The Challenges of Scripts in Bible Translation and Publishing Software

Extracted text from Daud_-Festschrift_Kees_de_Blois__after_proofs.pdf:

Of Translations, Revisions, Scripts and Software Contributions presented to Kees de Blois

Kees de Blois

The Challenges of Scripts in Bible Translation and Publishing Software Daud Soesilo

Preface I had the privilege of working with Kees F. de Blois for many years. Kees, who hails from the Netherlands, joined UBS after serving as a missionary linguist with the Reformed Mission League in the Netherlands Reformed Church. I think Kees has had the most job titles and designations during his tenure with UBS, ranging from Translation Consultant (TC), Regional Translation Coordinator (RETCO) for Africa, Regional Translation Coordinator (RETCO) for the Americas, Acting Regional Secretary for the Americas, Coordinator of the Translation Computer Resource Group, Interregional Consultant for Computer-Assisted Translation & Publishing Services, and most recently, UBS Interregional Consultant for Translation & Publishing Computer Services. If Kees could have a second tenure, he would have been given a new title, Global Institute for Computer Assisted Publishing (ICAP) Coordinator, a title which his successor will bear and a confirmation of the scope of the role which Kees has played so very well. I have been with Kees at various Area and Global Translation meetings and computer-related meetings at Regional and Global level, including the Open Scripture Information Standard (OSIS) initiative at the Society of Biblical Literature Annual meetings. Kees and I were co-presenters at the Bible Technologies Conference at the Pontifical University Urbaniana, Vatican City in 2002. We worked closely together in the Translation Computer Resource Group (TCRG), later called Translation & Publishing Computer Resource Group (TPCRG), and which has been modified again and is now called Global ICAP Advisory Group (GIAG). At the UBS Triennial Translation Workshop (TTW) in Iguau Falls, Brazil, in 2003, Kees and I made a joint plenary presentation on Bible revision. The Bible revision paper was then expanded with contributions from a few TC colleagues, was approved by COTP, and adopted by the UBS Global Board in 2005 as the UBS Guidelines for Revision of Scripture Translations. The subject of revision guidelines was one of the plenary topics at the TTW in Mombasa, Kenya, in 2006. Kees and I were co-convenors for the UBS Global Revision Workshop in Heliopolis, Cairo, 2007. Kees has always been coming to the Asia Pacific Area Translation Consultation (ATCON), in order to provide Paratext and Translation Management System (TMS) training to our Asia Pacific Translation team. 73

I consider Kees a gentle and trusted colleague, and a true friend, a real team player who is always willing to lend a hand. Even when we see things from different angles, he will be cool-headed and will listen attentively to a different point of view until we arrive at a meeting point. All in all, Kees is a true servant leader, who always serves with cultural sensitivity. It is to him that this paper on the challenges of scripts in Bible translation and publishing softwareparticularly in the Asia Pacific Areais dedicated. An overview of the scene The Asia Pacific Area of the Bible Society fellowship ranges from Mongolia to New Zealand and from Afghanistan to Fiji. It is the birthplace and home of major religions of the world including Judaism, Christianity, Islam, Buddhism, Hinduism, Sikhism, Shinto, Taoism, Bahai, Zoroastrianism, and the home of the three most populous nations (China, India and Indonesia). It is also the home of more than half of the worlds languages (3,579 languages or 51.8%). Due to this diversity, it is also the home of numerous complex non-Roman scripts and modified Roman scripts. These scripts have posed a formidable challenge to Bible translation and typesetting software. During 2010 the Asia Pacific Area has been involved in providing translation consultancy and computer assisted publishing services to 272 active translation projects. The five national Bible Societies that have the most translation projects are India, South Pacific, Indonesia, Myanmar and Papua New Guinea. In addition to the Roman scripts, there are six other script groupings in the Asia Pacific Area: a. Chinese, Japanese and Korean (CJK) b. Mainland Southeast Asia c. South Asia and Indic scripts d. West Asia (Arabic-based script) e. Modified Roman scripts (Pollard and Fraser scripts) f. Vertical Syriac script. On the whole, the challenge we face is two-fold. Firstly, unless the script is Unicode compliant, the translation team cannot key in the translated texts into the Paratext Bible translation software, and the project will therefore not benefit fully from the extensive manuscript checking tools and the comprehensive biblical languages and translation resources in the Paratext suite. Secondly, the translation project will not be able to benefit from Publishing Assistant in Paratext, which is extremely useful for layout and interfacing with typesetting software such as Adobe InDesign. The lack of Microsoft support to a number of these non-Roman and modified Roman scripts makes it more difficult to find a quick solution. For smaller language groups the development of script solutions is not considered to be commercially viable. However, whatever the font issues that we face, 74

