Richard Ishida
Internationalization Lead,
W3C

Addison Phillips
Globalization Architect
Lab126 (Amazon)

This tutorial track provides an introduction to the topics of internationalization, localization and globalization. Attendees will understand the overall concepts and approach necessary to analyze a product for internationalization issues, develop a design or approach, and deliver a global-ready solution. The focus is on architectural approaches and general concepts, but will include specific examples and exercises.

Part I focuses on characters, character encodings, and the basics of Unicode.

Thomas Milo
Partner, DecoType

Richard Ishida
Internationalization Lead,
W3C

The tutorial will provide you with a good understanding of the many unique characteristics of non-Latin writing systems, and illustrate the problems involved in implementing such scripts in products. It does not provide detailed coding advice, but does provide the essential background information you need to understand the fundamental issues related to Unicode deployment, across a wide range of scripts. It has also proved to be an excellent orientation for newcomers to the conference, providing the background needed to assist understanding of the other talks! The tutorial goes beyond encoding issues to discuss characteristics related to input of ideographs, combining characters, context-dependent shape variation, text direction, vowel signs, ligatures, punctuation, wrapping and editing, font issues, sorting and indexing, keyboards, and more. The concepts are introduced through the use of examples from Chinese, Japanese, Korean, Arabic, Hebrew, Thai, Hindi/Tamil, Russian and Greek. While the tutorial is perfectly accessible to beginners, it has also attracted very good reviews from people at an intermediate and advanced level, due to the breadth of scripts discussed. No prior knowledge is needed.

Addison Phillips
Globalization Architect
Lab126 (Amazon)

Su Liu
AIX Globalization Architect,
IBM

Keywords: Unicode, AIX, UNIX, Code Set, Conversion, Operating System

Craig R. Cummings

Michael G McKenna  

Senior International Engineering Manager, Zynga

Tex Texin
Chief Globalization Architect, Rearden Commerce, Inc. 

Jim DeLaHunt
Principal
Jim DeLaHunt & Associates

Michael G McKenna
Craig R Cummings Senior International Engineering Manager, Zynga

This tutorial will cover the next level of detail of what Unicode is, and how it is used in the real world. The modules of the tutorial include: The Unicode standard - what are the "Guiding Lights", or design principles behind Unicode? A tour of Unicode's structure, encoding forms, behavior, technical reports, database, and how to use the Unicode Standard. Implementation according to Unicode - a walk through the details of attributes, compatibility, non-spacing characters, directionality, normalization, graphemes, complex scripts, surrogates, collation, regular expressions and other aspects according to the Unicode Standard and associated Technical Reports. Unicode and the Real World - an overview of Unicode-based internationalization development libraries and implementations supporting Unicode in web servers, application servers, browsers, C/C++, Java, PHP, SQL, and various operating systems. On-going programs - how Unicode is evolving and how you can participate in its future. Pointers to other sessions at the conference that dive deeper on particular topics are highlighted throughout.

Loïc Dufresne de Virel
Localization Strategist
Intel

Co-authors / presenters:
Michael Kuperstein, Intel, Localization Engineer
Octavio Ramos, Intel, Validation Lead

John Emmons
Senior Software Engineer
IBM

ICU is a very popular internationalization software solution. However, while it vastly simplifies the internationalization of products, there is a learning curve.

The goal of this tutorial is to help new users of ICU install and use the library. Topics include: Installation (C++ libraries, Java .jar files, Java SPI for JDK integration), verification of installation, introduction and detailed usage analysis of ICU's frameworks (normalization, formatting, calendars, collation, transliteration). The tutorial will walk through code snippets and examples to illustrate the common usage models, followed by demonstration applications and discussion of core features and conventions, advanced techniques and how to obtain further information. It is helpful if participants are familiar with Java, C and C++ programming. Issues relating to ICU4C/C++ as well as ICU4J (Java) will be discussed. After the tutorial, participants should be able to install and use ICU for solving their internationalization problems.

