On Wed, 9 Jul 2003, Wu Yongwei wrote:
> (excluding the desktop, which I prefer KDE). But I did have some bad
> experience with Windows Gtk applications running on Chinese versions of
> Windows. Not for functionality, but for UI. You are right that they do
> care about Asian languages, but the problem seems that they may not have the
> hands to test on Asian language platforms. At least not on Simplified
> Chinese Windows. Not their fault, I must add. Ah, I cannot bear setting
I have no experience with Windows Gtk, but it could well be due
to the fact that Win32 APIs come in two flavors, 'A'(NSI) APIs and 'W'
APIs. MS recommened a few different paths to support both pre-Unicode
("ANSI"-based ) Windows (Win 9x/ME) and Unicode-based Windows
(Win2k/XP). One of them is to use 'MSLU'(Microsoft Layer for Unicode?)
with pure 'W' APIs (not using 'A' APIs at all). Mozilla developers
once considered this approach, but gave it up because it led to a
dillemma. To make Mozilla run under Win 9x/ME, Mozilla developers have to tell
Mozilla users to install MS IE 5.x or later (or MS Office or other programs
that have license to bundle MSLU dll with themselves). Obviously,
it doesn't make much sense to ask users to install its competitor before
using it (needless to say, the reality is that virtually MS Win users
have MS IE installed so that we don't have to worry...). There may be
other reasons that MSLU path was not taken that I don't know of.
What Mozilla ended up doing is to write our own wrappers and function
pointers for two dozen or so of Win32 APIs that get pointed to
either A APIs or W APIs according to the run-time detection of the
OS (Win9x/ME vs Win2k/XP). Mozilla's transition to this is not yet
complete (see http://bugzilla.mozilla.org/show_bug.cgi?id=162361 and
http://www.mozilla.org/releases/mozilla1.4/known-issues-int.html)
It's likely that Win32 Gtk is still dependent on 'A'NSI APIs. However,
this is a pure speculation and could well be completely wrong.
> Linux locale to Chinese, which makes the desktop too ugly to me. Rationale:
> The good intent of Open Source developers may not result in understanding
> the requirements of Asian users owing to lack of native
> developers/testers/users.
That's a bit strange. My desktop under ko_KR.UTF-8 locale is not so bad.
Anyway, it's not yet as pretty as that of Win32.
I think it's not so much due to defects in programs as due to the lack of
high-quality fonts. These days, most Linux distributions come with free
truetype fonts for zh, ja, ko, th and other Asian scripts. However,
the number and the quality of fonts for Linux desktop are still
inferior to those for Windows.
> There seems little sense now arguing the virtues of UTF-8 and UTF-16.
> Technically they both have advantages and disadvantages. I suppose we have
> presented enough of them in this discussion.
Let me just add my last comment...
If MS had decided to use UTF-8 (instead of coming up with a whole new set of
APIs for UTF-16) with 'A' APIs, Mozilla developers' headache(and that of
other opensource developers) mentioned above would have been a lot easier
to cure :-) Of course, this is just one aspect of UTF-8/'A' APIs vs
UTF-16/'W' APIs and there are many other things to consider in case of Win32.
Jungshik
--
Linux-UTF8: i18n of Linux on all levels
Archive: http://mail.nl.linux.org/linux-utf8/