each national Bible Society still has the responsibility to provide Scriptures in printed and other electronic formats for those who need them. We will now consider a number of examples of the challenges that we face with non-Roman and modified Roman scripts. Chinese Japanese Korean (CJK) Chinese and Japanese have no word breaks, all words are connected in a manner similar to the Greek uncials, and this poses a special challenge to text processing, typesetting and concordance building. Chinese (Mandarin) Used in mainland China (simplified script) and in Hong Kong and Taiwan (traditional script), as well as in the Chinese Diaspora worldwide. Number of speakers: 1.3 billion. Available translations: Chinese Union (1919), Todays Chinese Version (1979), Chinese Union with New Punctuation (1988), and Revised Chinese Union Version (2010). Text inputting: Paratext 7. Chinese is Unicode compliantthe font names are Ming, MSungHK. Typesetting: Publishing Assistant, InDesign and Adobe Creative Suite 4. At the moment the issue for Mandarin Chinese which is still unresolved is a vertical typesetting solution.

(traditional) (simplified) Pin Yin (Romanised Mandarin Chinese) It is interesting to point out that the Bible Society of Singapore is currently preparing a diglot Chinese Pin Yin (in interlinear formatsee below) and English New International Version. Available edition: Chinese Pin Yin Bible (2005 by Morrison Christian Literature). Text inputting: Hanyu Pinyin Convertor, Paratext 7. It is Unicode compliant the font name is Arial. Typesetting: Publishing Assistant, InDesign.

John 1:1 (Revised Chinese Union Version)

y b l h n de h uy

Mat 1:1

d w i de z s n

y s

j d de jip

75

Taiwanese Used in Taiwan using modified Roman script. More recently they have developed a newer writing system using Han (Mandarin) Chinese script combined with modified Roman script. Number of speakers: 17 million in Taiwan, and about 35 million elsewhere. Available edition: Todays Taiwanese New Testament (2008) Text inputting: Paratext 7. Typesetting: Publishing Assistant, InDesign. There are eight phonetic signs which have to be placed on the left hand side of the letter. This was what used to have to be keyed in manually: Chhong-se 'i-ch^eng chiu u To, To kap Si~ong-te t^ong-ch~ai, T~o s~i Si~ong-te. Recently, local IT support staff improved the method of keying in the texts in Paratext so that the team members can view them right away, that is, without having to wait for the IT staff to convert them to viewable format. John 1:1 Chhng-s -chng chi T, T kap Sing-t tng-chi, T s Sing-t. A similar challenge was also faced for the Hakka and Cou languages in Taiwan. Japanese Used in Japan and among overseas Japanese. Number of speakers: 120 million. Available translations: Meijimotoyaku (1880), Taisho Revised Version (1917), New Interconfessional Version by Japan Bible Society (1987). Japan Bible Society has just started to work on a new Japanese translation. Text inputting: Paratext 7. Japanese is Unicode compliant, the font name is MS Mincho. Typesetting: Publishing Assistant, InDesign.