Presenter:

Laura Welcher
Director of Operations, The Rosetta Project

The Rosetta Project at The Long Now Foundation is working to build an open public digital collection of all human language as well as an analog backup that can last for thousands of years - The Rosetta Disk.

In the "long now," the goal is long-term storage and access to information on the scale that both supports and transcends individual human societies and civilizations. In the "here and now" the project serves to support and amplify the importance of the world's nearly 7,000 human languages, the vast majority of which are endangered and, if current trends continue, likely to go extinct in the next 100 years.

The Rosetta Project shares the Unicode vision of a world where people can use communication technology on their own terms - in their own language. According to World Internet Statistics, over 80% of all web communication is in about ten languages, with over half in either English or Chinese. The remaining 20% represent "everyone else" including about 400 languages with speaker populations above 1 million, which collectively comprise about 95% of everyone on earth. Because of essential technologies like Unicode, we are poised to see this breadth of human languages flourish online and on mobile devices, providing for these languages a critical new domain of language use in the modern world. I will present several efforts underway at The Rosetta Project including the "Language Commons" that rely on Unicode as an essential technology in building the multilingual Web.

Addison Phillips
Globalization Architect
Lab126 (Amazon)
Richard Ishida
Chair, Activity Leave
W3C Internationalization WG

In this presentation we'll explore the changes that are available today, the status of these standards, and the challenges that remain.

Software Engineer, Google Inc.

Andy Staudacher Software Engineer, Google Inc.

John Tamplin

Software Engineer, Google Inc.

Katsuhiko Momoi

Staff Test Engineer, Google Inc.

Mark Davis, Sr. Internationalization Architect, Google Inc.

In this presentation we present some innovations on pseudo locale uses within Google with particular focus on standardizing pseudo locale naming scheme based on BCP 47, defining a small set of standard pseudo locales and their exact definitions, and how they can be used for automated checks for detecting major internationalization and bidirectionality issues. To spread good practices in internationalization testing across the industry, we recently released an open source pseudolocalization Java library. Standardizing pseudo locales is essential for this purpose as well.

We begin with a brief introduction to pseudolocalization concepts, its current merits and limitations. We argue that standardizing the naming scheme (BCP 47 compliant) and defining a few standard pseudo locales offers many benefits for Google's development environments and similar ones elsewhere.

A project can host other projects and expect to see across the components exactly the same pseudo locale methods used with the standard pseudo locales: "en-psaccent" (the LTR pseudo locale) and "ar-psbidi" (the FakeBidi pseudo locale). We will demo a BidiChecker tool that can run on the FakeBidi locale.

These innovations in the use of pseudo locales allow us to detect localizability issues early via automated checks, encourage development of new tools and tests on them, and help improve internationalization quality of products.

Martin J. Dürst
Profesor
Aoyama Gakuin University
Shunsuke Oshima
Master Course Student
Aoyama Gakuin University

The Unicode Bidirectional Algorithm (UBA) is tailored for running text such as letters and newspaper articles. However, it is not suited directly for structured formats such as XML, HTML and programming languages. The source is often reordered in unpredictable ways that are unrelated to the logical structure of these formats, and therefore, source editing was essentially impossible. In this paper, we present a solution to this problem and its implementation in the text editor Emacs.

Emacs is a very flexible and extensible text editor providing an integrated environment for a wide variety of development tasks. Extensibility is based on Emacs Lisp, which we also have used for our research. While there have been some experimental implementations of bidirectional rendering in Emacs in the past, a full implementation of the UBA has only become available recently in Emacs version 24.

In earlier research, we implemented a Web-based simulation for XML and XHTML source rendering (IUC28) and a JavaScript-based experimental editor (IUC32). These implementations, however, were standalone and did not reach the level of practical usability.

