Mozilla Languages Enabling Projects Status Report
Contact: Frank Tang <ftang@netscape.com>Discussion: mozilla-i18n@mozilla.org or netscape.public.mozilla.i18n
Last Modified Date: April 30, 1998
This page summarize the activities and issues for the current Mozilla national language enabling/enhancement work. We gather open source developers to help us to add/enhance a particular writing system/language support for Mozilla.
Beside the work we list here, there are also some people working on translating Mozilla User Interfaces into different languages. You could their activity at Mozilla localization projects page.
Highlight of This Week
Discussion
-
Customize Encoding Menu
Serge Orlov and Jungshik Shin suggest we do not display the item if the necessary font support is not available. ftang concern about the case for dynamic font. Alexey Pushkin suggest we group items into sub menu by language group. Erik van der Poel concern about sub menu idea will necessary time of clicking and suggest we do customized encoding menu support like VC++ did. (I wonder do they, or Wang, have a patent for that ... ) Leong Kok Young agree with Erik. No conclusion yet. Need more discussion. Need design specification. Need prototype to prove concept. Need XP/FE implementation.
-
Contextual Language Support
Samphan Raruenrom, Sandeep and Namita Sibal, Jungshik Shin, Jon Babcock <jon@knaji.com> discuss the need for Contextual Text layout support for Thai, Indian languages, Korean, and (even) Chinese. A lot of presentation for different language (which is good). Shree suggest we take a look at IITM Multilingual Software and see whether we can use their API instead of redesign it. However ... Need people to lead the investigation. Need more discussion. Need someone start to prototype for one language.
-
Use Process Code in Layout
Hiro Yoshioka suggest us to measure the performance of using process code in layout. ftang and Hovik Melikyan give some feedback. No conclusion and real performance measure information yet.
-
Enhance Mozilla Cyrillic support?
Great accomplishment here this week!!! Alexey Pushkin report cp1251 does not work on UNIX which lead ftang to fix a conversion rule bug. Serge Orlov send out change according to Alexey Pushkin's "decide WinCharSetID depend on font availability" idea. Great Job, Cyrillic Folks !!!
-
Public Specification for Hebrew Enabling
JKL < jklnet@usa.net> and Jonathan Rosenne discuss the availability of public Hebrew specification.
Patches
- [April 26, 1998] Armenian WinFE Support (by hovik@moon.yerphi.am) - ftang is currently reviewing it. Need WinFE owner's approval. Hovik is working on the UNIX change. We need MacOS expert to help.
- [April 24-27, 1998] XFE Cyrillic conversion bug (by sorlov@con.mcst.ru and ftang@netscape.com) - fixed and checked in to 1998-04-29 Mozilla source
- [April 28, 1998] Thai charset support, esc. Win32 (by samphan@email.ksc.net) - ftang have reviewed and commented on it. Samphan is working on a new patch. Probably need help to build on UNIX, OS/2 and Macintosh.
- [April 29, 1998] XFE Automatic Cyrillic Font Charset Selection (by sorlov@con.mcst.ru and ftang@netsape.com) - waiting for your review/QA before check in.
- [April 30, 1998] XFE Locale Independent FORM button/selection handing (by ftang@netscape.com) - waiting for XFE owner's approval and your review/QA before check in.
Documentation
[Nothing This week, hope we can have something next week.]Misc.
- Need Help to verify FAST_MULTI turned on layout code -
We plan to turn on the layout performance work in the near future. But
we need your help to verify the behavior for multibyte charset page.
Read ftang's posting for details:
I would like to ask you particular to try this for us because you are experts on CJK issues. I appreciate your help. Other folks on the list, welcome to join to help us.
I want to ask our layout group to turn on the layout optimization work they did for CJK (currently turn off as default). But I need you guys help me check the quality.
How To Do That ?