Korean Used in South Korea, North Korea and among overseas Koreans. Number of speakers: 72 million. Available translations: John Ross NT (1887), Old Korean Version (1911), Korean Revised Version (1938, 1961), New Korean Revised Version (1998), Common Translation (1977), Revised New Korean Standard Version (2001, 2004), New Korean Standard Version(1993), and Catholic New Translation (2005). 76

John 1:1 (New Interconfessional Version)

Text inputting: Korean Hangul Word Processor, Paratext 7. Korean is Unicode compliant, the font name is Hangul (also known as Human). Typesetting: Publishing Assistant, InDesign. John 1:1 (New Korean Revised Version)

Mainland Southeast Asia There are four more scripts in this subregion that have no word breakThai, Burmese, Khmer and Lao. Thai Used in Thailand and among overseas Thais. Number of speakers: 60 million. Available translations: Old Thai Bible (1883), Thai Bible (1971), Todays Thai Version (1985), Revised New Testament (2004). Thailand Bible Society is currently completing a revision of the Thai Old Testament. Text inputting: Paratext 7. Thai is Unicode compliant, the font name is Garuda and other Thai fonts. Typesetting: Publishing Assistant, InDesign. Thai language specifications (including non-breaking syllables) are handled in language definition ThaiThSp and Stylesheet usfm_tbs.sty The remaining issue is in the non-Roman compositor. There appear to be compatibility problems with InDesign CS2 and CS4. The Thai TAZI project team that is currently working on Old and New Testament portions is still facing some problems in handling the conversion from the old Asia Pacific standard format markers, but they are being progressively dealt with manually. John 1:1 (Revised Version)

F F F F F Genesis 1:1 (Thai TAZI)

F

F

Urak Lawoi This literally means sea people. It is a Malay-based language using Thai script, spoken by 6000 speakers on the west coast of Thailand. Available translations: New Testament published by Thailand Bible Society (1998). The translator is currently completing the translation of the Old 77

Testament. Urak Lawoi uses the underline character to mark phrase breaks. This has to be changed to extra space in typesetting, and it is currently being done in Finalchanges.txt. John 1:1

.

_

. .

Burmese Used in Myanmar. The script is also used in other languages of Myanmar, including Karen, Kayah, Shan and Mon. Number of Burmese speakers: 35 million. Available Burmese translations: Judson Bible (1835), Garrad Version (1928), Revision of Judson Bible by McQuire (1933), Common Language Version (2005). Work is currently underway to revise the Garrad Version and to prepare the Common Language Study Bible. Text inputting: using DOS, which can be pasted into Paratext only for viewing, not editing. The Burmese DOS font name is Burma.if4, the legacy font is Burma.PFM, while the Unicode font is PadaukOT.ttf. Typesetting: VTeX and Page Maker, and gradually Publishing Assistant and InDesign. Burmese script is the second hardest in Asia Pacific area, as far as Unicode compliance and typesetting solutions are concerned. The conversion made from the legacy font to the Unicode font PadaukOT.ttf for the Judson Version and Common Language Version seems to be 95% all right, but the rest has to be corrected manually. A locally developed Burmese Syllable Segmentation program allows for typesetting with Publishing Assistant and InDesign, and they seem to yield good results, though it has not been thoroughly investigated. Based on the latest ICAP input, there are three possible ways in which the Burmese texts could be typeset in Publishing Assistant: a. using the standard InDesign non-Roman support b. using a non-Roman composer with a TrueType Font c. using a non-Roman composer with a Graphite font John 1:1 (Common Language Version) feff ggfgf g ggfg g ffghh ggfgedg efeg g g g ffghh ggfgedgedgefggfghh 78

Sgaw Karen (Sgaw Kayin) Used in Myanmar. Number of speakers: 500,000. Available translations: Mason Version (1853), Common Language Version (2003). Currently a Study Bible is being prepared. Text inputting: Microsoft Word. The font names are Sawmu and Saw Ka Nyaw. At the moment Paratext or Publishing Assistant are not being used. John 1:1 (Mason Version)

vUxH;vUoDe>OwIuvkItdO0J, 'D;wIuvkIe>O Khmer (Central Khmer) Used in Cambodia, South Vietnam, and part of Thailand. Number of speakers: 17 million. Available translations: Khmer Old Version Bible 1954, 1962; Khmer Standard Version Bible 1998, 2005. The Revised Khmer Old Version is underway as well as Khmer TAZI (adapted from Khmer Old Version). Text inputting: Paratext 7. Khmer is Unicode compliant. The font names are Khmer Hammond, Khmer Arun. Typesetting: Publishing Assistant, InDesign and Adobe Framemaker. The remaining issue is in the non-Roman compositor. At the moment the following signs are used for word breaks (~) and syllable breaks (^); these codes still need to be made invisible. John 1:1 (Khmer Standard Version)

