of the cross-platform and Unicode compatible OpenType format in 1996, Qalmi’s fate was sealed as a remnant of superseded technology. Soon after the typeface was removed from Linotype’s library. In contrast, for DecoType legacy practice or technological shifts of typemaking had only marginal relevance. As the Dutch company started their work without prior typographical knowledge, it approached the subject without preconceived ideas and did not have to continue any established techniques. Instead, DecoType developed a research-based approach to Arabic type-making. Suspicious of conventional design methods and typeface models, DecoType pursued a two-pronged strategy in the making of its types: on one hand, it analysed the pre-typographical forms of the script, notably embodied in the Ottoman calligraphic school, and on the other hand it created dedicated technology for the unadulterated representation of the Arabic script with all its historically attested features. Whilst initially driven mainly by curiosity and without a specific application in mind, the partners of DecoType – Thomas Milo, Mirjam Somers and Peter Somers – soon turned their approach into a precise methodology bearing the hallmarks of scientific research. As relevant sources a precise corpus of material was defined, namely Ottoman manuscripts that had been penned by a calligrapher endowed with an iǧāzä, the Islamic certificate of competence (for an example see Figure 2.78 on page 170). Thus, only penmanship of the highest pedigree, tracing its lineage to the most renowned writing masters, was considered in the subsequent analysis of the script. In this systematic investigation of Arabic writing styles DecoType was guided and inspired by Mitchell’s work on Ruqʿä, as well as MacKay’s ideas for computer-aided composition, vicariously lending his work a degree of influence and applied success it had not achieved on its own. Indeed, DecoType’s work was defined by the fusion of expertise from diverse areas, notably linguistics, fine arts and (aeronautical) computer engineering, a mutually enriching environment that Plooij and MacKay may have lacked. DecoType’s collaborative team work resulted in typographical technology which not only proposed a radically new and different way of making Arabic

456

Chapter 7

type, but which also achieved a degree of usability and quality that continues to exert a lasting influence on the trade at large. This achievement is all the more remarkable when considering the power relations between an entrenched typographical industry, dominant and largely ignorant technology companies, and a micro-company with no leverage except the quality of its products. Although Linotype Ltd’s Sheeraz project and DecoType’s approach to Arabic type-making emerged from thoroughly dissimilar contexts – the former developed by the oldest typesetting machinery business as a speciality product for a discreet market segment, the latter as a wholly novel way of type-making from a handful of dedicated individuals without any business base – there are significant parallels between them. Indeed, both examples emerged at a time of upheaval induced by the profound technological changes microcomputers had brought about. Here, in two unrelated developments, the new possibilities computers provided for type-making and typesetting were employed more comprehensively than ever before. The combination of thorough, script-specific analysis with the directed use of computer technology achieved a radical departure from conventional techniques and provided the basis for the unequalled functionality of the resulting products. Yet despite these analogies – historical and conceptual – the differences between the two examples remain clear. When in the mid-1990s the Desktop Publishing Revolution was taking its toll on the balance sheets of traditional typesetting machinery makers, numerous businesses merged, scaled down or disappeared altogether. As IT companies displaced type-makers as the principal source of typographical equipment, the latter had to radically change their business models and shed all but their most essential assets. Linotype thus forfeited original design work and limited its activity to the marketing and distribution of its existing library through the German Linotype GmbH. In this process the British branch Linotype Ltd was closed in 1996, bringing an end to one of the most influential centres of nonLatin typographical developments of the century. The root cause of these cataclysmic changes in the trade’s structure also had profound reverberations on type as such, because typographical technology for PCs was effectively retrograde: with PostScript fonts the one-to-one correspondence between letterform and character was re-introduced, and the advanced capabilities of dedicated typesetting systems and tailor-made software packages became obsolete. Technologically advanced fonts like Sheeraz and Qalmi had to give way to the lowest common denominator of de facto industry standards, which were again designed from a Latin-centric perspective. PostScript fonts had much more stringent character repertoire limitations than

Complexities and Simplifications

457

type for photocomposition, and software lacked the capacity to handle many non-Latin scripts for decades to come. For DecoType, itself a product of the Desktop Publishing Revolution, such problems did not arise. Its technology was conceived independently of existing typesetting techniques – the algorithm and accompanying software could equally guide a typewriter’s bars as it could control the arrangement of digital outlines on a computer screen – and therefore avoided many of the constraints of conventional Arabic fonts. The architecture of the Arabic Calligraphic Engine provides flexibility on several levels: (1) type-making is largely liberated from concepts and metaphors that originated in Latin foundry type, providing more design options and improved efficiency than conventional type formats; (2) type composition benefits from increased precision of kerning and spacing, user-controlled diacritic and mark positioning, and variant and swash forms of letters facilitate better justification; (3) the ACE is agnostic of typeforms and formats and can easily be adapted to diverse rendering environments. Whilst this independence from conventional type formats provides the unequalled potency of the ACE, it also limited and limits the scope of applications for DecoType’s technology. This conundrum was shared by all discussed examples of radical innovation and constitutes a key issue for the evolution of Arabic type-making in the machine age. For the inventors of new techniques that diverged from established practice the perennial dilemma was to weigh an objectively better solution against its practical applicability. Ultimately typographical technology only succeeds if it gets used, no matter how correct or excellent its principles may be. Alternative approaches thus had to strike a balance between principled stance and actual application, by definition requiring compromise and flexibility. For Edward Plooij the historical circumstances prevented his success. It is difficult to imagine that any degree of compromise on his side could have persuaded the established type businesses to consider his ideas for Arabic typemaking. In the 1960s and 1970s the trade was still too entrenched to allow for any divergence from conventional techniques. By the time that Pierre MacKay commenced his work on Arabic computing, the walls around the closed sphere of type manufacturers were already crumbling, yet his work did not extend into the mainstream of the field; indeed, MacKay does not appear to have sought entry into this field and for a self-contained academic interest no compromise or adjustment was necessary. The Department of Typographic Development of Linotype Ltd, on the other hand, originated from the industry itself. The varied backgrounds of its staff from universities, art schools and computing, however, constituted a break from traditional patterns of the industry, and explain the unprecedented approach put forward in the development of Sheeraz. Its radi-

458

Chapter 7

cally new concepts emerged and remained, however, within the closed and protected environment of a dedicated typesetting system, allowing for liberties that vanished with the advent of PCs, desktop publishing and standardised font technology; with these liberties also vanished Sheeraz and Qalmi as their advanced concepts proved to be too complex and thus incompatible with the newly set norms. Against this background, DecoType’s work appears as the most sustained and successful effort in radical innovation. Due to the company’s tenacity, the exceptional quality of its products, and, perhaps counterintuitively, the apparent flexibility and willingness for compromise, DecoType could hold on to its design approach and continuously improve its products, achieving significant use and impact in the field. Its fundamental challenge of the status quo did not, as is often the case, remain on the level of critique, but it gave rise to concepts that found their way into mainstream technology and continue to influence the direction Arabic type-making takes today. Did Technological Advances Improve Arabic Type ? Irrespective of periods and technologies, a recurring feature in typographical history is that technological advances did not automatically lead to innovation in type design. From the introduction of movable type with its initial copying of manuscript forms, to the beginnings of mechanical composition and the reproduction of historical foundry type on the new machines, aesthetics were initially derived from precursors. Similarly, the development of founts for photocomposition, whether Latin or non-Latin, began by imitating the established models as the safest way to achieving acceptance by the reading audience.22 In 1985, at the beginning of the Desktop Publishing Revolution, the Swiss designers André Gürtler and Christian Mengelt assessed the degree of innovation during the photocomposition era soberly: The fact that the liberation of typography from the restraints of metal opened up undreamed of new creative possibilities, is a statement from the pioneer days of phototype which is still often and gladly quoted today. If we now consider the typography of every day with a critical eye, more than 30 years after the introduction of phototypesetting, we must

22

See Ross, The Printed Bengali Character and Its Evolution, 193.

Complexities and Simplifications

459

ask ourselves to what extent these possibilities have been used for a new creativity in typeface design as well.23 Yet this tendency could be expected in a practice-dominated field. Generally speaking, readers tend to prefer the habitual, not the unexpected, and for manufacturers the immediate concern was to make the most popular and economically successful types available for new technologies. Besides the commercial aspects that must feature prominently in a trade, existing typefaces were readily available, making their adaptation a pragmatic choice over new designs. But if for Latin script typography this tendency was criticised by designers in search of novelty, it remained an issue of little gravity. Indeed, the problem was minor, as one need look no further than Adrian Frutiger’s work during the photocomposition era to see genuine innovation. For non-Latin scripts, however, the situation was different. Because of the small number of typefaces, the range of choices for adaptations to new technologies was narrow. Moreover, non-Latin scripts were a marginal aspect of type businesses throughout most of the twentieth century, vividly illustrated by the fact that all but Latin fell within one category. Considering the wide range of scripts within this broad definition, each only receiving a fraction of resources that were dedicated to all non-Latin developments, the dimension of the advantage of Latin type becomes tangible. For Arabic, the tendency to copy existing type often entailed the repetition of design compromises that had originally been imposed by limitations of machinery: reduced character sets were copied, distorted proportions were repeated unnecessarily, design choices for specific circumstances were followed without assessment of their origins. In short, design practice left a lot to be desired. A case in point is the history of Simplified Arabic. From the first adaptation of Arabic to the Linotype, substantial concessions had to be made to accommodate the large repertoire of letterforms within the limited capacity of the machine. Yet further streamlining was sought for the sake of faster composition speeds – a notion closely entwined with 1950s typesetting technology – leading to the design of Simplified Arabic. Due to its commercial success, Simplified Arabic became the first Arabic fount that was adapted for every new technology, despite its roots in linecaster composition. And even though computers in the 1980s had little in common with linecasters, Simplified Arabic, known as Yakout from 1967, was still among the first 23

André Gürtler and Christian Mengelt, “Fundamental Research Methods and Form Innovations in Type Design Compared to Technological Developments in Type Production,” Visible Language XIX, no. 1 (Winter 1985): 125.

460

Chapter 7

typefaces that were converted to the PostScript format. Moreover, its ubiquity led to its copying by competitors, furthering the disproportionate influence of this historical design. In the 1970s and 1980s the US Compugraphic Corporation became notorious for its flagrant and overt plagiarism of typefaces by the established type-makers.24 As the company aspired to enter the Middle Eastern market, Arabic typefaces had to be offered with its machines. Cynthia Batty, Compugraphic’s former manager of the Department of Type Design and Development, recalled the situation pragmatically: It was all about selling typesetters – the type was the way to sell them, and in the non-Latins you had to have the standard metal faces or nobody would buy. Simple commercial situation.25 Therefore the company sought such ‘standard’ designs and implemented them on its technologically advanced, low-price phototypesetters. As discussed above (see page 335 and following), Compugraphic based its Arabic founts on the most successful typefaces of Linotype and Monotype respectively: whereas Linotype’s Yakout or Simplified Arabic had become established as the norm for newspaper and magazine typography, Monotype’s Series 589 had set the standard for quality book work; Compugraphic imitated both. By 1989 Linotype had obtained samples of Compugraphic’s range of four Arabic typefaces in eight fonts (Figure 7.9). The designs appear to be copies of Monotype’s Series 549 (or 589), and Linotype’s Yakout, Mariam and Kufi typefaces. Their fidelity to the apparent models varies, and their commercial success and the extent of their use with Compugraphic equipment are not known. However, the two text typefaces attained an unlikely prominence as a result of another technological shift. From 1982 the Belgian Agfa Gevaert Corporation held a majority share in the Compugraphic Corporation, and in 1988 Agfa acquired the entire company. Having thus secured access to its type library, Agfa was able to use and license fonts to others. In consequence, Compugraphic’s Arabic fonts were licensed to Microsoft for use in its Windows operating system. The typefaces retained their original names Simplified Arabic and Traditional Arabic, a legacy from design 24

Cynthia Batty, former manager of the Department of Type Design and Development at Compugraphic, defends this policy. According to Batty, the company approached the Linotype group and Monotype to license designs, but was not given permission. In Batty’s view this justified the plagiarising of the competitor’s libraries. Email message to the author, “Re: Fontographer,” 11 June 2013. The author is not convinced by this rationale.

25

Ibid.

Complexities and Simplifications

Figure 7.9

Sample of Compugraphic’s first Arabic typefaces, 1989, actual size. From top to bottom the shown fonts resemble these competitor’s designs: Yakout (regular and bold), Mariam (bold and outline), Monotype Series 549 (regular and bold), and Kufi (regular and outline). DTGC, folder Arabic PostScript ­developments.

461

462

Chapter 7

classifications that originated in long-obsolete technological constraints. Indeed, not only the names were carried-over from hot-metal machinery, but also design principles were copied without revisions: further to the simplified character set, the baseline of these fonts was set higher than the standard Latin baseline in order to fit within the vertical space of the Compugraphic EditWriter 7800’s exposure window (see also page 335 and following).26 The licensed fonts appear to have undergone little, if any, revision, and were first published with the Arabic version of Windows 3.1. From Windows 2000 they were bundled with all subsequent editions of the Os.27 Moreover, when the system fonts Arial and Times New Roman were extended to support the Arabic script, the required glyph complements were also taken from Compugraphic’s Simplified Arabic fonts, in what can only be described as flagrant design laxity.28 In effect, three nominally different typefaces on the Windows platform thus look identical for Arabic text (Figure 7.10). Further�more, the True­Type core fonts for the Web, Microsoft’s free font bundle for the World Wide Web, included Arial and Times New Roman, making these Arabic fonts available across most computers – in terms of role-model function and the establishment of typographical norms an unfortunate situation. Due to their proliferation and the resulting concerns for legacy document compatibility, and notwithstanding their apparent shortcomings, Simplified Arabic and Traditional Arabic remained the standard Arabic fonts through five Windows generations over the next 12 years and are still shipped with the most recent version of the OS.29 However, the copying of existing typeface designs and technical configurations that originated from specific machinery was not unique to the Microsoft Corporation, but a widespread phenomenon in the 1990s. When Glyph Systems 26

Boutros et. al., Talking About Arabic, 42–43. Boutros conveys that this was due to technical constraints which could only be changed with the OpenType format. This seems improbable as nothing in the TrueType format would have required such a baseline shift; inconsiderate copying of existing type including the copying of compromises from legacy technology is a more plausible explanation for such practice.

27

Boutros et. al., Talking About Arabic, 44, and “Simplified Arabic,” 2011, http://www.microsoft.com/typography/fonts/font.aspx?FMID=1879 (accessed 16 May 2017).

28

The glyphs are not completely identical but the differences are too slight to constitute a distinct design. Note that Western technology companies could only afford such laxity with regard to non-Latin scripts, as is readily apparent by the care that was taken with Latin type from the beginnings of desktop publishing.

29

Only in Windows 8, released in 2012, a newly conceived Arabic version of the system typeface Segoe UI superseded the old Compugraphic fonts as the default.