The problem of bidirectional source editing for structured formats such as (X)HTML and XML can be divided into three areas. The first area is the treatment of syntactically significant characters, for example the ubiquitous angle brackets in HTML and XML. The UBA classifies them as neutral so that they follow the direction of their surrounding text, including potential mirroring, which is appropriate for running text. However, in HTML and XML they define the overall structure of the markup and therefore have to be treated as strong.

The second area is the treatment of bidirectional control characters such as LRM and RLM in source editing. Without intervention, such characters can be entered either literally, in which case they show their effect but are invisible and therefore difficult to edit, or they can be entered in escaped form (e.g. ‎ or ‏) in which case they are visible but not effective. Ideally, they would be both visible and effective.

The third area is bidirectional markup such as the 'dir' attribute or the <bdo> element in HTML, which for ease of editing should be reflected in the layout of the element content during source editing.

All the above areas can be addressed by carefully placing additional bidirectional control characters into the source text. We already had worked out much of the details of this placement in our earlier research. The main difficulty with using additional bidirectional control characters is that they are not part of the actual source text and therefore have to be distinguished from the same characters when they are part of the source, and have to be carefully removed for operations such as copying and saving. We implemented two different ways of doing this. One way uses a special Emacs property to distinguish these characters so that they can be removed before the relevant operations. The other way uses Emacs overlays, which by definition are not part of the text proper. In many ways, this would be the ideal solution, but overlays currently are not taken into account for bidirectional rendering.

Another problem is that the inserted bidirectional control characters have to be recalculated for every single editing operation. Limiting insertion to the currently visible part of the text being edited makes sure that we achieve acceptable performance even for very long source files.

We are currently extending our implementation to work with TeX, and are looking into ways to fine-tune our implementation based on user feedback.

Joel Sahleen
Software Developer, Adobe Systems Incorporated

Tak Takahashi
Globalization Engineering Manager, Teradata

Below is a list of challenges we face in software localization and translation:

• For frequently updated/released applications, we need to "migrate" translation from older versions to newer versions, for example from Teradata CRM 6.1.1.2 to 6.1.1.5. This migration process may or may not require new translation by the translation companies, but the language pack may need to be specific for that particular release. We want to automate this translation migration process.
  
• Cost reduction is always a challenge for everyone. We want to reduce the external translation costs we pay to the translation companies. When we translate software or documentation, every word sent to the translation companies will cost us, even if it was translated before.
  
• Once we have localized, for example, a product-A, we would like to share or reuse translations for other products, without involving the translation companies.
  
• We would also like to improve the localizability test and avoid any functional problems caused by inappropriate translation or localization.
  

To address these issues, we have designed and implemented a new localization process using the translation repository and its supporting tools. The translation repository was designed as a relational database that consists of tables, columns, and rows, so that we can scan tables and submit a query by SQL to search English or translated strings.

At Teradata, we now generate and maintain translation repositories for any software localization. The translation repositories enable:

• Localization migration
• In-house translation sharing
  • Reuse and sharing of translation
  • Facilitate in-house translation
• Translation using common resource file format for any software translation
• Facilitate validation and corrections of translations
• Localization problem investigation
• Reduction of external translation costs
  

In 2010, we have also enhanced our tools/process to handle DITA documentations. Now, we are capable of pre-processing DITA XML documentations, pre-translating elements using translation repositories generated from the software, and reducing the external translation costs.

Murray Sargent III
Partner Software Design Engineer Microsoft

Microsoft Corporation

The Unicode Bidi Algorithm (UBA) can display parenthesized text in strange ways, such as a)b) instead of (a)b, as described in a recent blog post*. That post describes an algorithm to fix the display of many such cases and that algorithm shipped with Microsoft Excel 2007/2010. The problematic cases have different UBA directionalities for the two parentheses of a matched pair. The algorithm essentially says that in such ambiguous cases, use the paragraph (or embedding) directionality for both parentheses. In all cases, increment the levels of text runs inside by 2 when necessary to keep the text inside the parentheses. In this presentation we describe an enhancement of the algorithm to deal with cases in which the text inside the parentheses has a single directionality. For such cases, the directionality of the parentheses is chosen to be the same as the directionality of the text within them. This refinement handles a set of anomalous cases where, for example, parenthesized English text appears in right-to-left paragraphs.

