Re: Linux/Mozilla related Short-Term Contract

[shimpei@example.com]

Chris Sekiya wrote:
> Hmmm ... can anyone think of a Japanization project that had its work 
committed
> into the main source tree (I'll stipulate FSF Emacs, but look how long it took
> that to happen)?

rxvt is one example. OK, it's the only example I remember offhand, but you
only asked for one. :) Various gnome programs also have Japanese programmers 
working on them, but that probably doesn't count since gnome actually had i18n 
support as one of its design goals from the beginning (as opposed to the other 
desktop project, whose idea of i18n consisted of ISO-8859-1 until pretty
recently).

But really, a lot of the blame in bad l10n by Japanese patchers can be laid
at the feet of the original programmers who conveniently assumed byte=char.
It takes a lot of effort to fix that sort of an assumption.
It also doesn't help that most non-Asian programmers assume that "everybody 
speaks English" and consider acceptable multibyte support Somebody Else's 
Problem. It helps even less that few free software developers have anything 
resembling multibyte support in their environment for debugging (I had to
do intercontinental debugging for a couple of weeks to get Japanese support
on somebody's computer working enough to see why gtk was spewing mojibake on
a particular program).

Well, whatever. The end result is still sucky Japanese support on Linux that
keeps rotting on the vine because the maintainer doesn't want to look at
them. Speaking of which, maybe tonight I'll bother to figure out why the 
Japanized version of file(1) on Kondara thinks all JPEG files are "SJIS text 
files"....

Shimpei.