Complexities and Simplifications

Figure 7.10

463

Sample setting of Microsoft fonts Arial, Times New Roman, Simplified Arabic and Traditional Arabic (top to bottom). Note that Simplified Arabic features slight improvements over Arial and Times New Roman, notably because it has kerning definitions for the Arabic glyphs. Composition by the author.

c­ ommenced the development of Arabic software, it licensed Compugraphic’s type library and bundled it with the Arabic enabled version of Ventura Publisher. It was released in 1990 as the first Arabic desktop publishing application for Windows, and further perpetuated the use of these technically obsolete fonts.30 Through these successive poor choices and questionable practices, Simplified Arabic and Traditional Arabic, Compugraphic’s ill-informed clones of Monotype Series 549 and Linotype Yakout, probably became the most widely used and read Arabic typefaces in history. Clearly, in this prominent case advances in technology brought little progress in Arabic type design and the resulting typography. The history of Simplified Arabic is a particularly suggestive example of how innovation in technology and innovation in type-making were repeatedly out of step with each other. Indeed, throughout the machine age, novel methods also entailed regression for Arabic type design and typography. The Linotype brought economic and efficiency improvements, yet the quality of Arabic composition greatly suffered in comparison to its foundry type precursor. Monotype’s Series 549 set a quality standard in hot-metal composition that degraded in later adaptations to photocomposition. By the time the typeface was digitised and adapted to this latest technology, the repeated conversions had further harmed the quality of the original design (Figure 7.11).

30

Boutros et. al., Talking About Arabic 44. The inadequate quality of Arabic type bundled with Ventura Publisher was a central motivating factor for Thomas Milo to develop his own typographical technology.

464

Chapter 7

In contrast to Monotype’s often haphazard Arabic developments during the photocomposition era, Linotype-Paul’s programme demonstrated commitment to take advantage of the gradual improvements technology offered for new Arabic typefaces. Not only were existing designs updated, implementing improved fitting, kerning, vocalisation and larger character sets, but original designs like Osman, Mitra, Jalal, Nazanin, and Sheeraz were also produced. Arguably, Linotype’s work in the late 1970s and 1980s was amongst the most productive and innovative efforts in the history of Arabic type-making. From the 1960s, the technologically-induced erosion of the monopolies of traditional manufacturers opened up the field and enabled the emergence of competitors and new ideas. Of the numerous and increasingly frequent changes of technology, the ascent of digital composition had the most radical effect on the field since the beginning of mechanical composition. It brought the complete transformation of typesetting and type-making, and turned a trade that relied on industrial manufacturing processes into a profession which required minimal capital outlay. From the 1990s anyone with a personal computer and relatively inexpensive software could create type that was (in principle) on par with industry standards. The drawing offices and type manufacturing departments were no longer required when typefaces could be created digitally, resulting in the rapid demise of the former equipment ­manufacturers. Whereas this development has often been lauded for its democratic qualities, there has been criticism too. Kinross rightly pointed out that “if ‘democracy’ implies a spreading of power to the people, this is the wrong description for what is going on here: it is more a simple spreading of typography among the masses”.31 Moreover, the end of traditional type-makers also destroyed accumulated knowledge, quality standards, and savoir faire. The tabula rasa the Desktop Publishing Revolution brought to the world of type-making was as destructive as it was liberating. This was particularly felt in Arabic type-making, a domain too young to absorb this kind of competence-destroying technological change without damage. As has been noted above, during the 1980s and 1990s various new IT companies were busy developing basic right-to-left capabilities for software, generally with poor or no kerning, and no provisions for mark positioning, a technological state that was inferior to Monotype hot-metal composition from the middle of the century (see “Desktop Publishing Tools for Arabic” on page 406). And although the newly accessible means resulted in an increase of typefaces, their quality rarely matched those developed by typesetting manufactur31

Kinross, Modern Typography, 182.

Complexities and Simplifications

Figure 7.11

465

Monotype Naskh, specimen created with a 600 dpi Monotype Papermaster. From Monotype Typography Ltd, Library of Non-Latin Typefaces, n.d., ­reduced to 70% linear.

466

Chapter 7

ers for their proprietary equipment.32 Here, the spread of the tools for ­making type also entailed a bastardisation of the products, as access to computers could not substitute for experience, knowledge, and skill. Plagiarism and piracy – not new but much facilitated by digital fonts – played an important part in this negative trend, but were not solely responsible. For Arabic type-making the rise of the amateur type designer was more problematic than in the world of Latin typography. The field’s short (and little known) history, the narrow range of existing models, and the lack of design education and established quality standards opened the door to poor designs by self-professed experts, and customers lacking expertise to make informed choices. Indeed, the breakdown of the relations between customers and manufacturers was identified by Ross as a factor that contributed to the deterioration of quality standards.33 Whereas company representatives had kept remarkably close ties with clients, the demise of these professional networks entailed an anonymisation of type. The direct link between type-maker and user was severed when digital font files simply became elements of software packages, and feedback and revision in accord with customer needs became initially more difficult.34 Radical innovation generally happened along the margins of the trade, rather than in its established centre. Although the influence of such contributions on the bulk of applied typography was modest, the most far-reaching modernisations happened there. Indeed, most Arabic typefaces on the market today could technically be implemented on a Monotype hot-metal system, and in some cases would even work on a linecaster without alteration – a situation that speaks volumes about the actual advances technological ‘progress’ has brought for Arabic type-making. The accomplishments of type design as practised since the beginning of the digital era remained overwhelmingly on the level of stylistic, formal variation, and, apart from a handful of exceptions, did not venture into structural innovation. It is for this reason that Linotype Ltd’s Sheeraz and DecoType’s ACE are such towering achievements that advanced Arabic type through technological innovation. Neither of them could have been implemented with any prior technology, and whilst Sheeraz is now con32

In 1998 Ross wrote: “Now font tools allow a single individual with a personal computer the facility to design a font and compose with it without recourse to professional typefounders. In the field of non-Latin types the results have not been laudable, and this method is still eschewed by the major font manufacturers.” Fiona Ross, “Translating Non-Latin Scripts into Type,” Typography Papers no. 3 (1998): 75.

33

Ross, interview.

34

With the advent of the World Wide Web this difficulty was of course eliminated. Today communications between designers and users can be more direct than ever before.

Complexities and Simplifications

467

fined to the pages of history, ACE continues to provide unprecedented design potential for the makers, as well as the users of Arabic type.35 Its ­revolutionary qualities are, however, not readily apparent to the uninitiated. Because the essence of the ACE lies in its structure, rather than the immediately perceivable stylistic appearance – which incidentally is conventional in the best sense of the word – the novelty of DecoType’s work has remained opaque to casual observers and users. This comes as no surprise because the recognition of substance in novelty, a key competence for typographical designers, necessitates deep and prolonged immersion in the writing and printing culture of a script; and although lip service to research has become common currency in design discourse, there is little agreement as to what constitutes sound research in this emerging field. Whilst the practice of Arabic type-making and typography constitutes the lifeblood of this field, the dual pressures of economic concerns and time shortages often compromise the depth of engagement with the subject, resulting in ill-informed design decisions. A contributing factor in this dynamic stems from improvements of the tools available for digital type-making today. Recent years saw the development of more dedicated software than ever before, providing more choice, better tools and more easily accessible design options. The facility with which fonts for most scripts can be conceived now is remarkable, and constitutes an essential requirement for the advancement of non-Latin typography. Epilogue: Arabic Type-Making Today For Arabic type-making OpenType, today’s most prevalent font format, has significantly alleviated many limitations of earlier technologies.36 Several constraints which recurred throughout the various type-making and composition processes employed in the twentieth century do not apply anymore: no longer are character set reductions necessary as the number of glyphs a font may 35

Recently John Hudson drew inspiration from the concepts articulated for Sheeraz in the design of the typeface Aldhabi, commissioned by Microsoft and released with its Windows 8 OS in 2012. “John Hudson ISType 2012, June 15” (Istanbul, 17 October 2012), ­ http://vimeo.com/51580484 (accessed 16 May 2017).

36

Another so-called smart font technology is Graphite, developed by SIL International. “About Graphite,” SIL International, http://scripts.sil.org/cms/scripts/page.php?site_ id=projects&item_id=graphite_about (accessed 16 May 2017). SIL International is a US missionary institution that plays a significant role in the development and provision of typographical tools and typefaces for so-called minority languages and non-Latin scripts.

468

Chapter 7

contain is virtually unlimited, and at the discretion of the designer; composition speed no longer applies as a factor, as the direct correspondence between keying and output has long been dispensed with; vocalisation, a conundrum for Arabic type-making for much of its history, can now be implemented with comparable ease; and the kerning and spacing of fonts are only subject to the skill and tenacity of the designer. OpenType programming can provide typographical refinements, language-specific font behaviour and letterforms, and the mixing of various scripts and typefaces in one line can now be reasonably expected from page-layout software. Yet the belief that technological limitations can be considered a relic of the past is wholly unjustified. For despite the improvements and facilities that current tools provide the makers and users of type, the suggestion that Arabic type design can now flourish as an unfettered expression of creative powers is quite ill-informed. With the demise of proprietary typographical technology and the general use of standardised formats in computing, control over the final output also ceased. Broadly speaking, the separation between typemaking and the making of text processing software introduced complications that were previously unknown. For although standards are defined and compliance with them is generally encouraged and attempted, significant parts of the standards remain discretionary and depend on implementation. Thus, and despite standards compliance, the same font may be rendered differently across changing software environments – a problem that has been exacerbated through the proliferation of portable devices and the accompanying multiplication of operating systems, browsers and other type-rendering applications. Here, the OpenType specification has fundamental shortcomings, for although specific functionality may be defined as so-called ‘features’, fonts rely on text processing software for their interpretation. However, because the inclusion of feature processing in software is discretionary, support for this added functionality is inconsistent and may not be provided at all in some applications.37 For example, until recently Apple’s operating system did not recognise Arabic kerning, displaying fully functional fonts much inferior than was the case on Microsoft Windows; and Google’s Chrome browser even rejected all fonts that included OpenType functionality that is necessary for complex script render37

At the time of writing the JSTF table which is conceived to provide advanced justification control, a feature that may have particular usefulness for Arabic type-making, is part of the OpenType specification but it is not implemented in any text processing client, nor does any existing type-making software provide means for its definition. “JSTF - The Justification Table,” 4 January 2017, http://www.microsoft.com/typography/otspec/ jstf.htm (accessed 16 May 2017).

Complexities and Simplifications

469

ing, in effect blocking many non-Latin fonts from being displayed.38 As can be gathered from these examples, the constant evolution of software and ­computer capabilities preclude a definitive assessment of OpenType support; yet this structural deficiency of the format remains a problematic fact. Besides the issues arising from diverse standards implementations in software, type-making tools continue to shape how Arabic fonts are designed. There have been considerable improvements to the capabilities and the interfaces of tools, notably in the application Glyphs developed by Georg Seifert. Basic Arabic font functionality such as the selection of initial, medial and final shapes can be automated, the application allows for a right-to-left preview of text and, importantly, has a better interface for the definition of mark positions and diagonal joining behaviour than any other currently available software. The developers of Glyphs actively seek to improve its non-Latin capabilities and, in part through the generous help of its growing user community, have achieved good progress. Furthermore, the application provides several options that are designed to simplify and accelerate the design process, notably through the reuse of characters and parts thereof. As the market for non-Latin and specifically Arabic type has been steadily growing over the last decade, increasingly designers without prior experience, knowledge, and training develop script extensions to their original Latin designs. Encouraged by current type design software and its publicity, promising effortlessness in the making of complex script fonts, a mind-set appears to emerge which presumes that technological facility may compensate for a lack of script specific expertise. Because the technology to develop Arabic fonts without a steep learning curve is readily available, designers who are not willing or able to invest the years of learning and research necessary to master a foreign script are led to believe that their tools and some superficial ‘borrowing’ of design elements are sufficient for successful design. An assumption that could not be further from the reality of sound design practice. Whilst efficiency-improving automation and tools such as the software-assisted use of modules for the creation of typeforms have their undeniable utility, they also impose their logic on designers whose lack of script specific expertise makes them uncritical in their decisions. For example the distinction of typeforms that share similar skeletons, critical for legibility, often falls victim to excessive modularity in contemporary designs; more generally, there is an observable tendency towards static, rigidly aligned and geometrically constructed typeforms, betraying the formative power of current tools on the imagination and creativity of designers. 38

This problem has been addressed in a subsequent update of the browser.

470

Chapter 7

Another aspect of advances in type-making tools, namely the automation of select basic functionality and the in-built definition of font synopses, has the potential to be perceived as actual standards by undiscerning users. In view of the historical evidence this may be problematic, for indiscriminate copying of various kinds of legacy practice and character sets has been a recurring and demonstrably negative feature of Arabic type-making. Whereas the manual definition of these elements of a typeface necessitates considerable research and immersion with language and script, which in turn encourages the review and re-assessment of existing practice, the provision of ready-made solutions curtails innovation and risks the proliferation of errors and poor practice; especially so in a field that has not yet developed a rich body of (reliable) literature and a high level of typographical culture. Yet despite these, potentially ambiguous features, none of the current, commercially available tools provide comprehensive means for Arabic type-making. Defining characteristics of the Arabic script such as the justification of connected letter groups, diagonal connections, the changes of position and shape of marks in varying contexts, and the generally higher level of contextual dependence of letter configurations, are poorly provided for. While some of these characteristics can be partially achieved through laborious and often inefficient use of OpenType functionality, structural deficiencies of industry standards and design tools have so far precluded unfettered Arabic typemaking.39 Glyphs has achieved some noticeable improvements over earlier type-making software, yet the visual definition and control of contextual variation, and the interaction of the various semantic layers of the Arabic script – rasm, diacritics and marks – remain aspects that require further attention and ­amelioration. ∵ Apart from the tangible advantages and drawbacks of current digital technology, conceptual design questions also arise from these novel means. As 39

In a recent example John Hudson, probably one of the most technologically competent developers of OpenType fonts, failed to successfully implement fundamental shaping principles of common Urdu letters. The typeface Aldhabi, commissioned by Microsoft as a showcase of mainstream font technology, thus had to exclude support for approximately 65 million Urdu native speakers. See “John Hudson ISType 2012, June 15” (Istanbul, 17 October 2012).