This improved algorithm displays the vast majority of parenthesized text the way one would want, but it is fair to say that no simple algorithm can handle all cases. The UBA has the LRE, RLE, PDF, LRO, RLO Bidi control characters to force particular choices. In the event that any of these Unicode Bidi control characters are used, the Bidi parentheses algorithm is not used, because the assumption is made that the user has specific choices in mind. The simpler Unicode LRM and RLM control characters can be used with the algorithm if desired.
* http://blogs.msdn.com/b/murrays/archive/2010/05/07/bidi-paragraph-with-parenthesized-text.aspx

Matt Sanford
Tech Manager of International Team,    TwitterInc.

Social networks connect you to people you already know. Twitter's model of connecting people to what's most important to them at any given time has created a cross-language and cross-cultural network unlike any other. All of this creates changes to language and communication that present a unique localization challenge. With our continued expansion in Japan, our translation of twitter.com, and our work on supporting Tweets in any language, we've learned some valuable lessons we're excited to talk about. Twitter should work in any language, even if we don't have the resources to support a full twitter.com localization, and making that happen is a large undertaking. 

Shaopeng Jia
Senior Software Engineer
Google Switzerland GmbH

This is a 201 level talk compared to the talk I gave last year. It provides a more in-depth look at international phone number handling, and places focus on the new development of the library in the past year.

Roozbeh Pournader
Internationalization Engineer, Google

Norbert Runge
Test Engineer,
Google, Inc. 

For example, some pages are double-converted to UTF-8, turning 'ü' into 'ü'. Many pages have the wrong encoding declared, others contain a mix of encodings. I will show details for this and other examples.

Mark Davis
Staff Test Engineer & other titles, Google, Inc

Luke Swartz

Product Manager, Google

Thomas Milo
DecoType

Luke Swartz
Product Manager, Google

- Core i18n Libraries: Giving web applications access to internationalization libraries on par with those available to desktop applications.
- Resources: Allowing web applications to easily package and manage strings and other international resources.
- Localization: Making it easy to translate web applications' strings and other resources.

In this talk, we will explore each of these challenges, and show some potential solutions, which will help web applications become fully international.

Helena S Chapman
Program Director, IBM

We will also take time to examine where the data interoperability gaps are in an end-to-end content/data life cycle of a localization request. What roles can Unicode Consortium play in helping mature and drive the most appropriate adoption of its existing and future standards and assets. Most importantly, as a Unicode standard supporter, what you need to be aware in contributing to drive these standards to benefit the localization operation of your organizations directly or for your clients.

Adil Allawi
Technical Director,
Diwan Software Limited

For the past five years I have been the Iraq correspondent to Global Voices and had the chance to work closely with those on the leading edge of the Arabic social media revolution. In this presentation I will seek to tie this experience with the technology that underlies it.

I will review the history of the Arabic language on the web, its current problems and successes and make a comparison between the uses of social networking sites in Arabic and their equivalent uses in mono-directional languages.For the presentation, I also hope to include feedback from active users of social networks on how they use this medium, what are their frustrations and successes on their chosen devices.

Finally I will conclude with a review of how Global Voices reports on social networks in multiple languages and make recommendations on how social web sites can improve Arabic support for their users and those reporting the conversations.