1. Build mozilla if you have not.
2. Open lib/layout/laytext.c . Look at FAST_MULTI . Delete the surrounding comment so it will be defined. Look at -> 32 Change it from
31 /* 32 * Turn this define on to get the new multibyte parsing code 33 #define FAST_MULTI 34 */
To31 /* 32 * Turn this define on to get the new multibyte parsing code */ 33 #define FAST_MULTI 34 /* */
3. Build again, it will just compile that file and link. Should not take a long time.
What Should Be Check?
Any Multibyte Charset should layout exactly as the one do not turn this on. Please do not try to IMPROVE the way it is doing right now. Let's do that after we turn on the code path. It should have no affect on single byte charset behavior.
- Line Breaking
- Kinsoku Rule
Please let me know if you start to try it and also tell me your result [better come with some URL we could test. ] when you finish it. Thanks. I want to ask the layout group to turn it on by default by next Wed.
Frank Tang
Index by Languages
- Armenian Language Enabling
- Georgian Language Enabling
- Bi-directional Language Enabling: Arabic and Hebrew
- Indian Language Enabling
- Thai Language Enabling
- Korean Language Enhancement
- Chinese/Japanese/Korean Language Enhancement
- East European Languages Enhancement
- Cyrillic Languages Enhancement
- Armenian Language Enabling
Hovik Melikyan <hovik@undp.am> is trying to add ArmSCII-8 and ArmSCII-8A (the modification adjusted to MacOS) support for the mozilla source code. You can find more information about Armenian charset standard from here. He is also working on an English document describing Armenian character set, encoding, classification, sorting, etc. Hovik Melikyan describe himself as C/C++ UNIX/windows programmer. He may need your help in the Macintosh and OS/2 platform. He have a patch for Mozilla Armenian (Window only) support which could be found at http://www.freenet.am/Mozilla/ before we merge into Mozilla source. We need someone to help him to add similar change for Macintosh and UNIX now. Please contact us immediately if you could do this for us. Thanks.
- Georgian Language Enabling
Tomas Martin Widmann <viralbus@saratoga.daimi.aau.dk> said later he will help us doing non-Unicode base Georgian language enabling work.
- Bi-directional Languages Enabling: Arabic and Hebrew
Franck Portaneri <franck@langbox.com> volunteer to help the bi-directional text enabling work (with his company's encouragement!!!). He is familiar with Arabic CTL, X window and Motif. Join him if you are a good at Macintosh and Window. He is planning to work on a SGI IRIX 6.2 or 6.3 for this work.
Anoosh Hosseini <anoosh@gpg.com> (Author of PMosica) said he is very busy these days and cannot join the development activity. But he is welling to monitor the discussion and provide some feedback.
Craig Olinsky <olinskyc@intel.co.jpo> said he is willing to test the bi-di feature.
Yehavi Bourvine <YEHAVI@vms.huji.ac.il> said he will try to find someone to help us for Hebrew support.
Doron Shikmoni <doron@isoc.org.il> said he is willing to join the discussion.
Jonathan Rosenne <rosenne@netvision.net.il> list the following specification as Hebrew reference:
- RFC 2070- Internationalization of the Hypertext Markup Language
- HTML 4.0 Specification
- Middle Eastern Language Issues
- Indian Languages Enabling
Sandeep and Namita Sibal <sibal@sibal.com> is familiar with UNIX platform and is gathering people to form a Indian languages enable project for Mozilla source code. He will start with Linux build and target on Devanagari first. He is interested about how Thai and Arabic people solve their problem. He may put together a site focus on this Indian Mozilla project and we will make a link to his site.
Currently we have collect the following on-line information for Indian language support:
- Macintosh Indian Language Kit- Quick Reference Card (in PDF format)
-
Macintosh Indian Language Kit- User's Manual (in PDF format)
- Indian Institute of Technology, Madras- Multilingual User Interface Software referred by Shree <shree@usa.net>
- Portable Layout Services: Context-dependent and Directional Text (Open Group CAE Specification) by the Open Group
- Thai Language Enabling
Samphan Raruenrom <samphan@ksc.th.com> is trying to enable Thai on Win32 environment. Before he look at the mozilla source code, he already have a independent Thai support library and is considering to make it public available. He is familiar with Win32 on Windows95 and VC++ 5.0. He need someone to help him for X and Macintosh. He is currently study the layout line breaking code and try to figure out how to plug his Thai line breaking code into it.