Complexities and Simplifications

471

fonts may now contain many thousands of glyphs, designers have the unprecedented possibility to develop typefaces for multiple scripts under the umbrella of one, unified visual language. One of the first typeface projects to actively explore this novel approach was Lucida Sans Unicode, designed by the US ­designers Charles Bigelow and Kris Holmes in the late 1980s and first published with Windows 3.1 in 1993. In an article published the same year, Bigelow and Holmes discussed their approach and the challenges of the design task, and articulated – potentially for the first time – the notion of harmonised typeface design: By ‘harmonization’, we mean that the basic weights and alignments of disparate alphabets are regularized and tuned to work together, so that their inessential differences are minimized, but their essential, meaningful differences preserved. In this way the ‘noise’ of insignificant design artefacts and historical accidents is filtered out, leaving the ‘signal’ of significant character features amplified in comparison. Within a harmonized font, when text changes from Latin to Cyrillic, or from Greek to Hebrew, or when mathematical expressions or other symbols are introduced into text, the visual size, weight, and rhythm of the characters should not appear to change, should not jar or distract the reader, but the basic character shapes should nevertheless be distinctive and immediately recognizable.40 Bigelow and Holmes also expressed caution, recognising that “one could argue that regularization of design elements is homogenization”. To avoid this potential problem the designers learned principles of the various scripts that were included in Lucida, and researched their visual cultures. Deliberate steps were taken to avoid imposing “Latinate notions of ‘correct’ letter design”, in a laudable approach to culturally sensitive typeface design.41 Yet the rationale for the endeavour remained vague, summarised in “harmonization seems like a desirable goal”, and specific use scenarios which could benefit from such a design approach were only mentioned in passing. Indeed, the question remains, in what application is it desirable to minimise the differences between scripts? Certainly, the majority of continuous texts are predominantly in a single script, thus not even raising the prospect of visual noise. But even if a text incorporates multiple scripts, it would rarely benefit from a comprehensive 40

Charles Bigelow and Kris Holmes, “The Design of a Unicode Font,” in Electronic publishing, vol. 6 (University of Nottingham, 1993), 292.

41

Ibid., 299.

472

Chapter 7

suppression of distinctive features. For example foreign language quotations may call for typographical contrast, rather than uniformity, and dictionaries – the prime example of multi-script typography – depend on the distinctive treatment of different text categories. Generally speaking, it appears that harmonised design is more applicable in uses that are less text-oriented, as the priorities of branding, advertisement and identity design may favour more uniformity. It comes as no surprise, then, that over the last decade it is in these areas that Arabic type design is most sought after. Under the dual impetus of increasing international trade and communications, and the continued ascent of the Gulf economies, multi­national corporations and organisations are investing into unified visual appearances. Against the background of established Latin script corporate identities and the pre-eminent influence of Western design language, the adaptations for worldwide use are predictably one-sided, generally adopting aesthetics and codes that retain little regional cultural distinctiveness.42 On the other hand, the global interconnections of the Internet age provide unprecedented opportunities for designers from the Arabic script world, and have commenced a diversification of approaches and styles that is bound to contribute to the advance of Arabic typography. No longer do designers come principally from the West and a small strata of Lebanese society, but contributions from Pakistan to Iran, and from Palestine to Morocco start to propose divergent concepts and aesthetics. Arabic type-making today is as emancipated from ethnicity or nationality, as it is from economic power: this is demonstrated by Hasan Abu Afash who has been able to carve out a practice from Gaza, against all odds and electricity cuts. Yet, as his example shows, large parts of this Arabic script world are in various states of disarray, economic misery, oppression and ravaged by war and strife. Whereas the important cultural centres are mired in conflict, gagged by regimes and stifled economically, the current hubs of wealth and power of the Middle East have little cultural capital to offer. Therefore a considerable responsibility remains on the side of Western and international designers, agencies, and clients to approach the subjects of Arabic type-making and typography in the most informed and considerate way possible. Research and subsequent publications therefore have a central role 42

A comprehensive discussion of harmonised typeface design goes beyond the scope of this study; suffice it to note that it appears to be a distinctly modernist notion, implicitly advancing a decidedly Western perspective. For a longer discussion see Titus Nemeth, “‘Harmonised type design’ revisited,” in Mary C. Dyson & Ching Y. Suen eds. Digital Fonts and Reading, Series on Language Processing, Pattern Recognition and Intelligent Systems 1, 150–172 (Singapore: World Scientific, 2015).

Complexities and Simplifications

473

in providing the basis for a sound understanding of design issues and possible approaches. Here, it is heartening that, whilst far from mature, Arabic typographical design discourse is widening, and the value of substantive research is increasingly acknowledged. Indeed, amongst the efforts to establish typographical design as an academic discipline, research projects related to various aspects of non-Latin typography form a sizeable part of this endeavour; a trend that is likely to continue, not least because the refugee child of today will be the design student of tomorrow – may the present book be of some use for them. Although Arabic type-making remains limited by assumptions which derive from the properties of the Latin script, there is the real prospect that the information age’s promise of a post-mechanical era will hold true, and conclusively dispel the ghosts of distant technologies that still haunt Arabic typo­graphy today.

Annex 1

‘Arabic Composed on “Monotype” Machines’

Typescript of a draft text for a promotional publication created at the occasion of the launch of Monotype Series 549 in 1948. The Monotype Corporation Ltd, ­“Arabic Composed on ‘Monotype’ Machines,” n.d. MT, folder Arabic (early correspondence).

476

Annex 1

Annex 1

477

Annex 2

“Untitled Typescript”

Probably an earlier draft for the text shown in Annex 1. The Monotype Corporation Ltd, “Untitled Typescript,” n.d. MT, folder Arabic (early corres­ pondence).

Annex 2

479

480

Annex 2

Annex 2

481

Bibliography Published Sources Abou Rjeily, Rana. Cultural Connectives. New York: Mark Batty Publisher, 2011. Adler, Michael. Writing Machine: History of the Typewriter. London: George Allen & Unwin Ltd., 1973. Aman, Mohammed M. “Use of Arabic in Computerized Information Interchange.” Journal of the American Society for Information Science 35, no. 4 (1984): 204–10. Ansari, Hamied. Egypt: The Stalled Society. Albany: State University of New York Press, 1986. Askari, Hossein, and Shohreh Majin. “Recent Economic Growth in Iran.” Middle Eastern ­Studies 12, no. 3 (1 October 1976): 105–23. Atanasiu, Vlad. De La Fréquence Des Lettres et de Son Influence En Calligraphie Arabe. Paris: L’Harmattan, 1999. Atiyeh, George N. “The Book in the Modern Arab World: The Cases of Lebanon and Egypt.” In The Book in the Islamic world, edited by George N. Atiyeh, 233–53. Albany: State University of New York Press, 1995. Avery, Peter. “Printing, the Press and Literature in Modern Iran.” In The Cambridge History of Iran. Vol. 7. Cambridge University Press, 1991. Ayalon, Ami. The Press in the Arab Middle East: a History. New York and Oxford: ­Oxford University Press, 1995. ———. “Private Publishing in the Nahḍa,” in International Journal of Middle East Studies, 40 (2008), 561–577. ———. “Arab Booksellers and Bookshops in the Age of Printing, 1850–1914,” British Journal of Middle Eastern Studies, 37, no. 1, (April 2010), 73–93. Baḥarī, Hišām Tawfīq. Ṣaḥāfä al-Ġadu. Cairo: Dār al-Maʿārif Bimiṣr, 1968. Banham, Rob, and Fiona Ross. Non-Latin Typefaces: At St Bride Library, London and Department of Typography & Graphic Communication, University of Reading. London: St Bride Library, 2008. Bartimo, Jim, and Michael McCarthy. “Is Apple’s Laserwriter On Target? Delays in Apple’s Mac Office Put Burden of Success on Laser Printer.” InfoWorld, 7, no. 6, 11 February 1985. Beeching, Wilfred A. Century of the Typewriter. London: William Heinemann Ltd., 1974. Bennett, Pam. Practical Desktop Publishing. McGraw-Hill, 1991. Bigelow, Charles, and John W. Seybold. “Technology and the Aesthetics of Type. Maintaining the ‘Tradition’ in the Age of Electronics.” The Seybold Report on Publishing Systems 10, no. 24 (24 August 1981): 3–16.

484

Bibliography

Bigelow, Charles, and Kris Holmes. “The Design of a Unicode Font.” In Electronic publishing, 6:289–305. University of Nottingham, 1993. Blair, Sheila. Islamic Calligraphy. Edinburgh: Edinburgh University Press, 2008. Boag, Andrew. “Typographic Measurement: A chronology.” Typography Papers no.1 (1996): 105–22. ———. “Monotype and Phototypesetting.” Journal of the Printing Historical Society no. 2. New Series (Winter 2000): 57–77. Boag, Andrew, and Christopher Burke eds. History of the Monotype Corporation. Woodstock: Vanbrugh Press, 2014. Bobzin, Hartmut. “From Venice to Cairo: On the History of Arabic Editions of the Koran.” In Middle Eastern languages and the print revolution: a cross-cultural encounter: a catalogue and companion to the exhibition, 151–76. Westhofen: WVAVerlag Skulima, 2002. Bringhurst, Robert. The Elements of Typographic Style. Point Roberts & Vancouver: Hartley & Marks Publishers, 2004. Boileau, C. L. “The Intertype ‘Fotosetter’ Machine.” In The Penrose Graphic Arts International Annual, 47:91–94. London: Lund Humphries Publishers Ltd., 1953. Boutros, Mourad. Arabic for Designers. New York: Mark Batty Publisher, 2007. Boutros, Mourad et al. Talking About Arabic. New York: Mark Batty Publisher, 2009. Bullen, Henry Lewis. “Linn Boyd Benton – The Man and His Work.” The Inland Printer 70, no. 1 (October 1922). Burke, Christopher. “The Early Years 1900–1922.” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 4–13. ­———. Active Literature: Jan Tschichold and New Typography. London: Hyphen Press, 2007. Cantú, Stefania, and Paolo Daniele Corda. La scrittura araba e il progetto Decotype: dai manoscritti alla calligrafia informatica. Mergozzo: Sedizioni, 2013. Carter, Harry. “Fournier on Typefounding.” In The Manuel Typographique of Pierre-­ Simon Fournier le jeune., edited by James Mosley, 3:51–324. Darmstadt: Lehrdruckerei Technische Hochschule Darmstadt, 1995. Carter, Matthew. “Olympian, A New Type for Newspapers.” Interpresgrafik no. 7 (1971): 42–47. Carter, Sebastian. “The Morison Years and Beyond: 1923–1965.” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 14–25. Chejne, Anwar G. Arabic Language: Its Role in History. University of Minnesota Press, 1969. Cioeta, Donald J. “Ottoman Censorship in Lebanon and Syria, 1876-1908.” International Journal of Middle East Studies, 10, no. 2. (May, 1979): 167–186. Coakley, J. F. The Typography of Syriac. A Historical Catalogue of Printing Types,

Bibliography

485

1­ 537–1958. New Castle, DE & London: Oak Knoll Press & The British Library, 2006. Cost, Patricia A. The Bentons: How an American Father and Son Changed the Printing Industry. Rochester: Cary Graphic Arts Press, 2011. Coulmas, Florian. The Writing Systems of the World. Oxford & New York: Basil Blackwell, 1989. Daines, Michael. “Some Aspects of the Effects of Technology on Type Design.” In Computers and Typography, 76–84. Oxford: Intellect Books, 1993. Datla, Kavita. “A Worldly Vernacular: Urdu at Osmania University.” Modern Asian Studies 43, no. 05 (2009): 1117–48. DecoType. “Typesetting Islamic Calligraphy.” Amsterdam: DecoType, 1989. ———. “DTP Naskh: The DecoType Professional font series for Microsoft Windows.” Amsterdam: DecoType, 1995. Derman, Uğur. Siegel des Sultans. Osmanische Kalligrafie aus dem Sakıp Sabancı ­Museum, Sabancı Universität Istanbul. Berlin: Deutsche Guggenheim, 2001. Diocles. On Burning Mirrors / Ed. by G. J. Toomer. Sources in the History of Mathematics and the Physical Sciences 1. New York: Springer Verlag, 1976. Dreyfus, John. “Foreword.” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 3. Dvorak, John C. “Kiss Your ASCII Goodbye.” PC Magazine 11, no. 15 (15 September 1992): 93. Dwiggins, William Addison. WAD to RR, A Letter About Designing Type. Cambridge, MA: Harvard College Library, Department of Printing and Graphic Arts, 1940. Esterson, Simon, and John L. Walters. “Robin Nicholas.” Eye Magazine 21, no. 84 ­(Autumn 2012): 68–78. Fahim, K, J. Habibion & F. Vittor. English-Persian Dictionary of Legal and Commercial Terms. Leiden: Brill, 1989. Fedirka, Sarah. “Reorienting Modernism: Transnational Exchange in the Modernist Little Magazine Orient.” English Language Notes 49, no. 1 (Spring/Summer 2011): 77–90. Felici, James, and Ted Nace. Desktop Publishing Skills : a Primer for Typesetting with Computers and Laser Printers / James Felici, Ted Nace. Reading, MA: AddisonWesley, 1987. Gabeyan, Hrant. “Modern Developments in Arabic Typesetting.” In Middle Eastern languages and the print revolution: a cross-cultural encounter: a catalogue and companion to the exhibition, 216–21. Westhofen: WVA-Verlag Skulima, 2002. Gacek, Adam: The Arabic Manuscript Tradition. A Glossary of Technical Terms & ­Bibliography. ­Leiden: Brill, 2001. Garth, William W. IV. Entrepreneur. A Biography of William W. Garth, Jr., And the Early History of Photocomposition. Self-published, 2002. Geerhold, William. “The Rear Guard Press.” Saudi Aramco World, February 1967.