Andrew Swerdlow

Co- Author:
Nav Jagpal
Technical Program Manager, Google

Michael S. Kaplan
Program Manager,
Microsoft Corporation

Hangul has had a long history from the 1446 document that first described the underlying Jamo to the latest Jamo additions to the Unicode Standard. This presentation will do a whirlwind and only mildly irreverent tour of that history in the form of a presentation to Sejong to explain what has happened, highlighting the use, encoding, and re-encoding of one of the more perfect alphabets, imperfectly handled, in this or any other age.

Behnam Esfahbod
Yahya Tabesh
Sharif University of Technology

Perso-Arabic script, the second-most used writing system in the world, has many unique properties which have made its computation harder than some other scripts. In this paper we introduce an abstract model for the typography of Perso-Arabic script which exhibits the hidden properties of the script and makes the typographical computation of Perso-Arabic text possible. This properties have been ignored in most of the recent works on Perso-Arabic script, specially in the international standards.

Each letter in Perso-Arabic script, in any of its cursive forms, is constructed from a Base Shape and some of the them are accompanied with some Auxiliary Shapes (for example: Dots, super-script and sub-script Alefs, Madda, Hamza). These Auxiliary Shapes may appear on the top of the Base Shape, on the bottom, and sometimes on the head or the tail. Also, some other Auxiliary Shapes may accompany each letter as a separate Unicode character (for example: Harakats, Shadda, Sukun) which also will be positioned above or below of the Base Shape.

Base Shapes and Auxiliary Shapes work as the building blocks of the Perso-Arabic script. These visual properties are consistent in all of the writing styles of Perso-Arabic script (like Naskh, Thulth, Nasta'liq, and Tahriri). In this Abstract Model, we have encoded these properties such that a series of Shapes can be computed for any Unicode string.

Also, we introduce a metric distance based on our Abstract Model, the Shape Distance, for strings of Unicode characters. This metric makes it possible to compare Perso-Arabic strings based on their actual appearance, regardless of what writing system or font is used. The Shape Distance works such that strings with very similar appearance would have a distance close to zero, and big difference in the appearance results to large numbers. For example, two words with similar letters which only differ in one auxiliary part (like a dot) would have less distance than strings with letters that have different Base Shapes.

And finally, we have studied two Persian text corpora based on the properties of the words, the letters, and the Base Shapes and Auxiliary Shapes. We show that the distribution of Base Shapes and Auxiliary Shapes follows the same the pattern as distribution of the letters.

The first important application for our model is the security of Perso-Arabic domain names. By the introduction of Internationalized Domain Names (IDNs) and looking at the future of the internationalized Internet, security of domain names (at both the TLD level and the registry level) has became a serious concern of ICANN and various ccTLD ang gTLD registries. The Abstract Model and the Shape Distant algorithms can be used to calculate the similarity of Perso-Arabic domain names with very high accuracy.

The second application for our model is the font industry, it can be used in font generation, alteration and verification programs. Using this model, some parts of the glyph generation, classification, and table generation in Perso-Arabic fonts can be automated.

Another application of this model is the font rendering engines which at the moment depend on ArabicShaping table in UCD and the tables provided in font tables. The character data provided in ArabicShaping table is incomplete for some use cases, and the font tables cannot always be trusted to be complete or accurate. This model can help in better Perso-Arabic text rendering and increasing the stability of font rendering engines for this script.

## Notes

1. "Typography" here means "the general character or appearance of printed matter."

2. Perso-Arabic script is called "Arabic" in Unicode standard. Also some other names has been introduced, like "Arabetics". In this paper we use "Perso-Arabic" as the name of the script to distinguish it from the Arabic language.

Andrew Glass
Program Manager, Microsoft Corporation  

Su Liu
IBM

Keywords: Unicode Data, Multiple Versions, Network, Conversion

Pim Rietbroek
John Hudson
Brill / Tiro Typeworks

A scholarly publisher needs to be able to publish any text in any European language from any period, and to do so within traditional canons of typographic quality and sophistication, as expected by their authors and readers. This presentation describes the demands this makes on the publisher, and on editors and authors, and looks at how these demands are met through careful specification of requirements, standardisation on Unicode text encoding, and development of extensive and typographically sophisticated OpenType fonts.