Craig Olinsky <olinskyc@intel.co.jpo> said he is willing to test the Thai feature.
(Draft) Tasks List: (by Samphan Raruenrom)
-
Thai charset support :
- auto-select possible?
- default cross platform font family? e.g. DB ThaiText?
- handle font size incompatibility, e.g. too-small, due to implicit assumption?
- use font encoded in different charset, e.g. THAI_CHARSET, ANSI_CHARSET.
- full Unicode support?
- promote use of encoding Meta tag.
-
Thai line wrapping :
- need work on Layout.
- cross platform solution.
- prepare for NGLayout now.
-
front end :
- Win32.
- Macintosh.
- X.
- Win16? Win32s? possible?
-
parser :
- perform sequence-correction on load.
- handle character references mapping unsupported by ISO-8859-11, but supported by platform font charset, e.g. (c) (r) tm available in Mac?
-
caret position - cursor movement, selection :
- character left, right : base on cell.
- word left, right : base on Thai word break.
- double-clicked word selection.
- context-sensitive keyboard language switching.
-
editor :
- character insert : do sequence-check, reject illegal.
- character delete : one cell at a time.
- character backspace : one character at a time (not cell).
-
spelling-check :
- need better spelling-list.
- multi-lingual : at least US/Br English, use existing code.
-
other Thai locale support :
- sorting : using Thai collating order.
- formatting?
-
low-level - especially for non-Thai GUIS (MS, MAC, X) :
- glyph shaping : document charset to glyph charset transform, should work for both display and print.
- keyboard map : already available for all GUIS, add native support, e.g. automatic switching.
- localized user interface?
-
Thai charset support :
- Korean Language Enhancement
Jungshik Shin <jshin@pantheon.yale.edu> (Author of Hangul and Internet in Korea FAQ ) said he is not good at programming but he will try to put together a Korean enhancement wish list. The current wish list is :
- Korean IME support in non-Korean MS-Windows version
- Korean Mail/News encoding change.
- JOHAB encoding/font support in X window version to support Unicode 2.0
- X Window Locale independent form handling
- Display European character correctly in Korean MS-Window system
- Combining jamos support in Unicode
- CJK Languages Enhancement
Craig Olinsky <olinskyc@intel.co.jpo> said he is interested in input methods and non-latin script issues. He is familiar with windows, C++/MFC but mostly in Win32 SDK. He suggest we work on Furigana ("ruby") support, and "Platform Independent IME support". He probably will spend some time to help us fix some Japanese specific bugs.
Leong Kok Yong <kokyong@irdu.nus.edu.sg> in National University of Singaport is trying to found a project with several students to enhance CJK support for Mozilla. They are good in Input Method support. Let them know if you have some good idea.
- East European Languages Enhancement
Tomas Brodsky <brodsky@cfar.umd.edu> already submit some fix for Mozilla X window Latin 2 support. I think he could help us to test and enhance Mozilla single byte support on UNIX version. He is currently look at Mozilla X window form handling.
- Cyrillic Languages Enhancement
Andrey A. Chernov <ache@nagual.pp.ru> (Author of RFC 1489 "Registration of a Cyrillic Character Set") is working on KOI8-R enhancement. He already send some patch to us to add KOI8-R locale name mapping for FreeBSD.
Serge Orlov <sorlov@con.mcst.ru> is currently try to improve our Cyrillic support and send us several patches already.
Dima V. Fedorov <fedorov@cpmet.ufpel.tche.br> said he want to contribute and waiting for ideas.
Linda Valdez [Need to find her email address] said she want to help Cyrillic enhancement.