486

Bibliography

Gilmour, David. Lebanon: The Fractured Country. 2nd rev. ed. London: Sphere Books, 1987. First published 1983 by Martin Robertson, Oxford. Glass, Dagmar. “Arabic Newspapers and Periodicals in the Arab World (1828–1928).” In Middle Eastern languages and the print revolution: a cross-cultural encounter: a catalogue and companion to the exhibition, 207–15. Westhofen: WVA-Verlag Skulima, 2002. Glass, Dagmar, and Geoffrey Roper. “The Printing of Arabic Books in the Arab World.” In Middle Eastern languages and the print revolution: a cross-cultural encounter: a catalogue and companion to the exhibition, 177–206. Westhofen: WVA-Verlag Skulima, 2002. Gonczarowski, Jakob. “Industry Standard Outline Font Formats.” In Visual and ­Technical Aspects of Type, 110–25. Cambridge University Press, 1993. Griffith, Chauncey. “The Linotype Development of Type Faces.” In The Annual of Book Making 1927 · 1937, no page numbers. New York: Colophon, 1938. Gürtler, André, and Christian Mengelt. “Fundamental Researh Methods and Form Innovations in Type Design Compared to Technological Developments in Type Production.” Visible Language XIX, no. 1 (Winter 1985): 123–48. Hamm, Roberto. Pour une typographie arabe: contribution technique à la démocratisation de la culture arabe. Paris: Éditions Sindbad, 1975. ———. “De la calligraphie à la typographie,” Communication et langages no. 46 (2ème trimestre 1980): 34–49. Hammond Type-Writer Company. Advertisement, The American Journal of Semitic Languages and Literatures, 16, no. 1 (October 1899). Hanaway, William L., and Brian Spooner. Reading Nastaʻliq: Persian and Urdu Hands from 1500 to the Present. Costa Mesa, CA: Mazda Publishers, 2007. Hanebütt-Benz, Eva, Dagmar Glass, and Geoffrey Roper eds. Middle Eastern Languages and the Print Revolution: a Cross-Cultural Encounter: a Catalogue and Companion to the Exhibition. Westhofen: WVA-Verlag Skulima, 2002. Harris, Nigel. The End of the Third World: Newly Industrializing Countries and the Decline of an Ideology. London: Penguin Books Ltd., 1990. Hammam, M.Y. “History of Printing in Egypt”, Gutenberg-Jahrbuch 1951, 156-159. Heath, Leslie G., and Ian Faux. A Manual for Phototypesetting. London & Manchester: Lithographic Training Services Limited, 1978. Heinrich, Herbert. “The Discovery of Galvanoplasty and Electrotyping.” Journal of Chemical Education 15, no. 12 (December 1938): 566–75. Hobsbawm, Eric. The Age of Empire: 1875–1914. London: Vintage, 1989. ———. The Age of Extremes. London: Michael Joseph, 1994. Hosking, R. F., and G. M. Meredith-Owens, eds. Handbook of Asian Scripts. London: The British Museum, 1966. Hourani, Albert. Arabic thought in the liberal age, 1798–1939. London, New York,

Bibliography

487

­Toronto: Oxford University Press, 1962. Howe, Ellic & Harold E. Waite. The London Society of Compositors (re-established 1848): a centenary history. London, Toronto, Sydney, Melbourne: Cassell & Co., Ltd., 1948. Hunziker, Hans Jürg. “Aspects of Arabic Script Reform.” Typographische Monats­blätter, 4 (1985): 1–36. Intertype Ltd. “Untitled.” Interludes I, no. 1 (January 1931): 1. ———. “With the Editor.” Interludes III, no. 2 (April 1933): 8–9. “ITC Arabic? Yes.”, in U&lc, Upper and lower case, the international journal of t­ ypographics, 15, no. 3, August 1988, 8–11. Kahan, Basil Charles. Ottmar Mergenthaler: The Man and His Machine: a Biographical Appreciation of the Inventor on His Centennial. New Castle, DE: Oak Knoll Press, 2000. Kinross, Robin. “The digital wave”, Eye, vol. 2, no. 7 (1992): 26–39. ———. Modern Typography: An Essay in Critical History. 2nd ed. London: Hyphen Press, 2004. First published 1992 by Hyphen Press, London. ———. “Übergangsmomente: Schrifttypen von 1968 bis 1997.” In postscript. Zur Form von Schrift heute, 14–23. Ostfildern-Ruit: Hatje Cantz, 2004. Knuth, Donald E. TEX and METAFONT : New Directions in Typesetting. Bedford, MA: Digital Press and American Mathematical Society, 1979. ———. The METAFONT Book. Reading, MA: Addison Wesley, 1986. ———. The TeXbook. Reading, MA: Addison Wesley and the American Mathematical Society, 1986. ———. Digital Typography. CSLI lecture notes 78. Stanford, CA: CSLI Publications, 1999. Knuth, Donald E., and Pierre A. MacKay. “Mixing Right-to-Left Texts with Left-toRight Texts.” TUGboat 8, no. 1 (April 1987): 14–25. Krek, Miroslav. “The Enigma of the First Arabic Book Printed from Movable Type.” Journal of Near Eastern Studies 38, no. 3 (1 July 1979): 203–12. ———. “Some Observations on Printing Arabic in America and by Americans Abroad.” Manuscripts of the Middle East 6 (1992): 71–87. Kupferschmidt, Uri M. “On the diffusion of ‘small’ Western technologies and consumer goods in the Middle East during the era of the first modern globalization.’ In A Global Middle East: Mobility, Materiality and Culture in the Modern Age, 1880–1950, edited by Kozma, Schayegh, Wishnitzer, 229–262. New York: I. B. Tauris, 2015. Lane, John A., R. Breugelmans, and Jan Just Witkam. The Arabic Type Specimen of Franciscus Raphelengius’s Plantinian Printing Office. Leiden: The University Library, 1997. Lanston Monotype Machine Co. The Monotype System. A Book for Owners & Operators

488

Bibliography

of Monotypes. Philadelphia: Lanston Monotype Machine Co., 1912. Legros, Lucien Alphonse, and John Cameron Grant. Typographical Printing-Surfaces: The Technology and Mechanism of Their Production. London: Longmans, Green and Co., 1916. Lieser, Gerhard. “Production of High Quality Arabic Texts on a CRT Filmsetting ­Machine.” In Advances in computer-aided literary and linguistic research: proceedings of the Fifth International Symposium on Computers in Literary and Linguistic Research held at the University of Aston in Birmingham, UK from 3-7 April 1978, 201–13. Department of Modern Languages, University of Aston in Birmingham, 1979. Linotype & Machinery Ltd. It Can Now Be Revealed. Altrincham: Linotype & Machinery Ltd., n.d. [1946]. ———. “Printers Must Face the Present, Here There, Everywhere – Except for the Printer at Home.” Linotype Matrix II, no. 1 (Spring 1948): 1. ———. “Many Stages, Chapter IV in The Matrix Story.” Linotype Matrix II, no. 6 (Spring 1950): 7. ———. “The House of Commons, The Times and Linotype.” Linotype Matrix II, no. 10 (Summer 1951): 1. ———. “Government Press of Pakistan, A Valuable Asset to a Developing Nation.” Linotype Matrix II, no. 11 (Autumn 1951): 6. ———. “Linotype Organizations Are Making Valuable Contributions to the Typographic Resources of the East.” Linotype Matrix II, no. 14 (1953): 1. ———. “Millions of People Benefit from the Advantages of Linotype Composition in Arabic.” Linotype Matrix II, no. 23 (October 1955): 2. ———. “TTS for Book Composition Offers More Than Double the Speed Without Any Loss of Quality.” Linotype Matrix II, no. 29 (June 1958): 3. ———. “Ettela’at, First Persian-Language Newspaper to Employ Mechanical Composi-tion in Iran.” Linotype Matrix II, no. 29 (June 1958): 2. ———. “Linotype Are First in the World to Produce a System of Simplified Arabic for Mechanical Composition from One Magazine.” Linotype Matrix II, no. 32 (November 1959): 5. ———. “From the Stork’s Beak to Benton’s Punchcutter.” Linotype Matrix II, no. 33 (May 1960): 2. ———. “Now Available: 3 Sizes of Mrowa-Linotype Simplified Arabic.” Linotype Matrix II, no. 34 (December 1960): 2. Luna, Paul. Understanding Type for Desktop Publishing. New York: Blueprint, 1992. MacKay, Pierre A. “Setting Arabic with a Computer.” Scholarly Publishing 8, no. 2 (January 1977): 142–50. Marshall, Alan. Du plomb à la lumière : La Lumitype-Photon et la naissance des industries graphiques modernes. Paris: Éditions de la Maison des sciences de

Bibliography

489

l’homme, 2003. Marzolph, Ulrich. “Early Printing History in Iran (1817 – ca. 1900).” In Middle Eastern languages and the print revolution: a cross-cultural encounter: a catalogue and companion to the exhibition, 249–68. Westhofen: WVA-Verlag Skulima, 2002. Massoudy, Hassan, and Isabelle Massoudy. Abcdaire de la calligraphie arabe. Paris: Flammarion, 2003. Mergenthaler Linotype Company. Specimen Book of Type Styles. New York: ­Mergenthaler Linotype Company, 1915. ———. “Al-munaḍḍadä al-ʿarabiyyä”. New York: Mergenthaler Linotype Company, 1929. ———. Linotype Machine Principles. Brooklyn, New York: Mergenthaler Linotype Company, 1940. Meynet, Roland. L’écriture Arabe En Question, Les Projets de l’Académie de Langue Arabe Du Caire de 1938 à 1968. Hommes et Sociétés du Proche-Orient 3. Beirut: Dar El-Machreq, 1971. Milo, Thomas. “Arabic Script and Typography.” In Language, Culture, Type, edited by John D. Berry, 112–27. New York: ATypI Graphis, 2002. ———. “Authentic Arabic: A Case Study. Right-to-Left Font Structure, Font Design and Typography.” Manuscripta Orientalia 8, no. 1 (2002): 49–61. ———. “Ali-Baba and the 4.0 Unicode Characters: New input and output concepts under Unicode,” TUGboat 24, no. 3 (2003): 502–511. ———. “Arabic Amphibious Characters. Phonetics, Phonology, Orthography, ­Calligraphy and Typography.” In Vom Koran zum Islam, edited by Markus Groß and Ohlig. Berlin & Tübingen: Verlag Hans Schiler, 2009. ———. “Balancing Arabic and Latin typography,” Book 2.0, no. 2 (2011): 239–254. Milo, Thomas, and Mohamed Zakaria. “Beautiful Book, Made to the Highest Western Standards.” Al-Computer, Communications & Electronics, 2006. Moginet, Stefan F. Du calame à l’ordinateur : l’évolution graphique de l’écriture arabe. Gap: Atelier Perrousseaux, 2009. Mokarzel, Mary. Al-Hoda, 1898–1968: The Story of Lebanon and Its Emigrants Taken from the Newspaper Al-Hoda. New York: Al-Hoda, 1968. Mirdehghan, Mahinnaz. “Persian, Urdu, and Pashto: A Comparative Orthographic Analysis.” Writing Systems Research 2, no. 1 (2010): 9–23. Mitchell, T.F. Writing Arabic, A Practical Introduction to Ruq’ah Script. London: Oxford University Press, 1953. Monotype Corporation Ltd. Monotype Newsletter. Holiday Number 7, 1934. ———. Monotype Newsletter. no. 33, (October 1937). ———. “Backmatter.” Monotype Newsletter no. 39 (May 1939). ———. “Arabic in All Its Glory.” The Monotype Recorder 42, no. 2 (Spring 1961): 24. ———. “‘By-Passing Gutenberg’ in Karachi.” The Monotype Recorder 42, no. 2 (Spring

490

Bibliography

1961): 13–14. ———. “Metal to Film: What Is Involved in the Transition?” The Monotype Recorder 42, no. 2 (Spring 1961): 1–2. ———. “Greatest Exotic Language Triumph for ‘Monotype’.” The Monotype Recorder no. 3. New Series (October 1981): 1. Moran, James. “Filmsetting, Bibliographical Implications.” The Library XV, no. 4. Fifth Series (1960): 231–45. ———. Stanley Morison. His Typographic Achievement. London: Lund Humphries Publishers Ltd., 1971. Muench, Roger. “The origins of modern filmsetting: the Uhertype: a research report.” Journal of the Printing Historical Society no. 3, New Series (2001): 21–40. Naqvi, Jamil A. “Nastaliq – the Elegant Urdu Script – Its Origin and Progress from the Sixth Century to the Present Day.” The Monotype Recorder no. 3. New Series (October 1981): 1–4. Nemeth, Titus. “Complex Dutch Arab Complex.” TYPO no. 36 (summer 2009): 86–93. ———. “Simplified Arabic, a New Form of Arabic Type for Hot Metal Composition.” Typography Papers no. 9 (2013): 173–89. ———. “‘Harmonised type design’ revisited,” in Digital Fonts and Reading, Series on Language Processing, Pattern Recognition and Intelligent Systems 1, edited by Mary Dyson & Ching Suen, 150–72. Singapore: World Scientific Publishing, 2015. Neumann, Christoph K. “Book and Newspaper Printing in Turkish.” In Middle Eastern languages and the print revolution: a cross-cultural encounter: a catalogue and companion to the exhibition, 227–48. Westhofen: WVA-Verlag Skulima, 2002. Osborn, J. R. Letters of Light: Arabic Script in Calligraphy, Print, and Digital Design. Cambridge, MA & London: Harvard University Press, 2017. Orcutt, Reginald. Merchant of Alphabets. London: Jarrolds Ltd., n.d. [1945]. Pfiffner, Pamela. Inside the Publishing Revolution: The Adobe Story. Adobe Press, 2002. Phillips, Arthur H. Computer Peripherals and Typesetting: A Study of the Man-Machine Interface Incorporating a Survey of Computer Peripherals and Typographic Composing Equipment. London: Her Majesty’s Stationery Office, 1968. ———. Handbook of Computer Aided Composition. New York and Basel: Marcel Dekker Inc, 1980. Plooy, Edward B. “A New System of Photo-Composing Arabic Script.” Quaerendo IV, no. 4 (1974): 330–32. Porter, Andrew. The Nineteenth Century, The Oxford History of the British Empire, vol. III. Oxford University Press, 1998. Proudfoot, Ian. “Mass producing Houri’s moles, or Aesthetics and Choice of Technology in Early Muslim Book Printing,” in Islam: Essays On Scripture, Thought and Society, A Festschrift in Honour of Anthony H. Johns, 161–89. Leiden: Brill, 1997. Reinhart, A. Kevin. “Arabic on the Macintosh: Overview and Review.” Middle East

Bibliography

491

Studies Association Bulletin 26, no. 2 (December 1992): 159–66. Richards, George Tilghman. The History and Development of Typewriters. 2nd rev. ed. London: Her Majesty’s Stationery Office, 1964. Richards, John F. The Mughal Empire. Cambridge University Press, 1996. Ritter, Hellmut. “Über einige Regeln, die beim Drucken mit arabischen Typen zu beachten sind.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 25, no. 100. Neue Folge (1950): 577–80. Robinson, Francis. “Technology and Religious Change: Islam and the Impact of Print.” ­Modern Asian Studies 27, no. 1 (February 1993): 229–51. Romano, Frank. History of the Linotype Company. Rochester, New York: Rochester Institute of Technology, RIT Press, 2014. ———. Machine Writing and Typesetting: The Story of Sholes and Mergenthaler and the Invention of the Typewriter and the Linotype. Saleh, New Jersey: GAMA, 1986. Roper, Geoffrey. “Fāris Al-Shidyāq and the Transition from Scribal to Print Culture in the Middle East,” in The Book in the Islamic world, edited by George N. Atiyeh, 209–33. Albany: State University of New York Press, 1995. ———. ed. Historical aspects of printing and publishing in languages of the Middle East: papers from the Third Symposium on the History of Printing and Publishing in the Languages and Countries of the Middle East, University of Leipzig, ­September 2008. Leiden: Brill, 2014. ———. ed. The History of the Book in the Middle East. Farnham: Ashgate, 2013. Ross, Fiona. “Translating Non-Latin Scripts into Type.” Typography Papers no. 3 (1998): 75–86. ———. The Printed Bengali Character and Its Evolution. Richmond, Surrey: Curzon, 1999. ———. “Non-Latin Typesetting in the Digital Age.” In Computers and Typography 2, 42–53. Bristol: Intellect Books, 2002. ———. “Selected Exhibits from the Non-Latin Type Collection, University of Reading.” In Non-Latin scripts, from metal to digital type. London: St Bride Library, 2012. ———. “The Type Design Process for Non-Latin Scripts.” In Non-Latin scripts, from metal to digital type, 125–54. London: St Bride Library, 2012. Safadi, Yasin Hamid. Islamic Calligraphy. London: Thames & Hudson, 1978. ———. “Printing in Arabic.” The Monotype Recorder no. 2. New Series (October 1981): 2–7. Safwat, Nabil F. The Art of the Pen: Calligraphy of the 14th to 20th Centuries. London & Oxford: The Nour Foundation in association with Azimuth Editions and Oxford University Press, 1996. Said, Edward W. Out of Place: a Memoir. New York: Vintage Books, 2000. Saunders, David. “Two Decades of Change 1965–1986.” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series