Pim Rietbroek (Brill) and John Hudson (Tiro Typeworks) present an overview and selection of short case studies from their five-year project, illustrating some of the challenges encountered in texts ranging from transliterated ancient Egyptian and Sumero-Akkadian to linguistic descriptions of present-day endangered languages, with stops among ancient Greek acrophonic numerals and the not-always-helpful Unicode encoding unifications and disunifications.

Michael S. Kaplan
Program Manager, Microsoft 

Leandro Reis
Senior Program Manager, Globalization. Adobe Systems

Mike McKenna Senior International Engineering Manager, Zynga

Paul-Henri Arnaud
Senior Process Analyst, Autodesk

Peter Constable

Senior Program Manager, Microsoft Corp.

The next version of Microsoft Windows is in development and promises to provide great new opportunities for the Windows developer ecosystem. In addition to introducing new app development paradigms, Windows 8 also adds a lot of new multilingual and globalization functionality. This talk will provide an overview of additional functionality that developers can leverage to develop world-ready desktop or Metro-style apps for Windows.

Peter Constable
Senior Program Manager, Microsoft Corp.

Steven R. Loomis
Software Engineer, IBM

Adil Allawi
Technical Director,Diwan Software Limited 

Over the following 25 years the font has been rebuilt several times over, extended to the full Arabic Unicode range, ported to every kind of device. The latest incarnation is now a web font for every major browser. The design needed to be relevant for user interfaces, printing and small screens yet still appear compatible with Roman fonts. Support for the full Unicode Arabic range, meant adding an extra 1500 glyphs together with the relevant tables for Arabic shaping, ligatures, justification and kerning.

Along the way I will explain the features need for a modern typeface to be useful in a world where data may be exchanged across different standards; How to approach the development of user interface fonts and the new standards for web fonts. I will discuss the importance of embedding semantic information into a font to allow unique identification of its glyphs and allow equivalence to be found in other fonts.

The presentation will conclude with a discussion about the future for international typefaces as they make their way into open standards and become part of the content of the worldwide web.

Doug Felt
Google

Steven Loomis
Software Engineer
IBM

.

Markus Scherer
Unicode Software Engineer, Google Inc.

Mark Davis
Sr.Internationalization Architect,
Google Inc.

ICU has been improving on the Java formatting framework, adding support for such message variants in both its C++ and Java versions. In addition, other aspects of message formatting have been simplified. This session explains the challenges, approaches, and new functions and capabilities.

Co-authors: Markus Scherer & Mark Davis

Anshuman Pandey                Ph.D. CandidateUniversity of Michigan

This presentation will discuss the role of Unicode in the development of minority and endangered languages in South Asia, with particular focus on India. The talk will begin with an update on the support for South Asian scripts in Unicode 6.0 and an analysis of how well the Standard covers these scripts today. It will then present five current character-encoding projects for writing systems used by minority language communities in India and Nepal and discuss the conditions of these languages, the linguistic requirements of these communities, and the potential that Unicode offers for increasing education and literacy using the language. The talk will then describe the role of Unicode in the Government of India's National e-Governance Plan (NeGP) and how the plan has the potential to facilitate institutional and governmental support for endangered and minority languages. The talk will close with a presentation of several case studies of writing systems being created today in India by minority language communities and what such ongoing activity will mean for Unicode. The presentation will conclude by discussing the current projects of the University of California - Berkeley's Script Encoding Initiative, which is working with the user communities to propose these scripts for encoding into Unicode.

Michael Kuperstein
Localization Engineer, Intel

Martin Jansche
Staff Software Engineer, Google Inc.

