Raph Levien <raph@acm.org>
11 Jul 1999
levien.com Gnome homeThis page keeps track of the effort to bring high quality text processing to Gnome.
I'm reworking the Gnome-Print interfaces to have nice font and text support. The new draft interfaces include:
Owen is working on a gscript interface for Gtk+. We are trying to get these unified as much as possible.
www.czyborra.com is an excellent resource for learning more about Unicode, glyphs vs. characters, internationalization issues in general, and alternative sexuality.
To see how Microsoft does it, check out this overview and this reference doc.
To see how Apple does it, check out the ATSUI Reference.
The Open Group has some information about editing in Thai and other complex scripts. The standard sucks, but the spec has some nice material in it.
The Omega Project has somewhat similar goals within the context of TeX. Thanks to Manne Miettinen for hte link!
ConScript contains many scripts invented in modern times. It will be interesting to support at least some of these as well. Nuq Nekh!
Specific scripts
Arabic
The Art of Arabic Calligraphy by Mamoun Sakkal is a beautiful and informative introduction to Arabic script.
CJKV
Ken Lunde's book, CJKV Information Processing is the definitive resource for all texts Chinese, Japanese, Korean, and Vietnamese. If only a book of this quality were available for the other scripts of the world, the task of creating a global text system would be much easier.
Malayalam
The Monotype WorldType has a description of Malayalam, including font sample and glyph repertoire.
Misc
The Merriam-Webster online dictionary has a nice table of 5 non-roman alphabets, covering Hebrew, Arabic, Greek, Russian (Cyrillic), and Sanskrit (Devanagari).
The Monotype WorldType catalog is a very rich resource, covering a great many scripts. Unfortunately, these pages contain only textual descriptions, font samples, and glyph repertoires; scholarly information about the alphabets is missing.
Help Wanted
Our goal with Gnome Text is to provide easy input and high quality rendering for all scripts of the world. Of course, this is a monumental task, and there is no way we can do this alone.
Fortunately, Gnome is a free software project, and thus very well suited to the kind of large-scale collaboration that true global text processing requires. If you have expertise with an unusual script, you can help in a number of ways.
First, send us links to scholarly resources for non-Latin scripts. We have the Unicode book and know about basic text formatting, but are more or less lost when it comes to the fine points of justifying Arabic text with tatweel or which Devanagari ligatures to use when. As mentioned above, Lunde's CJKV book is the definitive resource for CJKV scripts.
Second, we need fonts. Roman Czyborra's Unifont project provides us with low quality placeholder glyphs for a great many scripts, but what we need are free, high quality fonts. These fonts need to be GNU-licensable so that we can reencode, repackage, and refine them. I'm converging on Adobe Type1 format for glyph shapes and hinting, but fortunately most other formats can be converted to Type1 fairly easily (although the hinting does need to be redone).
Third, we will need coders to help with language-specific and script-specific modules. A good example here is Dov Rosenfeld, whose FriBidi library for the Unicode bidirectional algorithm is proving invaluable.
Fourth, we need examples of superb typography. My goal for the output of GnomePrint is to match the highest quality book printing being done today, not merely a legible reproduction of the characters. Recommendations of books with top-notch typography are quite welcome, as are <shamelessbegging>copies of the books themselves</shamelessbegging>. I can't find a copy of "Typography of Devanagari" by Bapurao S. Naik (referenced in the Unicode book). This sounds like it would be quite useful!
Finally, we will need good eyes to review our work. It's entirely plausible that our early efforts look fine to American eyes, but just somehow wrong to people deeply steeped in the language and culture. We need people who can criticize and refine these results.
I feel that this is one of the more exciting projects in the free software community today. A band of collaborators with the right skills and chemistry can make it happen!
If you're interested in helping, a good place to start might be the gtk-i18n list.