492

Bibliography

(1997): 26–35. Schlesinger, Carl. The Biography of Ottmar Mergenthaler. New Castle, DE: Oak Knoll Books, 1989. Seybold, John W. Fundamentals of Modern Composition. Media, PA: Seybold Publications Inc., 1977. ———. “Digitized Type: What Is It? What Does It Mean for Typesetting and Word Processing?” The Seybold Report on Publishing Systems 8, no. 24 (27 August 1979): 1–17. ———. “Mergenthaler Linotype’s CRTronic: A Review and Status Report.” The Seybold Report on Publishing Systems 10, no. 8 (22 December, 1980): 1. ———. The World of Digital Typesetting. Media, PA: Seybold Publications Inc., 1984. ———. “An Imagesetter for the Rest of Us: Apple’s LaserWriter.” The Seybold Report on Publishing Systems 14, no. 9 (28 January 1985): 3–26. Shaw, Graham. Printing in Calcutta to 1800: A Description and Checklist of Printing in Late 18th-Century Calcutta. London: The Bibliographical Society, 1981. ———. “Non-Latin Scripts and Printing Technologies: Triumphs and Tribulations.” In Non-Latin scripts, from metal to digital type, 11–34. London: St Bride Library, 2012. Sheehan, Edward R. F. “The Most Powerful Journalist in the World, His Newspaper and His Book.” In The Cairo documents: the inside story of Nasser and his relationship with world leaders, rebels, and statesmen. New York: Doubleday, 1973. Sheppard, Graham. “Urdu Nastaliq, 1981.” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 31. Silcock, E. “‘Monotype’ Photo Typographical Composing Machine.” In Penrose Annual, 44:102. London: Lund Humphries & Co. Ltd., 1950. Smith, Jason. “Book Review.” Eye Magazine 12, no. 45 (2002): 82. Smitshuijzen AbiFarès, Huda. Arabic Typography: a Comprehensive Sourcebook. London: Saqi Books, 2001. ———. Typographic Matchmaking: Building Cultural Bridges with Typeface Design. Amsterdam: BIS Publishers, 2006. Smitshuijzen, Edo. Arabic Font Specimen Book. Amsterdam: De Buitenkant, 2009. Southall, Richard. “A Survey of Type Design Techniques before 1978.” Typography Papers no. 2 (1997): 31–59. ———. Printer’s Type in the Twentieth Century: Manufacturing and Design Methods. London & New Castle, DE: The British Library & Oak Knoll Press, 2005. ———. “Technical History of the Monotype Composing Machines,” in History of the Monotype Corporation, edited by Andrew and Boag and Christopher Burke, 303–414. Woodstock: Vanbrugh Press, 2014. Spencer, Herbert. The Visible Word. 2nd rev. ed. London: Lund Humphries in association with the Royal College of Art, 1969. First published 1968 by Royal College of Art, London.

Bibliography

493

Steinberg, S.H. Five Hundred Years of Printing. rev. ed. by John Trevitt. London & New Castle, DE: The British Library & Oak Knoll Press, 1996. Tachau, Frank. “Language and Politics: Turkish Language Reform.” The Review of Politics 26, no. 2 (1 April 1964): 191–204. Taymour, Mahmoud. “A New Arabic Script to Facilitate the Use of the Diacritical Points Essential for Correct Speech and Writing.” The Islamic Review (October 1951): 16–20. Thompson, John Smith. History of Composing Machines. Chicago: The Inland Printer Company, 1904. Tiknūlūǧīā al-Kahrabāʾ. The Ministry of Education, Saudi Arabia, Ernst Klett, Stuttgart, and Interpart, Stuttgart, 1979. First published 1972 as Fachkunde für Elektroberufe, ed. Robert Arnold, by Ernst Klett, Stuttgart. Tosh, John. The Pursuit of History. Aims, Methods, & New Directions in the Study of Modern History. 2nd rev. ed. London and New York: Longman, 1991. First ­published 1984 by Longman, London. ———. The pursuit of history: aims, methods and new directions in the study of history, 6th rev. ed. Abingdon-on-Thames & New York: Routledge, 201). First published 1984 by Longman, London. Tracy, Walter. “The Flourishing Reed.” In Alphabet: International Annual of L­ etterforms. London: James Moran Limited, 1964. ———. “Advances in Arabic Printing.” Bulletin (British Society for Middle Eastern S­ tudies) 2, no. 2 (1975): 87–93. ———. “Arabic Without Tears.” In The Penrose Graphic Arts International Annual, 68: 121–26. London: Northwood, 1975. ———. “Composing Room Days (and After).” Printing Historical Society Bulletin 40, no. Winter 1995/96 (1995): 3–15. ———. Letters of Credit: A View of Type Design. Boston: David R Godine Publishers, 1987. Tripsas, Mary. “Surviving Radical Technology Change through Dynamic Capability: Evidence from the Typesetter Industry.” Industrial and Corporate Change 6, no. 2 (1997): 341–77. Twyman, Michael. Printing 1770–1970. An Illustrated History of Its Development and Uses in England. London: Eyre & Spottiswoode, 1970. The Unicode Consortium. The Unicode Standard. Worldwide Character Encoding. Vol. 1. Reading, MA: Addison-Wesley, 1991. van Krimpen, Jan. “On Preparing Designs for Monotype Faces.” In Type & Typography. Highlights from the Matrix, the review for printers & bibliophiles. New York: Mark Batty Publisher, 2003. Previously published in Matrix 11 (1991). Vervliet, Hendrik D. L. “Cyrillic & Oriental Typography in Rome at the End of the Sixteenth Century: An Inquiry into the Later Work of Robert Granjon; 1578–1590.”

494

Bibliography

In The Palaeotypography of the French Renaissance: Selected Papers on Sixteenthcentury Typefaces, 2:427–74. Leiden: Brill, 2008. Vilhjálmsson, Gunnar. “Ruq’ah – the new style.” The Recorder, no. 1 (2014): 107–16. De Vinne, Theodore Low. The Practice of Typography. A Treatise on the Processes of Typemaking, the Point System, the Names, Sizes, Styles and Prices of Plain ­Printing Types. New York: The Century Co., 1902. Walker, Victor E. “Mr V. E. Walker, Chairman and Managing Director of Linotype & Machinery Ltd., Writes to Friends in the Trade and Outlines the Company’s Aims and Prospects.” Linotype Matrix II, no. 3 (Spring 1949): 1. Wallis, Lawrence W. A Concise Chronology of Typesetting Developments 1886–1986. ­London: The Wynkyn de Worde Society in association with Lund Humphries, 1988. ———. Typomania. Upton-upon-Severn: Severnside Printers Limited, 1993. ———. “Dates with Phototypesetting.” In Typomania, 45–61. Upton-upon-Severn: Severnside Printers Limited, 1993. ———. “From Frame to Desktop: a Century of Typesetting.” In Typomania, 66–79. Upton-upon-Severn: Severnside Printers Limited, 1993. ———. “Forty Years On, a Personal Retrospective and Prospective of Typesetting Technology.” In Typomania, 80–85. Upton-upon-Severn: Severnside Printers Limited, 1993. ———. “Monotype Time Check.” The Monotype Recorder, One Hundred Years of Type Making: 1897–1997, Centenary Issue, no. 10. New Series (1997): 46–55. Wilson-Davies, Kirty, Joseph St John Bate, and Michael Barnard. Desktop Publishing. 3rd ed. Publisher’s Guide Series. London: Blueprint, 1989. First published 1988 by BSP Professional, Oxford. Winslow, Charles. Lebanon: War and Politics in a Fragmented Society. London & NY: Routledge, 1996. First published 1988 by BSP Professional, Oxford. Yosmaoğlu, İpek K. “Chasing the Printed Word: Press Censorship in the Ottoman ­Empire, 1876–1913,” The Turkish Studies Association Journal, 27, no. 1/2 (2003), 15–49. Zapf, Hermann. About Alphabets: Some Marginal Notes on Type Design. Cambridge & London: M.I.T. Press, 1960. ———. Alphabet Stories: A Chronicle of Technical Developments. Bad Homburg: Mergenthaler Edition, Linotype GmbH, 2007.

Patents Benton, Linn Boyd. Printing-Type. US Patent 290,201, filed 8 May 1883, and issued 18 December 1883.

Bibliography

495

———. Punch Cutting Machine. US Patent 332,990, filed 29 February 1884, and issued 22 December 1885. Haddad, Selim S. Types for Type-Writers or Printing-Presses. US patent 637,109, filed August 12, 1899 and issued November 14, 1899. Hanson, Ellis. Control System for Typesetting Arabic. US Patent 3,513,968, filed 24 January 1967, and issued 26 May 1970. Holloway, Timothy, Michael Fellows, and Fiona Ross. Printing. UK Patent 2208556A, filed 12 August 1987, and issued 5 April 1989. Khattar, Nasri Najib. Arabic Alphabet. US Patent D.157,983, filed 3 October 1947, and issued 4 April 1950. Lakhdar, Ahmed. Perfectionnements aux caractères d’imprimerie d’écriture arabe. French Patent 1.145.687, filed 29 December 1954, and issued 29 October 1957. Lamberti, Sergio. Means For Controlling Typographic Composing Machines. UK ­Patent GB1162180 (A), filed 24 December 1966, and issued 20 August 1969. Linotype & Machinery Ltd. An improved manner of and a type font for printing in Urdu or like script. Indian Patent 27030, filed 5 December 1939, and issued 4 July 1940. Marks, George Croydon, Chartered Patent Agent on behalf of the Hammond TypeWriter Company. Improvements in Type-writers. UK Patent GB190110460 filed 20 May 1901, and issued 22 June 1901. Pierpont, Frank Hinman. Improvements in or relating to the Casting and Composing of Single Types in Semitic Languages. UK Patent 28,692, filed 6 December 1910, and issued 7 December 1911. Sholes, Christopher Latham, Carlos Glidden, and Samuel Soulé. Type Writing Machine. US Patent 79,265, issued 23 June 1868. Plooij, Dr. Edward Bernard. Apparatus for Optically Projecting Characters onto a Light Sensitive Surface. UK Patent 972,071, filed 13 December 1960, and issued 7 October 1964. ———. Inrichting voor het optisch projecteren van tekens op een opnemer. Dutch Patent NL123701, filed 14 December 1959, and issued 15 September 1967. ———. A Method and Apparatus for Typewriting or Composing Arabic or Related Writing. UK Patent 1,176,523, filed 10 January 1967, and issued 7 January 1970. Plooy, Dr. Edward Bernard. Inrichting voor het machineschrijven of zetten van Arabisch en daarmee verwant schrift. Dutch Patent NL113873, filed 17 July 1959, and issued 15 June 1967. Westover, George. Improvements relating to Methods and Apparatus for Photographically Composing Text-matter. UK Patent 499,859, filed 27 April 1937, and issued 27 January 1939.

496

Bibliography

Online Sources Batty, Cynthia. to ATypI member discussion list, Email. “Re: [ATypI] History”. 17 February 2012. Becker, Joseph D. “Unicode 88.” The Unicode Consortium, 1988. www.unicode.org/ history/unicode88.pdf (accessed 16 May 2017). Center for Language Engineering, “Valid Ligatures of Urdu,” 13 January 2011, http://www. cle.org.pk/software/ling_resources/UrduLigatures.htm (accessed 16 May 2017). Concept Software Private Ltd. “InPage Urdu,” n.d. http://inpage.com.pk/ (accessed 16 May 2017). Encyclopædia Britannica Online, “Egypt.” https://www.britannica.com/place/Egypt (accessed 16 May 2017). ———. “Muhammad Hassanein Heikal.” https://www.britannica.com/biography/ Muhammad-Hassanein-Heikal (accessed 16 May 2017). ———. “Reza Shah Pahlavi.” https://www.britannica.com/biography/Reza-ShahPahlavi (accessed 16 May 2017). ———. “World War II.” https://www.britannica.com/event/World-War-II (accessed 16 May 2017). “Ethnologue. Languages of the World.”, www.ethnologue.com (accessed 16 May 2017). Heijdra, Martin J. “The Development of Modern Typography in East Asia, 1850-2000”, The East Asian Library Journal 11, no. 2 (2004): 100–168, https://library.princeton. edu/eastasian/EALJ/heijdra_martin_j.EALJ.v11.n02.p100.pdf (accessed 16 May 2017). Hudson, John. “Microsoft Typography - Windows Glyph Processing: An OpenType Primer,” November 2000. www.microsoft.com/typography/Glyph%20Processing/intro.mspx (accessed 16 May 2017). ———. to ATypI member discussion list, Email. “Re: [ATypI] a New Arabic Typography Book”. 31 March 2011. “Incunabula Short Title Catalogue,” British Library, http://www.bl.uk/catalogues/istc/ index.html (accessed 16 May 2017). “John Hudson ISType 2012, June 15.” Istanbul, 17 October 2012. http://vimeo. com/51580484 (accessed 16 May 2017). Klett Gruppe. “Klett Gruppe - The Klett Group - Dedicated to Education.” www.klettgruppe.de/home/the+klett+group/dedicated+to+education.765.htm (accessed 16 May 2017). Leonidas, Gerry. “The Next Ten Years,” 12 March 2013. http://leonidas.org/2013/03/12/ thenexttenyears/ (accessed 16 May 2017). MacKay, Pierre A. Interview by Dave Walden, 2007. http://tug.org/interviews/mackayp.html (accessed 16 May 2017). McMurtrie, Douglas. The Ludlow System of Slugline Composition. Chicago: Ludlow