Speech Internationalization at Google

Pedro J. Moreno, Linne Ha, and Martin Jansche*
Google, Inc.  (*corresponding author)

Internationalization (i18n) and localization of software that processes spoken input or output is faced with challenges that differ from those found in many other software i18n projects. In this talk we describe our experience with internationalizing Google Voice Search, a speech-to-search service available on many popular mobile devices. First launched for US English in November 2008, Google Voice Search is currently available in more than 20 languages.

Our lack of precision in the number of available languages illustrates one the main challenges of this project: working definitions of "language" and "dialect" vary depending on context. Focusing purely on technical aspects, our notion of language is driven by the current limitations of speech recognition technology. For example, Google has developed speech recognition models for several variants of spoken English, including American, British, Indian, Australian, and South African English. But our support for English does not end here. Many other languages have been influenced by English: Our Cantonese recognizer was explicitly designed to deal with the many English loanwords that are in daily use in Hong Kong. The languages of Europe all borrow English words liberally, despite the efforts of local language academies.

This kind of linguistic diversity makes the challenge we face even harder, since it increases the number of language projects we have to deal with. This is compounded by the fact that internationalizing speech technology is hard because bringing up a recognizer in each new language is a separate development effort, requiring significant amounts of data, compute power, and engineering time. We are fortunate to work in an environment where enormous compute power is available and can be easily harnessed. Of the other two factors, we will assume that engineering time is always limited and at best, we can reduce the accidental complexity of the very complex task of building recognition models. Since our development approach is heavily data-driven, most of our recent progress has been in the area of data acquisition, where significant changes have been made to the tools and processes which we use to acquire linguistic data as needed for building recognizers.

We generally need three kinds of data: spoken utterances plus textual transcriptions; pronunciation dictionaries; and large amounts of text. We have crowd-sourced several aspects of our data collection efforts. For example, we generally need spoken examples for a given target language, from a variety of speakers, and under varying environmental conditions. We have built tools that make it easy to collect such data from volunteers. We have also made process changes to allow us to manage the data collections remotely while monitoring progress and assessing data quality. This has allowed us to collect acoustic data many times faster than before.

The collection of pronunciation information cannot be crowd-sourced as easily since the current process requires a certain amount of linguistic expertise. Volunteers are asked to transcribe words into crude phonetic representations. For some languages, e.g. Spanish, the pronunciation of a words is readily apparent from its orthography and can be expressed algorithmically. In those cases we use ICU transforms to give us word pronunciations, perhaps combined with a brief pronunciation dictionaries of exceptional or foreign words.

While text resources are often easily available, this does not hold for all languages. For example, Voice Search is currently available in South Africa in English, Afrikaans, and Zulu. For Zulu, the amount of text available in electronic form is considerably more limited than for the other languages. We'll share our thoughts on what could be done to encourage the creation and dissemination of data for under-resourced languages.

Martin Raymond
Script Information Engineer/Editor, SIL International

This session will cover some of the needs ScriptSource has been designed to meet, as well as the challenges encountered in bringing together information from different sources and creating the data associations to make it as meaningful as possible. There will be a demonstration to show how easy the ScriptSource User Interface is to navigate and to illustrate the main functions, including adding information about a script. The use ScriptSource makes of CLDR data will also be demonstrated, and the plans for a more extensive interface with the CLDR will be discussed.

Mark Davis
Markus Scherer 
Google

ICU uses a number of innovative algorithms and data structures to handle internationalization, balancing tradeoffs between performance and data footprint. This presentation covers some of the more interesting of these structures, and their applicability beyond internationalization. The discussion includes: the new trie structure for string lookup in ICU, effective use of inversion lists and inversion maps, compact character mapping tables, transliteration mappings, and others. Code that supplies and uses these structures is part of ICU, the Java/C/C++ open-source Unicode enablement library.