Lao Used in Laos and Thailand. Number of speakers in Laos: 5.5 million, and in Thailand: 12 million. Available translations: Bible (2005). Text inputting: Paratext 7. Lao is Unicode compliantthe font name is Phetsarath OT. Typesetting: Publishing Assistant, InDesign. Remaining issues: Once word breaks can be marked, a wordlist can be created and the project will benefit from the spell check as well as other Paratext checking tools.

79

John 1:1

              .  

South Asia and Indic scripts There are a large number of scripts in this group. They include languages in India, Sri Lanka, Bangladesh, Pakistan, Nepal and Tibet (China). Two of the major Indic scripts are Tamil and Malayalam. Tamil Used in India, Sri Lanka and among overseas Tamils. Number of speakers in India: 60 million. Available translations: Old Version Bible (1791), Common Language New Testament (1995). Text inputting: Ileap. Typesetting: PageMaker 7. The font names are TT-VALLUVAR and TT1VALLUVAR, but they are not Unicode compliant. Remaining issues: conversion to Unicode and typesetting solutions. John 1:1 , , . Malayalam Used in India and among overseas Malayalis. Number of speakers in India: 33 million. Available translations: Old Version Bible (1841), Common Language New Testament (2000) . Text inputting and typesetting: PageMaker 7. The font names are MLTTKarthika and ML1-TTKarthika, but they are not Unicode compliant. Remaining issues: conversion to Unicode and typesetting solutions.

    .

John 1:1 ; ; Other Indian languages, including Marathi, Telegu, Kannada, Hindi, Bengali, Sanskrit (Devanagari), Oriya and Gujarati, face similar problems, namely script conversion to Unicode and typesetting solutions.

80

Marathi

Telegu OV

Kannada

Hindi

Bengali

Sanskrit (Devanagari)

Oriya

Gujarati

Nepali Used in Nepal, India and Myanmar, and among Nepalis in the Diaspora. Number of speakers in Nepal: 20 million. Available translations: Bible (1997-98, 2005, 2009). Nepal Bible Society is currently working on the revision of the Nepali New Revised Version. 81

Text inputting: Paratext. Typesetting: Publishing Assistant, InDesign. The font name is Nepali Bible, and it is Unicode compliant. John 1:1 Tibetan Used in Tibet (China), India and Nepal. Number of speakers in Nepal: 2 million. Available translation: Bible (1995-96). Text inputting: Paratext. Typesetting: Publishing Assistant, InDesign. The font name is Tibetan Youtso Paratext, and it is Unicode compliant. Jonah 1:1 West Asia The West Asia group of script includes all Arabic-based scripts such as Urdu, Farsi and Dari. The hardest one for the Asia Pacific script conversion and typesetting solution is Urdu. Urdu Used in Pakistan and India, and Urdu-speaking people in the diaspora. Number of speakers: 150 million. Available translations: Old Version Bible (1843), Revised Version Bible (1930). Pakistan Bible Society is currently revising the Urdu Version. Text inputting: Urdu Inpage 2.93. Typesetting: CorelDraw. The font name is Jameel Noori Nastaleeq, which is Unicode compliant. Remaining issues for the Urdu Protestant Bible are, among other things, correcting the placement of diacritics and a typesetting solution. Remaining issues for the Urdu Catholic Bible: script conversion and typesetting solutions. A solution for the Urdu Nastaleeq would also help solve the challenges in Pashto (Afghan), Pashto Pakistani, Saraiki, Muslim Sindhi, Hindu Sindhi, Baluchi and Punjabi. John 1:1 , ,

Farsi (Persian) Used in Iran, Afghanistan, Tajikistan, and Farsi-speaking people in the diaspora. 82

Number of speakers: 100 million. Available translations: Old Version Bible (1838), Todays Persian Version (2009). Text inputting: Paratext 7. Typesetting: Publishing Assistant and InDesign CS3. The font names are BZar and Times New Romanthese are Unicode compliant. John 1:1 .