Bibliography

497

­Typograph Company, 1927. http://archive.org/details/McMurtrieLudlowSlugline (accessed 16 May 2017). Microsoft Typography. “Simplified Arabic,” 2013. www.microsoft.com/typography/ fonts/font.aspx?FMID=1879 (accessed 16 May 2017). ———. “JSTF - The Justification Table,” 27 March 1998. www.microsoft.com/typography/otspec/jstf.htm (accessed 16 May 2017). Milo, Thomas. “Some Comments on the Arabic Block in Unicode,” 1999. www.unicode.org/L2/L2001/01329-milo-com.pdf (accessed 16 May 2017). Monotype GmbH. “Lotus® Arabic Font Family,” www.linotype.com/1183/LotusArabicfamily.html (accessed 16 May 2017). ———. “Palatino® Arabic Font Family,” www.linotype.com/286269/PalatinoArabicfamily.html (accessed 16 May 2017). ———. “Yakout Font Family,” www.linotype.com/1618/yakout-family.html (accessed 16 May 2017). NYTimes.com “Kemal Says Latin Alphabet Will Oust Arabic in Turkey.” The New York Times, 11 August 1928. http://select.nytimes.com/gst/abstract.html?res=F20615 F83E58127A93C3A81783D85F4C8285F9&scp=4&sq=arabic+alphabet&st=p (accessed 16 May 2017). ———. “Princeton to Print Textbooks in Arabic; University First in America to Be Equipped to Turn Out Such Work on Its Own Press.” The New York Times, 19 October 1930. http://select.nytimes.com/gst/abstract.html?res=F70E1EFA3A5C1 1738DDDA00994D8415B808FF1D3&scp=2&sq=linotype+arabic&st=p (accessed 16 May 2017). ———. “Arabic Paper Here Now 50 Years Old; Editor of Al-Hoda Recalls First Linotype Installed in 1912 as Significant Milestone.” The New York Times, 9 August 1948. http://select.nytimes.com/gst/abstract.html?res=F70613FC3F5C1B7B93CB A91783D85F4C8485F9&scp=2&sq=al+hoda&st=p (accessed 16 May 2017). OED Online, “Morphology, N.,” (Oxford University Press, n.d.), http://www.oed.com/ (accessed 16 May 2017). OIC Research Centre for Islamic History, Art and Culture. “About Us.” www.ircica.org/ about-us/irc380.aspx (accessed 16 May 2017). Officer, Lawrence H., and Samuel H. Williamson. “Five Ways to Compute the Relative Value of a U.K. Pound Amount, 1270 to Present.” MeasuringWorth, 2017. www. measuringworth.com/ukcompare/ (accessed 16 May 2017). ———. “Seven Ways to Compute the Relative Value of a U.S. Dollar Amount, 1774 to Present.” MeasuringWorth, 2017. www.measuringworth.com/uscompare/ (accessed 16 May 2017). Phinney, Thomas W. “TrueType & PostScript Type 1: What’s the Difference?” truetypetypography.com, 1 October 1997. www.truetype-typography.com/articles/ttvst1. htm (accessed 16 May 2017).

498

Bibliography

Reichardt, Hans. Chronik der Schriftgiesserei Stempel. Frankfurt am Main, 2002. www.schriftenservice-d-stempel.de/ChronikStempel.pdf (accessed 16 May 2017). Research Centre for Islamic History, Art and Culture (IRCICA). “8th International Calligraphy Competition Dedicated to Badawī Al-Diranī.” 2009. www.ircica.org/ content_images/8th-International-Calligraphy-Competition-ENG.pdf (accessed 16 May 2017). Roper, Geoffrey. “The History of the Book in the Muslim World.” In The Oxford Companion to the Book, edited by Michael F. Suarez and H. R. ­Woudhuysen. Vol. 1. Oxford Reference Online, 2010. www.oxfordreference. com/view/10.1093/acref/9780198606536.001.0001/acref-9780198606536-e3986?rskey=PCys83&result=7&q=qur%27an (accessed 16 May 2017). Rosetta Type Foundry. “About,” n.d. www.rosettatype.com/about (accessed 16 May 2017). Ross, Fiona. “Non-Latin Type Design at Linotype,” 2002. http://tntypography.eu/resources-list/non-latin-type-design-linotype-fiona-ross/ (accessed 16 May 2017). Rūznāmah Kayhān (Kayhan Newspaper). n.d. www.kayhan.ir/ (accessed 16 May 2017). Sakkal, Mamoun. “A Brief Survey of Proposals to Simplify Arabic Script”. Sakkal Design, 2008. http://sakkal.com/articles/simplified_arabic/survey.html (accessed 16 May 2017). Searle, Steven J. “Unicode Revisited,” http://tronweb.super-nova.co.jp/unicoderevisited.html (accessed 16 May 2017). Topping, Suzanne. “The Secret Life of Unicode: A Peek at Unicode’s Soft Underbelly,” 1 May 2001. http://tetrud.com/Billys/Lima/The%20Secret%20Life%20of%20 Unicode.pdf (accessed 16 May 2017). Tradigital Stuttgart GmbH. “TRADIGITAL - Philosophy,” www.tradigital.de/about.htm (accessed 16 May 2017). Unesco. Assistance to the Institute for Studies and Research on Arabization (IERA) Terminal Report. Paris, France, 25 November 1985. UNESDOC. http://unesdoc. unesco.org/ulis/cgi-bin/ulis.pl?catno=67225&set=51692385_2_339&gp=1&lin=1 &ll=1 (accessed 16 May 2017). ———. “Morocco Answers a Royal Call to the Classroom.” The UNESCO Courier, March 1958. http://unesdoc.unesco.org/images/0007/000781/078161eo.pdf ­(accessed 16 May 2017). ———. EFA Global Monitoring Report. Youth and Skills: Putting Education to Work. Paris, 2012. www.efareport.unesco.org/ (accessed 16 May 2017). Unicode. “Chronology of Unicode Version 1.0.” History of Unicode. www.unicode.org/ history/versionone.html (accessed 16 May 2017). Université Mohammed V, “Institut d’Études et de Recherches pour l’Arabisation,” 2011, http://www.um5.ac.ma/um5r/fr/content/institut-des-etudes-et-des-recherchespour-l%E2%80%99arabisation (accessed 16 May 2017).

Bibliography

499

Unpublished Sources Brée, Holger. to Rory Cowan, Telex, 30 June 1978. Private collection Hrant Gabeyan. Cowan, Rory. to Hrant Gabeyan, Letter, 16 February 1978. Private collection Hrant Gabeyan. De Baerdemaeker, Jo. “Tibetan Typeforms: An Historical and Visual Analysis of ­Tibetan Typefaces from Their Inception in 1738 up to 2009.” PhD thesis, University of Reading, 2009. Fellows, Michael. Interview by the author. Audio recording, 9 August 2013. Private collection of the author. Gabeyan, Hrant. to Walter Tracy, Letter, “The ‘Kashida’,” 31 January 1967. Private collection Hrant Gabeyan. ———. to Peter Schacht, Letter, 28 December 1978. Private collection Hrant Gabeyan. Hanson, Ellis. to Arthur H Walker, Letter, 22 December 1966. Private collection Hrant Gabeyan. Reef, Steven. to Carey et. al., Memorandum, “Second Revision for Meeting with Hrant Gabeyan,” 14 July 1978, Private collection Hrant Gabeyan. Ross, Fiona. Interview by the author. Audio recording, 13 April 2013. Private collection of the author. Savoie, Alice. “International cross-currents in typeface design: France, Britain and the USA in the phototypesetting era.” PhD thesis, University of Reading, 2014.

Index People ʿAbbās, Khedive 30 ʿAbd al-ʿAzīz, Sultan 412 Abi Raad, Joseph 409 ad-Dīrānī, Muḥammad Badawī 243 Afash, Hasan Abu 472 Ainsworth, C. A. 124 al-Assad, Hafez 243 al-Ǧamālī, Badr 254 al-Mustaʿṣimī, Yāqūt 129 al-Šidyāq, Fāris 31 Andrews, Walter 324 Atatürk, Mustafa Kemal 83 Baḥarī, Hišām 47, 200, 354–356 Barrett, G. C.  95, 274–275 Barrett, Gillian 259 Banks, Jack 48 Batty, Cynthia 390, 460 Beattie, E. 235 Bender, Harold H. 151 Benton, Linn Boyd 38–43, 134 Bigelow, Charles 471 Bisley, Anthony 258–259 Bolander, Tage 275–279 Boothman, Martin 308, 376, 394 Boulos, Adly 329–330, 333–334 Boutros, Mourad 388 Brainerd, Paul 389 Bruce, David 37 Bryant, Frank 190, 195, 278–279 Burke, Jackson 48–49, 87–88, 91, 110–114, 119, 218 Burney, Mohammad Elias 139 Caesar, Richard J. 96–97, 241, 247, 268, 305–308, 441 Chaignaud, Paul 91 Calderwood, Jillian 247 Church, William 37 Coates, Ros 259–260, 366, 374–375, 378 Collier, L. A. 95, 225 Cortbawi, Friar 225–226 Cowan, Rory 335–337 Danforth, William 261 Davidson, Arthur Clark 88

Day, C. T. 180, 219–222, 346–347, 356 Dib, George 243 Dreyfus, John 226 Dubash, Ronnie 378 Dunraven, Earl of 131 Dwiggins, William Addison 56 Eger, Jürgen 366, 375, 378 Elsdon, Paul 374 Emery, Edward 193–195, 232, 238, 264 Fārūq, King 87 Fellows, C. 213, 275 Fellows, Michael 251, 256, 358, 361, 367–371, 375–376, 382, 394, 454 Firmage, E. A. 95, 182, 213–214, 220–222, 225 Fletcher Rogers, D. G. 126, 226 Frutiger, Adrian 101, 459 Fuʾād, King 82, 161–162, 446, 449 Gabeyan, Hrant 99, 129, 191–203, 233–236, 240–254, 328–339, 355–356, 439–444 Garth, William Jr. 189, 262, 282 Gauerke, Carl 88 Genoux, Jean-Baptiste 36 Genower, A. E. 254, 376–378 Geschke, Charles 384 Goodall, Percy 158, 166, 180, 220 Goodman, Ralph 128, 202, 231–234, 277, 297–298, 301, 306 Goulding, John 95, 222–223, 226 Govil, Hari Govind 151 Greisner, Walter 255 Griffith, Chauncey 48–49, 69–72, 112, 150–152 Gutenberg, Johannes 9, 24–25, 35–37, 138, 218, 281–282, 406, 428 Habis, Gabriel 409 Haddad, Selim, S. 74–75 Haghighi, Hossein Abdollahzadeh 298–303, 315–318, 444, 448–450 Hahl, August 41 Hamm, Roberto 255 Hanna, George 423 Hanson, Ellis 189–190, 193, 196–197, 201–202, 262

Index Harrison, R. L. S. 277 Hattersley, Robert 38 Heikal, Mohamed Hassanein 187–188 Henry, George 363–366 Higonnet, René A. 205, 261–262, 274 Hiltmann, Reinhard 366 Holloway, Timothy 255, 308, 311, 314, 318, 357–358, 361–362, 365–368, 371–376, 379, 394–395, 398, 444, 454 Holmes, Kris 471 Homayoun, Dariush 297 Hörl, Herrmann 334–335 Hunziker, Hans-Jürg 101–102 Hussein, Osman 243–249, 254 Ini, Joyce 264, 267, 270 Ismaʿīl, Khedive 30 Jacobi, Moritz von 36 Jamil, Ahmad Mirza 346–352 Jaroudi, Nabih 114–116, 126, 129–131, 293 Jesmanowicz, Stanislaw 411, 427 Kauffmann, Paul 265 Khambatta, Feroze K. 369, 376–378 Khattar, Nasri Najib 86–91, 113, 281 King, Thomas 152 Knuth, Donald E. 327, 453 Kyte, Derek 357 Lacoche, Claude 264–267, 270 Lakhdar Ghazal, Ahmed 86, 90–103, 172, 253, 256, 264–265, 268, 329, 339 Lamberti, Sergio 193–197 Lanston, Tolbert 131–134, 138 Laubach, Frank 87 Lequint, P. 95 Lieser, Gerhard 329–331 Lindell, Joseph 156–160, 164–166, 173–182, 219–222, 442–444 MacKay, Pierre A. 323–327, 334, 378, 452–457 Mackey, Joseph T. 69–71 Mandel, Ladislas 96, 264–272 Massoudi, Farhad 293–295, 298, 301–302, 309–310 Mazunder, H. L. 140 McIntosh, Ronald 236 Mergenthaler, Ottmar 41 Milo, Thomas 410–425, 429, 455 Mitry, Chafik 173–177, 214 Mokarzel, Naoum 50–51, 55 Mokarzel, Salloum A. 49–52, 55, 59, 65, 74,

501 436, 441 Morey, Walter 183 Morison, Stanley 93, 138, 182 Morley, Sarah 260, 358, 361–362, 368 Moshkriz, S. 292 Moyroud, Louis 205, 261–262, 274 Mrowa, Kamel 87–88, 93, 109–116, 119–124, 128–131, 187, 226, 444 Muḥammad ʿAli 27–30 Mühendisoğlu, Ohannis 61, 160, 412 Mümtaz Efendi 352 Mustafa İzzet Efendi 352–353, 412, 425 Müteferrika, İbrahim 26–28 Nahas, Michael Adib 123, 173, 192, 201–202, 233, 290–294, 297 Nasser, Gamal Abdel 187–188, 192 O’Brien, Walter 264–265 Orcutt, Reginald 48, 288 Pahlavi, Rezā Shāh 287, 305 Panossian, Dikram 260, 394, 444 Panter-Brick, Keith 96 Parekh, Ish 372, 375–377 Parker, Michael 73, 201–202, 233–235, 243, 251–255, 444 Pierpont, Frank Hinman 140 Pillai, R. V. 140, 144 Plooij, Edward Bernard 233, 274–283, 313, 324, 331, 378, 418, 421, 443, 452–457 Pollock, Peter 249 Poore, C. A. 216, 226 Purdy, Peter 236 Quadri, Abdul Wadood 151 Reef, Steven 337–339, 409 Roethe, Marshall 88 Ross, Fiona 258–260, 361, 365–372, 375–376, 396, 444, 454, 466 Said, Edward 74 Saʿīd, Khedive 30 Saiyid, Matlubul Hasan 346–351 Sālim, ʿAbd ar-Rāziq Muḥammad 278 Sarkis, Ḫalīl 31 Saunders, B. 197, 212, 223, 349 Schacht, Peter 337–338 Seifert, Georg 469 Senefelder, Alois 150 Sholes, Christopher Latham 73, 132 Silcock, Elfrid 166–168, 175–176, 211 Simpson, Ron 270–272 Sintès, Gérard 91, 96–99

502 Sit, Thomas 251 Smith, Bernard 261 Somers, Mirjam 411–417, 425, 455 Somers, Peter 411–417, 425, 455 Stevens, D. 225 Stolk, Robert 425 Stone, Sumner 390 Stückradt, Dr. V.,  376 Tan, Margaret 255–258 Taqlā, Bishāra and Salīm 33 Taymour, Mahmoud 85 Tracy, Walter 48, 88–91, 96–102, 111–114, 119, 122, 183, 193–194, 197, 200–201, 226–227, 233–235, 240–259, 275–283, 288, 292–298, 301, 304–310, 314–318, 357, 367, 443–444, 452–454