Co-authors: Markus Scherer & Mark Davis

Ken Lunde
Senion Computer Scientist, Adobe Systems Incorporated

Today, one can easily argue that for a single font to adequately serve multiple CJK locales, it must include more than one glyph per CJK Unified Ideograph code point. Such fonts are referred to as Pan-CJK fonts, because they serve the needs of more than one CJK locale. I have predicted that at some point, years or decades into the future, cross-cultural interaction will evolve into initiatives whose aim is to genuinely unify CJK Unified Ideographs across all CJK locales. This is likely to have the effect of making a single glyph acceptable for all CJK Unified Ideographs. The Chinese standard designated GB 18030 is actually a step in this direction, mainly because it specifies a single glyph for each CJK Unified Ideograph code point.

This presentation will explore the history and development of ideographs and Han Unification, and draw conclusions based on the presenter's own experience developing CJK fonts and working with CJK character set standards.

Roy Boney, Jr.
Language Technologist, Cherokee Nation

Joseph Erb

Educational digital media specialist, Language Technology Program at Cherokee Nation Education Services Group

Jeff Edwards

Language Technologist,

Cherokee Nation

One problem of the digital globalization of communication is that it has the potential to erode the already endangered cultures and languages of indigenous peoples. These communications rarely occur in the indigenous languages of minority cultures, if at all. It does not need to be this way, and proper adoption of the technology by the community is paramount for success.

This presentation will discuss the efforts undertaken by the Cherokee community to address these problems. It will discuss the adoption of the Cherokee syllabary into Unicode, the usage of the Cherokee syllabary in modern computing systems, the mobility of the Cherokee language in various hardware platforms and social web presences, and the adoption of Cherokee language technology by the Cherokee community.

The presentation will be by Roy Boney, Jr. and Joseph Erb, Language Technologists of Cherokee Nation Education Services Group.

Stuart Gill
Member of Technical Staff, Google Inc.

Peter Edberg
Senior Software Engineer, Apple Inc.

Markus Scherer
Unicode Software Engineer, Google Inc

This presentation will provide a brief overview of ICU, with emphasis on the recent updates in ICU 4.8, including the latest support for Unicode 6.0 and CLDR 2.0, collation reordering for better customization and reduced collation data, plural and gender in messages, and other changes (see http://icu-project.org/download/4.8.html). The presentation will also touch on ICU's planned direction for 5.0 and future releases.

Zhang Zhoucai
CEO, Beijing UniHan Digital Tech Co.Ltd 

Zhang Chiyi
COO, Beijing UniHan Digital Technology Co., Ltd.

As Unicode/CJK standard developer and Unicode implementer in Chinese e-publishing industry, the author summarized their experiences and lessons in Unicode based e-publishing, and points out that, besides the ordinary requirements for general content provider on the web-cloud, Chinese Ancient Classics have some special requirements for Cloud-Publishing for better reading and usage experiences, which include but are not limited to

• super-CJK Font on Web-Cloud and its alternative solution,
• super-CJK Hanzi Handwriting Recognizer on Web-Cloud ,
• a Dictionary of Simplified-Traditional-Variant Hanzi on the web, and especially,
• an on-the-web OLD TERM CHECKER would be very much desired in order to help reader to judge and understand strange Chinese term's attributes - a dynasty year, a name of an officer, a name of a place, or a name of a person?

In addition, Author will introduce their implementation of web-cloud based handwriting recognizer (UniHan Q-Pen) and the above mentioned TERM CHECKER built in their UniHan Classics data base on the web.

Co-Author:
Zhang, Chiyi, COO,
Beijing UniHan Digital Technology Co., Ltd

Tex Texin
Chief Globalization Architect, Rearden Commerce, Inc.

This presentation recommends specific data values that are likely to identify internationalization problems in software intended for global markets.

Based on years of global software experience, these data values are useful in functional or linguistic QA tests of internationalized software. The data value recommendations include character encoding, postal address, locale and other data types typically used in software and will assist in finding common internationalization problems. This presentation will offer specific test suggestions.

Program is subject to change.