Dari Used in Afghanistan, Pakistan and Iran, and Dari speaking people in the diaspora. Number of speakers: 20 million. Available translations: Bible (2008). A complete revision is currently underway. Text inputting: Paratext 7. Typesetting: Publishing Assistant and InDesign CS3 Middle East version. The font names are XB Zar, which is Unicode compliant. John 1:1 .

Modified Roman scripts In the Asia Pacific area, we have a range of modified Roman scripts, including Maori of New Zealand and Tokelau of Tokelau and New Zealand. These require minimal modification from Roman (that is, macrons on the vowels are the only changes required). John 1:1 (Maori) I te tmatanga te Kupu, i te Atua te Kupu, ko te Atua an te Kupu John 1: 1 (Tokelau) Nae i te kmataga te Kupu, ma nae i ei fakatahi ma te Atua, ma ko te Kupu ko te Atua Motu and Koita of Papua New Guinea use a macron on the letter g/G. John 1:1 (Motu) Matamana negana ai Hereva vada e nohova; una Hereva na Dirava ida e nohova; una Hereva na Dirava. The Yuna and Pa of Papua New Guinea use an underscore under the vowels. 83

Matthew 5:2 (Pa) Yebo:ma he: pa:ka: maka:la:la: ko:bo: hyala: ba: sa akaka:la: yo daya:ma. Some other modified Roman scripts are very complexlike the Pollard and Fraser scripts, which are used in minority languages in mainland China. Pollard script was invented by Samuel Pollard and his coworkers, including a Chinese pastor. It is now used in the Yi and Miao language groups, and each requires its own font. Samples of various languages using Pollard script are shown below. Yi Wuding-Luquan (Black Yi) Used in China. Number of speakers in China: 300,000. Available translations: New Testament (1948, China Christian Council). Text inputting: Microsoft Word. Typesetting method: MS Word converted to pdf. The font name is Black Yi, it is a legacy font, and not Unicode compliant, so at the moment the team cannot benefit from Paratext. Genesis 1:1 aK W cu c j, yM s m m k m j& ^ Dry Yi (Gan Yi) Used in China. Number of speakers in China: 100,000. Currently there are no translations available. Text inputting: Microsoft Word. Typesetting method: MS Word converted to pdf. The font name is Gan Yi, it is a legacy font, and not Unicode compliant, so at the moment the team cannot benefit from Paratext. Genesis 1:1 t c dU dUu,y so m m k d m X Yi (Awu, White Yi) Used in China. Number of speakers in China: 20,000. Currently there are no available translations. Text inputting: MicroSoft Word. Typesetting method: MS Word converted to pdf. The font name is White Yi, it is a legacy font, and not Unicode compliant. At the moment the team cannot benefit from Paratext. John 1:1 g X d b , d yi: som d& j , d X x yi: som p 84

East Lisu Used in China. Number of speakers in China: 200,000. Available translations: New Testament (2009, China Christian Council). Text inputting: MicroSoft Word. Typesetting method: MS Word converted to pdf. The font name is Lippo, it is a legacy font, and not Unicode compliant, so at the moment the team cannot benefit from Paratext or Publishing Assistant. John 1:1

Fraser script was invented by a Karen preacher from Myanmar and improved by James Fraser, a missionary to Myanmar from 1908. It is used for the Lisu languages. Lisu Used in Myanmar, China and Thailand. Number of speakers in Myanmar: 350,000. Available translations: Standard Version (1968), Common Language Version (1997). Text inputting: Microsoft Word. The font name is Tag Torah Lisu. At the moment they are not able to use Paratext or Publishing Assistant. John 1:1 (Standard Version) \a J_ \a Zch E^ Q_ @^qHe Ibn Qbo @^qHe P\ ZcqRn @_ Ea Zjm Ib D\n Qbo @^qHe P\ ZcqR U Qbo West Lisu Used in China, India, Myanmar, Thailand. Number of speakers in China: 800,000. Available translation: Bible (1986, China Christian Council) Text inputting: Paratext. The font name is Lisu, and it was created by a missionary kid who was born and grew up in Thailand. It is a legacy font, not Unicode compliant, but it works well for display in Paratext. At the moment, the ICAP team is waiting for Microsoft to provide support for the characters above the basic multilingual plane, that is, above the 65,536 code points. Until the conversion to Unicode font is possible, the team cannot benefit from Publishing Assistant.