Index Tueni, Walid 439 Turner, C. G. 222 Turner, Douglas 189, 296 ʿUṯmān, Caliph 162 van Krimpen, Jan 137, 158, 207 Vesey, E. A. 225, 308, 315 Walker, Arthur Henry 59, 96, 195–197, 202, 234, 238–240, 278–279, 304 Walker, Victor E. 71–72, 104 Warnock, John 384 Watkins, Peter 96–98, 305, 371–372 Weller, D. 274–275 Wicks, Frederick 37 Zaīdān, Ǧurǧī 31–33 Zapf, Hermann 289–290, 308, 322

Subjects Academy of the Arabic Language 82–87, 90, 437 Aden 128 Adobe Illustrator 400, 408, 432 Adobe InDesign 429–433 Adobe Photoshop 408, 432 Adobe Systems Inc. 384–394, 398–402, 408, 429–433, 444 Adobe Type Manager 401 Advanced Typography Program 256 Agfa Gevaert Corporation 460 Alahram (typeface) 289, 308 Aldus Corporation 387–389, 408, 415, 428 Algeria 29, 103–104, 128, 255 allograph 19–21, 405, 414, 438 al-Nashir al-Maktabi 407–408 Altrincham 233 Altsys Fontographer 391, 398, 460 America see United States Amsterdam 410–411, 416–420, 424–426, 434 Amsterdam Typefoundry 274 Antoine Homsy 191–192 Apple Advanced Typography 406 Apple Inc. 385–387, 390–391, 397–402, 406–408, 414, 418, 429, 468 Apple Laserwriter 385–387, 390–392, 398, 414 Apple Macintosh 385–387, 391, 396–400, 407–409, 418

Apple Macintosh Plus 408 Arabic 90 (typeface) 438–439 Arabica (typeface) 270–272 Arabic Calligraphic Engine 414, 418, 421–434, 454, 457, 466–467 Arabic Calligraphic Printer 421–422 ArabicXT 409–410 archigrapheme 19–20, 425, 432 Arial (typeface) 462–463 Armenia 49, 191, 254, 260, 443 ASCII 402, 405 ASV Codar (typeface) 90–92, 100, 103, 253 Azhar Madrasa 162 Badr (typeface) 254–256, 279, 333, 400, 443 Bahrain 128, 173–174, 180, 422 Baltimore 41 Basel 36, 306, 328 Baskerville (typeface) 213, 304 Beirut 30–32, 86–89, 96, 110, 123, 171, 192, 202–203, 219, 225–226, 233, 243, 247–249, 268–269, 409 Belgium 460 Bengali 10, 13, 45, 49, 61, 72, 137–140, 443 Benton, Waldo & Co. 41 Bézier 385, 391 bitmap 320–322, 330, 417–418, 421 Blower Linotype 41 Bobst Graphic 102 Bodoni (typeface) 213, 304

Index Bombay see Mumbay Boston Institute of Technology 151 Brill Publishers 4, 19, 25–27, 274, 416–417 Britain see United Kingdom Būlāq press see Egyptian Government Press Cairo 27, 30–32, 54, 74, 82–84, 90, 140, 156, 160–162, 165–168, 191–192, 195, 201, 233, 254, 335 calligrapher 7, 22, 108, 129, 199, 243–245, 249, 278, 287–290, 297, 315, 325, 331, 337, 352–353, 368, 439, 455 calligraphy 4–5, 14, 18, 21–22, 25–26, 84, 101, 111–113, 149, 160–162, 170, 186, 217–218, 243–245, 256, 264, 268, 276–282, 284–285, 346, 350–353, 365, 412–414, 418–422, 425, 445, 454–457 Cambridge, UK 49 International Conference on Arabic-English Bilingual Computing 422 Cambridge University Press 154 Canada 45, 131, 409, 422 caster 37, 63, 93, 132–134, 137–139, 145, 150, 155, 166, 184, 197, 209–213, 222, 264, 297, 354–355 Cathode Ray Tube 180, 236–238, 250, 307, 319, 323, 330–331, 340–344, 361, 381 character set 60, 68, 142–145, 155, 158, 180, 266, 442 character shape 57, 63, 145, 148, 168, 182, 207, 272, 301, 312–313, 325, 341, 385, 394, 442, 471 Cheltenham 259, 361 colonialism 33, 68, 104–106, 109 compositor 37–38, 43–45, 173, 186 Compugraphic Corporation 99, 189–191, 195–197, 201, 262–265, 268, 333–339, 385, 390, 409, 444, 460–463 Compugraphic EditWriter 335–336, 339, 409, 462 Constantinople see Istanbul copyfitting 222 Crimea 29 CRT see Cathode Ray Tube CRTronic 342–343, 358–361, 366, 374–378, 382, 393 Cyrillic 138, 431, 471 Damascus 354, 363 Dār al-Maʿārif 47, 156, 173, 442–444 Dār al-Qalam 243–245, 330 Deberny & Peignot 264–265, 269

503 DecoType 62, 113, 348, 410–434, 454–458, 466–467 DecoType Authentic Naskh Server (typeface) 429–431 DecoType Authentic Ruq’ah Server (typeface) 427–429 DecoType Emiri (typeface) 62 DecoType Naskh (typeface) 424–425, 429, 433 DecoType Nastaleeq Press (typeface) 434 DecoType Nastaliq (typeface) 348, 433 DecoType Professional Naskh (typeface) 424–428 DecoType Setter 429–431 DecoType Thuluth (typeface) 424, 429 Delairco 269–272 Desktop Publishing 81, 327, 389, 398–400, 406–410, 415, 421, 424–430, 434, 456–458, 462–464 Devanagari 138–139, 213, 220 diacritical signs 15, 19–23, 116, 142–143, 148, 152, 168–170, 179, 243, 250, 272, 290, 333, 365, 371–374, 394–395, 409, 434–436, 457, 470 Didot (typeface) 115, 129, 144, 206 Digiset See Hell Digiset Diwan Software Ltd. 407–408 Dr.-Ing. Rudolf Hell GmbH 322–323, 328 Drukkerij Rob Stolk B.V. 425 DRUPA 274, 277, 366 D. Stempel AG 234, 250, 255, 258, 279, 289, 311, 314, 321, 392–393 Dubai 1, 128 duplexing 69–72, 89, 115, 135, 220, 300–301 Dymo Graphics Systems 269 Egypt 26–33, 51, 54, 59–60, 69, 82, 85–87, 99, 122, 128–129, 140, 145, 155–158, 165–168, 173–177, 187, 191–192, 202, 214, 219, 234, 253–254, 264, 269, 275–278, 289–290, 328–329, 442–443, 446, 449 Egyptian Government Press 25–27, 59–60, 140, 156–160, 168, 187, 253–254, 417, 442, 446 Egyptian Ministry of Education 85 Elektron caster (Linotype) 189, 197, 202–203, 264, 298 encoding 374, 402–405, 416, 430–431 English Church Missionary Press 31 Ernst Klett Verlag 327–333, 340

504 Filmsetting see Photocomposition Fokker 411, 417 Fonderie Typographique Orientale 89, 354–355 font synopsis see character set formal writing 21 foundry type 23–24, 34, 38, 47, 49, 56, 61, 63–65, 91, 99, 103, 133–134, 140, 149, 151, 158, 160–163, 168, 170–171, 178, 219, 229, 253, 262, 289, 292, 351, 354, 414, 421, 425–426, 437, 441, 446, 448–449, 457–458, 463 fount synopsis see character set Fraktur 138 France 27–29, 32, 64, 90, 96–98, 104, 113, 205, 261, 264, 385, 408, 445 frisket 249–250, 308, 314, 321–322, 374 Gaelic 138 Garamond (typeface) 304 Gaza 472 Geeza (typeface) 408 Germany 36–37, 41, 45, 76, 80, 110, 138, 150, 156, 234, 259, 262, 274, 288–292, 323, 327–329, 342, 355, 362, 375, 378, 448, 456 Gerrit Rietveld Academie 411 Ghossoub (typeface) 264 glyph 402–404, 424, 429, 432, 462–463, 467, 470 Glyphs (software) 469–470 Glyph Systems 409, 462 Google Inc. 468 Government Printing Press of Sindh 378 grapheme 15, 19, 347 Graphic Arts Research Foundation 262 Graphite (font format) 467 Greece 29 Greek (type) 138–140, 471 grid 230, 233–235, 238, 251, 279, 307, 311, 314–316, 321 Gujarati 139 Haas foundry 99, 103 half-form 67–68, 76, 143, 355 Hammond Type-Writer Company 75–76, 91 handwriting 17, 21, 145, 152, 279 Hattat (software) 325, 453 Hattersley (typesetting machine) 38 Hebrew 15, 49, 59, 138–140, 206, 398, 432, 471 Hell Digiset 282, 322–323, 327–331, 340, 359 Hewlett-Packard 414, 417–418 Hewlett-Packard LaserJet 414, 418

Index Hindi 148 HP see Hewlett-Packard Hungarian Revolution 258 Hungary 26, 258 Hyderabad 139–142, 151–156, 446, 449 Hyderabad Government Press 140–142, 154–156, 446, 449 hyphenation 186, 190–191, 237, 387 IBM 76, 79–81, 87, 276, 386–387, 396, 400 imagesetter 344, 383–387, 393, 398 Imprimerie Catholique de Beyrouth 225 Imprimerie Ideale 97–98 India 25, 32, 63, 72, 79, 107, 114, 139–140, 144, 148–151, 154–155, 225, 259, 264, 346, 398, 446, 449 industrialisiation 35, 68, 106, 109 InPage 352 Institut d’Études et de Recherches pour l’Arabisation 95, 100–102 Interludes (publication) 47, 155 International Photon Corporation 261–264, 267–269 Interpart 328–329, 333–336 Intertype Abridged Arabic (typeface) 120, 123–125, 188 Intertype (companies)  Harris Intertype 96–98, 103  International Typesetting Machine Company 47, 122  Intertype Corporation 47, 116, 120–122, 288, 295  Intertype Ltd 47, 155 Iran 216–217, 255–256, 285–293, 297–298, 301, 304–308, 314, 318, 448–451, 472 Iran Chap 288–290, 297–298, 301, 305, 308, 314, 318 Iraq 69, 128, 202, 353 Ireland 138 Israel 249 Istanbul 31, 48, 52, 84, 467, 470 Jalal (typeface) 256–258, 343, 396, 400, 464 Japan 205 Jawi 270–272 Jordan 128 journalism 31, 51, 55–56, 188, 203 Justape 191, 195–201, 319 justification 38, 41–43, 55, 99, 134, 172, 184–197, 201, 209, 222, 231, 237–238, 250, 256, 272, 316, 337–339, 365, 387, 407, 457,

Index 460, 468–470 Karachi 218, 222, 346–347, 371, 378 Kashida 186–190, 193–196, 199–201, 328, 331, 335–337, 365–366, 424 Kastenbein (typesetting machine) 38 Katib (software) 325, 453 kerning 54, 61–65, 137, 148, 152–154, 163, 166–168, 171, 175, 179–180, 209, 231–235, 244–247, 256, 276, 289, 294, 312, 333, 355, 358–362, 365, 394, 430, 435–436, 442, 455–457, 463–464, 468 keyboard 43, 46, 74–80, 93, 113, 118–119, 124–128, 132–135, 140–141, 156, 166, 176–177, 182–187, 193–195, 198–201, 210–211, 229–230, 237, 256, 266, 270–272, 276, 282, 302–304, 335–337, 342, 349, 368, 376, 407, 429, 436, 442–443 KITAB (Klett/Interpart Typography for Arabic Basis DP-Program) 282, 327, 330–335, 359 Klett See Ernst Klett Verlag Koenig (printing press) 36 Koran See Qurʾān K.S. Paul Ltd 235–236 Kufi (typeface) 400, 408, 460–461 Kuwait 128 Lahore 153, 213, 371 Lakhdar Naskhi 2 (typeface) 98–99 Lanston Monotype Corporation 131–132, 138 Lasercomp see Monotype Lasercomp Laserfonts (font format) 393 LaserJet see Hewlett-Packard LaserJet LaserWriter see Apple Laserwriter Layout Ltd. 409 Lebanon 1, 30–31, 49–52, 55, 88, 104, 109–113, 116, 128, 131, 202, 225–227, 234, 247, 269, 290, 330, 354, 409–412, 439–441, 472 legibility 14, 25, 61, 68–69, 83, 145, 148, 272, 309, 312, 435, 439, 469 Leiden 274, 410 Letraset 7, 100, 258, 390–391, 407 Letraset FontStudio 391 letterform 10, 13, 17–25, 61–67, 74–76, 86, 100, 111, 114, 124, 128, 142–143, 153–155, 171, 201, 222, 300–301, 310, 323, 333, 342, 363, 368, 372, 378, 411, 427–429, 452–456, 459, 468 lettering 7, 21, 89, 100, 114, 188, 244, 258, 268, 294, 311, 329, 346, 355, 364, 411–413, 419

505 Lexar 100, 103 Libya 128 Linasec 190–191 linecaster 14, 41–43, 46–47, 54–56, 59–64, 68–69, 98–99, 113, 119, 122, 148, 151, 154, 171–175, 178, 183–190, 202–203, 209, 225–226, 229–231, 272, 288–289, 294, 300, 329, 436–439, 446, 459, 466 Linocomp 96–97, 235, 238–241, 255–258, 305, 392, 439–441 Linofilm  Linofilm Quick 231–238, 277  Linofilm Super Quick 237  Linofilm V-I-P 97–99, 237–241, 244–259, 319–322, 328, 340, 358–359, 392 Linotron  Linotron 101 26, 203, 254, 352, 385  Linotron 303 133, 254–257, 306–307, 311, 314–316, 392, 448  Linotron 404 320, 340  Linotron 606 160, 187, 340–341, 358–359, 392, 398  Linotron 505c 96–97, 235–252, 255–258, 305, 308–309, 319, 328–330, 333, 439–441 Linotronic 312, 376–378, 385, 393, 450 Linotype (companies)  Linotype GmbH 342, 362, 374, 456  Linotype Ltd. 321, 358, 362, 366–368, 374–375, 378, 382, 394, 421, 454–457, 466  Linotype & Machinery Ltd. 4, 11, 47–48, 69, 99, 105–109, 112, 118–120, 192, 202, 235, 241, 303, 328, 441  Linotype-Paul Ltd. 99, 235–238, 253–260, 303, 306–308, 314, 328, 357, 361, 366, 463  Société Linotype Française 156 Linotype Bulletin, The (publication) 49, 52 Linotype Matrix (publication) 39, 57–59, 71, 103–108, 116–121, 184, 288 Linotype Ruq’ah (typeface) 352–353, 356–367, 396, 427–429 Linotype Series  2 with 3 (typeface) 71, 115, 120, 126, 129, 173, 185  8 with 9 (typeface) 124–126, 129, 290 literacy 21, 32, 70, 83, 93, 103, 110, 436 Lithography 26–28, 34, 139, 150, 153, 161–163, 173, 206, 217, 223, 231, 261, 286–287, 319, 448 Lithomat 205, 261–262 logic (computing) 190, 194–199, 237,

506 319–321, 324, 349, 372–374, 413, 443 London 27, 36, 43–45, 96, 105, 110, 132, 184, 194, 233, 258–259, 418–419 Lotus (typeface) 253–257, 394, 398–400 Ludlow Typograph 329, 354–355 Lumitype-Photon 205, 261–262 Malaysia 269–270 Malta 31 Mamluks 254 Manhattan Graphics 407 manuscript 14, 22–26, 34, 45, 63, 113, 128, 138, 149–150, 160–163, 170, 186, 286, 294, 300, 314, 370, 413, 425, 429, 438, 448, 451, 455, 458 Mariam (typeface) 400, 460–461 matrix 35–43, 48, 56–63, 67, 89, 93–95, 98–99, 103, 115–116, 119–120, 128, 134–137, 140, 143–145, 155, 158–160, 163–164, 167, 171–179, 184–186, 196, 200–201, 207–216, 220–225, 234, 237, 245, 262, 272, 288–294, 298–303, 319, 354–356, 414, 436 Matrix Case Arrangement 135, 140–141, 144–145, 155, 175–177, 225 Mergenthaler International 276 Mergenthaler Linotype Company 41–42, 45, 47–49, 51–53, 55–57, 59, 62–63, 65–66, 68–69, 71–74, 88–89, 101, 107, 111, 114, 122, 132, 150–151, 155, 201, 218, 233–235, 237–238, 276, 321, 392, 441, 444 Microsoft Arabic Typesetting (typeface) 432 Microsoft Corporation 390, 396, 400–406, 421–432, 460–463, 467–470 Microsoft-DOS 407, 421 Microsoft Office 424, 430 Microsoft Windows 425, 428, 431, 468  8 462, 467  10 428, 431  3.1 400–401, 462, 471  NT 406 Microsoft Word 425–427 Middle East Liaison Office 96, 201–203, 233–234, 253, 260, 363, 371, 444 Mitra (typeface) 255–256, 306, 311–318, 400, 450–451, 464 Mohammed V University 95 Monophoto 210–219, 222–227, 233–234, 245–247, 269, 274, 344 Monotype Corporation Ltd. 95, 126, 132, 137–140, 156–158, 166, 211, 214–216, 222–229,

Index 274, 344–347, 442 Monotype Lasercomp 344–347, 352, 368, 377–378, 383 Monotype Recorder (publication) 24, 48, 105, 131, 137–138, 206, 211–212, 215–219, 346–352 Monotype Series  485 (typeface) 139–148, 152–158, 213, 216–217, 223, 351, 437, 446–447  527 (typeface) 160, 163–169, 173, 176–178, 182, 213, 218–222, 226, 246–247, 253, 271– 272, 333, 417, 437, 440–442, 446, 460–463  567 (typeface) 170, 177–178, 181–182, 213–216, 222–223, 338, 441–442, 460  640 (typeface) 93  641 (typeface) 93  642 (typeface) 93  649 (typeface) 182, 218–223 Morocco 86, 90–103, 128, 472 morphology 10, 13, 15, 17, 19, 22, 26, 86, 324, 347–348, 363–365, 367–368, 404, 411–413, 421, 437, 452 mould 37, 43, 134, 140, 144, 148, 176–178 Mrowa-Linotype Simplified Arabic (typeface) 116, 119–121, 124–131, 192, 226, 438–441 Mughals 148–149 Multilingual Scholar 415 Mumbay 139–142 Museum of Egyptian Antiquities 254 Nadeem (typeface) 408 Nahḍa 30–32, 82 Nasḫ 2, 23, 63, 139, 152, 160, 163, 170, 265, 288, 292, 311, 333, 354, 363, 412–413, 416, 423, 426–427, 429, 446, 448 Naskh Regular (typeface) 432–433 Naskh style see Nasḫ Nasriphot (typeface) 265–270 Nastaʿlīq 2, 26, 108–109, 139, 149–154, 218, 223, 256, 285, 288, 292, 340, 346–347, 350–351, 368–378, 398, 432–434, 446–448, 451, 454 Nastaliq style see Nastaʿlīq Nazanin (typeface) 306, 316–318, 400, 444, 448–451, 464 newspapers 13, 28–39, 47–52, 55–56, 59, 63–64, 72–73, 83, 93, 107–112, 116, 119, 125–129, 149–150, 175, 183–189, 200, 203, 218, 226, 231–232, 238, 251, 259, 264–266, 270, 281, 287–290, 294, 305–308, 311–318,

Index 323, 340, 344–347, 350–358, 376–378, 405, 421, 437–439, 450, 454–455, 460  al-Ahram 33, 47, 89, 122–125, 187–197, 200–203, 235, 238, 252, 256, 264, 289, 329, 354–357, 443  al-Ǧawāʾib 31  al-Hayat 109–110, 116–118, 122, 252  al-Hoda 49–52, 55–56, 65, 74, 82, 436  Eastern Morning News 38  Ettelaʿat 235, 288–297, 301–305, 308, 315, 318, 448–450  Ḥabl al-Matīn 287  Iran 287  Jūrnāl al-Khidiw 28  Kayhan 251, 288, 306–308, 311, 314–318, 450  Le Moniteur Ottoman 29  Lisān al-Ḥāl 31  Nawa-I-Waqt 376–380  New York Tribune 41–42  Taqvim-i Veqayi 29  The Daily Jang 218, 347, 350  The New York Times 51, 83, 122, 187, 203  The Times 27, 36–38, 184  Utusan Melayu 270–273  Waqāʾiʿ Miṣriyyä 28–31 New York 41, 49–52, 55, 74–75, 87, 151 Noori Nastaliq (typeface) 352 Object Linking and Embedding 427–432 OpenType (font format) 402–406, 429, 432, 455, 462, 467–470 Oryx 399–400 Osmania University 139, 152 Osman (typeface) 243–246, 249–254, 305, 308–311, 314, 443, 464 Ottoman Empire 1–2, 24–29, 32, 50, 61, 83, 113, 170, 253, 284, 352–354, 410–413, 422, 436, 455 PageMaker 387–389, 401, 408, 415, 428–432 Pakistan 107–109, 148, 153, 213, 216–219, 223, 346–351, 357, 367–371, 376–378, 421, 455, 472 Palestine 69, 139, 155, 247, 472 pantograph 39–42, 56–58, 207, 212, 414 Paris 31, 64, 156, 202, 264, 269–272 Pashto 145, 152–154, 288–298, 301, 304 pattern (type-making) 39, 56–58, 88, 163, 208, 212, 216, 348–349, 442 penmanship 21, 455

507 pen-stroke 324–325, 421 Percy Lund Humphries 154 Persian 2, 13–15, 23, 59, 81, 91, 148–149, 154, 170, 256, 284–311, 314, 318, 366, 396, 416, 425, 448–451 personal computer 320, 385–386, 393, 396, 400, 405, 418, 421, 464 Philadelphia 50, 131–134 Photocomposition 6, 91, 96–97, 100–102, 128, 170, 180, 183, 204–219, 222–244, 251– 253, 256, 261–265, 270–273, 276–277, 282, 305–306, 310, 322–324, 330–331, 335–336, 339–346, 349–351, 357, 376, 383–384, 391, 394, 409, 443, 457–459, 463 photography (type-making) 112, 158, 205–216, 222–223, 229–238, 246–247, 262, 319, 370–371 photomechanical 229, 319–321 Photon Corporation 96–100, 103, 189, 205, 261–274 Phototypesetting see Photocomposition Photrop 261 Pivotal caster 34, 37, 42 PostScript (font format) 312, 382–401, 450, 455–456, 460–461 Press 29–30, 32, 55, 63, 286–288, 436  Arabic Press 33, 112 Princeton University 72, 151 Princeton University Press 72 PROVO (publication) 425 punch 24, 35–43, 56–58, 65, 207 punchcutter 22–24, 35–36, 41–42, 56–58, 71, 212 Qadi (typeface) 259–260, 398–400 Qahira (typeface) 408 Qajars 286–287 Qalmi (typeface) 455–458 Quark Software Inc. 409 QuarkXPress 406, 409–410 Qurʾān 22, 26, 149, 160–163, 168, 243, 286, 442, 446, 449 Rabat 91, 95, 101 Raed (typeface) 292–295 rasm 19–20, 23, 142–143, 170, 258, 470 Rawalpindi 378 Ready, Set, Go! 407 Remington  Remington Rand 87  Remington & Sons 73

508 Richard Austin 140 Roqua‘i (typeface) 355 Ruqʿä 2, 101–102, 108, 340, 352–368, 412–418, 421–423, 427, 432, 455 Ruqah style see Ruqʿä Russia 29, 36, 243, 258, 286, 448 Sanskrit 72, 139, 259 Saudi Arabia 128, 328, 331–335, 410 School of Oriental and African Studies, University of London 259 Screencheck 423 script reform 82–86, 90, 95, 101–103, 281, 437–438 Segoe (typeface) 462 Self-Spacing Type 38, 134 Seljuqs 254 Sellers & Co. 131 Seybold Report on Publishing Systems, The (publication) 320–323, 340, 385 Sheeraz (typeface) 282, 368–382, 398, 421, 454–458, 464–466 SIL International 467 simplification 74–75, 84, 87, 93, 103, 110–113, 123–129, 151, 172, 225–226, 239, 251, 257, 295–297, 300, 324, 331, 436, 439, 448 Simplified Arabic (typeface) 91–93, 104, 109–111, 116–131, 185–187, 192, 195–196, 199, 225–227, 232–233, 237–241 251–253, 265–266, 281, 285, 290, 293–295, 300, 311, 337, 438–439, 441–444, 459–463 Simplified Persian (typeface) 291–292, 295–301, 304–305 Six-Day War 202, 249 sort 22–24, 37–39, 59–61, 73, 133–134, 140, 143–145, 152, 160–162, 170–171, 179–180, 436 South Asia 149–150 Stephen Austin & Sons 277 Stereotyping 36, 272 Sudan 128, 155, 192, 202 Syria 29, 32, 49, 104, 128, 243, 249, 354 Syriac 49, 59 Tabriz 286 Tahir (typeface) 258 Tamil 138–139 Tanzimât 29 Tehran 217, 251, 287–290, 293–298, 306–308, 315 Teletypesetting 183–190, 193, 231 Telugu 139

Index TeX 327, 407 The Hague 278, 416 Tibetan 10, 76 Times New Roman (typeface) 462–463 Tradigital 335, 422 Traditional Arabic (typeface) 97–98, 173, 198–199, 240–243, 251, 425, 441, 460–463 TrueType (font format) 391, 401–402, 406, 423, 432, 462 TrueType GX (font format) 402, 406 TrueType Open (font format) 402 Tübingen 162, 259 Ṯuluṯ 170, 424 Tunis 202 Tunisia 128, 156, 265–266 Turkey 28, 48, 83–85 Type 1 (font format) 389–392, 455 Type 3 (font format) 389–392 type-casting 37–38, 107 Type Drawing Office (Linotype) 99, 234, 243, 308 Type Drawing Office (Monotype) 227, 442 typeform 7, 10, 76, 154, 186, 219, 300, 333, 339, 418, 451, 455–457, 469 type-making 1, 5–14, 20, 34–35, 46, 56–57, 69, 109, 203, 223, 241, 252, 265, 272–274, 282, 313, 318, 334–335, 339, 378, 382, 389–392, 406–412, 417–421, 425, 428–444, 451–458, 463–473 typewriter 73–81, 86–87, 91, 103, 110–111, 183–184, 276–278, 281, 413–417, 457  Brother 411–415  Olympia 76–77  Remington 73–74, 80–81, 87 typography 3–5, 8–10, 21, 24–25, 34, 45–46, 49, 54–55, 63, 82, 86, 90, 101, 106–107, 119–120, 122, 129, 133, 145, 150–151, 154, 160, 162, 168, 182–183, 196–197, 203, 206, 229, 231, 234, 255, 265, 274, 282, 285–289, 308, 320, 325, 327, 331, 333–334, 350–352, 357, 371, 389, 391–392, 404, 406–408, 410, 412, 416, 427, 430, 435, 437, 439, 443, 445–446, 450, 455, 458–460, 463–464, 466–467, 472–473 Typomatic 258 Uhertype 205 Unesco 93, 99–103, 277–278 Unicode 13–14, 400–406, 409–410, 431–432, 455, 471

Index Unified Arabic 87–89 United Kingdom 4, 11, 37–38, 43, 45, 52, 65, 69, 71–73, 76, 104, 107, 109, 131–132, 138–139, 154, 156, 158, 175, 213, 231, 233–234, 259, 275, 279, 298, 330, 349, 375, 407, 422, 444, 456 United Nations 90, 100–102, 277, 410 United States 11, 37, 40, 43, 49–50, 55–56, 69, 71–73, 75, 86–87, 90, 103–104, 112, 132, 150–151, 156, 197, 203, 205, 213, 234–235, 261–262, 265, 323, 403, 409, 417, 441, 460, 471 Unit shift 134 unit system 79, 135–137, 158–159, 185, 231, 235, 247, 315–316, 358–359, 394 University of London 259 University of Washington 324, 453 Urdu 13–14, 23, 107–108, 114, 139, 145, 148–154, 214, 218–219, 223, 346–352, 368, 377, 381, 446, 451, 454, 470 Ventura Publisher 409, 415, 463 VideoComp 323–327 Vienna 259 vocalisation 15–17, 20–24, 63, 73, 80, 84–85, 91–93, 98–99, 152, 162–165, 168–173, 176, 179–180, 216, 235, 241, 244, 249–253, 256,

509 264–268, 276, 312, 331–333, 358, 363–365, 379, 394–396, 407–408, 423–424, 435–436, 442, 464, 468 Wang Laboratories 190 Washington 51, 88, 131, 170, 284, 324, 453 Western hemisphere 6–8, 22–27, 33, 45, 49–51, 55, 69–70, 74, 82, 106, 109, 113, 154, 282, 303–305, 391, 406, 439–441, 462, 472 West Frontier Province Government Press Peshawar 154 WinSoft SA. 401, 408, 428–434 WinSoft Tasmeem 113, 415–416, 422, 425, 430–434 World War, First 33, 48, 52, 138 World War, Second 33, 48, 69, 72, 104, 111, 114, 144, 155–156, 205, 437 WYSIWYG 387 Xerox Corporation 384, 402, 409 XPress see QuarkXPress Yakout (typeface) 129–130, 253–257, 264–266, 338, 395, 398–400, 459–463 Yemen 103 Zamana Gallery London 419 Zilog Z80 413