85

Ruth 1:1 yi-sD-le bu tA bA mu xy, P: bu le jU:O m vy; nyi kwI yi w mU: kw ve; mD; Ko; do l lo= yu-t mU: bE-lE-vE kw ty su Ti ro ny yi R mLYO be yi R nyi: xU. tA ho: t siI mo-a-bD mU: kw Fo yA; mu jeO lo= Vertical Syriac Script In the country of Mongolia (also known as Outer Mongolia) Cyrillic (Russian) script is usedwhich is classified as a Roman script and is Unicode compliant (the font name is Arial Mon). The province of Inner Mongolia in China, however, uses vertical Syriac script (see the next page), which is written from left to right, and which is essentially the Uyghur scriptoriginally from the Aramaic (Syriac). It is not Unicode compliant. Mongolian (Halh) Used in Mongolia, Kyrgyzstan and Russia. Number of speakers in Mongolia: three million. Available translations: Bible (2000, 2003). A Bible revision is currently underway. Text inputting: Paratext 6. The font name is Arial Mon, it is Unicode compliant, and can benefit from Publishing Assistant. John 1:1 .

(Inner) Mongolian This language is similar to the one that is spoken in Outer Mongolia, but with a different writing system. Number of speakers in China: five million. Available translation: New Testament (1952). A new translation is currently underway. Text inputting: MicroSoft Word. It is not Unicode compliant, so the team cannot benefit from Paratext or Publishing Assistant. Conclusion I hope that this article has clearly shown the blessings and the difficulties of the diverse linguistic context of the Asia Pacific Area, and the challenges as well as the opportunities that we have because of the different non-Roman and modified Roman scripts in its living and vibrant languages. The issues that are still to be resolved range from a vertical typesetting solution for Chinese, to fixing the non-Roman compositor for a number of mainland Southeast scripts, to script conversions and typesetting solutions for 86

Burmese and related scripts, many of the Indic scripts, Urdu and related scripts, the Pollard scripts, the Fraser scripts, as well as the vertical Syriac script. In reality, the long list of remaining issues (related to the conversion from legacy to Unicode fonts and the quest for typesetting solutions) avails us with opportunities to do more research and technological cooperation with various like-minded organisations. A speedy solution is needed in order to facilitate the provision of the Scriptures in various scripts, in printed editions as well as electronic formats, including e-book and other interactive media. Without a 87

script solution, translation teams will not enjoy the full benefits of the comprehensive checking tools and resources of the excellent Paratext software, and will never benefit from the marvellous typesetting facilities in Publishing Assistant. Selected bibliography American Bible Society Library. Scriptures of the World. Online: UBS Intranet (Translations / TMS/SoW Report); http://www.worldscriptures.org/index2.html. Gordon, Raymond G. Jr., ed. Ethnologue: Languages of the World. 15th ed. Dallas: SIL International, 2005. Lam Thang, Khoi. Eagle in the Myanmar Bible. The Bible Translator 60 (2009): 195-200. Nida, Eugene A. The Book of a Thousand Tongues. Rev. ed. London: United Bible Societies, 1972. Soesilo, Daud. An Overview of Bible Translation History in Asia with Focus on the Regions of Chinese-Character Cultures, Journal of Biblical Text Research 19 (October 2006): 138-151. Soesilo, Daud. The Bible Translation in Asia-Pacific and Americas. Pages 163-181 in A History of Bible Translation. Edited by Philip C. Noss. Roma: Edizioni di Storia e Letteratura, 2007. United Bible Societies Translation Management System (TMS). Online Version.

88