Re: Linux & Unicode

From: Tague Griffith (
Date: Thu Dec 03 1998 - 15:02:53 EST

Glen :-

I think one of the problems is that you are bringing up a hot-button
issue for the Linux community. Linux is going through the same
acceptance-phase that many other communities (standards bodies, other
platforms) went through several years ago -- it's just comming much
later to the Linux community. There is alot of resistance to adopting
Unicode (particularly UTF-8 for the file system) because some people see
their file sizes growing by x3.

In general, I would say the grass roots Linux community is mostly
opposed to the adoption of Unicode and that the push is comming from
commerical vendors like the folks from Troll Tech (Qt) and Red Hat (Gtk
work). Linux is still a "hacker" OS and the internationalization
support, while there for some languages, is still very imature in
comparison to other platforms. Linux is in the stage that the MacOS or
Windows *.* was several years ago -- and "they" are still learning the
leasons that many of us have already learned.

Unlike a commerical company, Linux really has no cetral architect or
product requirements to "force" these changes. Because linux is
opensource, no one can just change the system to Unicode, you have to
have an existing Unicode distribution that is winning people over. Linux
is a conglomerate of different organizations each with their own agenda
-- often supporting mutliple platforms which not all may support
Unicode. Also, remember that alot of Linux is comming from
non-professional developers which might explain some of the strange
responses you are getting, but we get them on the Unicode list as well.

As far as Red Hat and GTk, I've found them to be pretty helpful. You
might want to look at, it talks
about GTK's plan for supporting Unicode. Once GTK supports it, GNOME as
a derivative should pick it up.


Glen wrote:
> While I certainly don't claim to be aware of *all* i18n issues, I'm well
> aware that the text encoding isn't the end of it. I do consider it the
> *beginning* of it, though.
> What's the point in even discussing issues like bidirectionality,
> composition, line breaking algorithms fancier than break-on-white-space,
> ruby, Unicode-savvy regular expressions, and so on with people who don't
> even consider requests for Unicode itself worthy of a response?
> As I said, I've had no luck eliciting any response at all from GNOME, KDE,
> RedHat, GTK, and others regarding Unicode. No responses whatsoever from any
> of them. Posting the question on a Linux newsgroup merely elicited a couple
> of "Micro$oft sux, you luser"-type replies. Not very enlightening. On
> Slashdot, I was informed by a European that Linux had the "concept of
> locales, so Unicode doesn't matter", and by the Chinese "developer of the
> Chinese version of KDE" that I didn't know what I was talking about because
> Unicode was useless to Chinese developers. Hmm.
> Now, people's willingness to respond to me isn't necessarily indicative of
> the level of their interest in, or committment to, Unicode, of course. I'm
> aware of that, which is why I have to assume that I may just be misinformed.
> I haven't been able to find any real evidence of any other sort that Unicode
> matters much to the Linux community, though, and lots of evidence that it
> doesn't.
> If you want to go beyond Unicode and ask them about support for Tibetan
> decomposition or Thai line breaking, be my guest. Good luck.
> With the notable exception of Arnt's wonderful response to my question
> (thanks, Arnt!), I haven't had much luck and don't see much enthusiasm for
> Unicode in the Linux community. It appears that Microsoft, Sun, IBM, Oracle,
> Xerox, etc., as truly multinational organizations, have a much more global
> mindset than the Linux community. They want a single, consistent, worldwide
> text encoding for what they see as a unified, global IT infrastructure.
> Linuxers, on the other hand, seem to see their machines as one-of-a-kind
> works of art, as individual expressions, not as cogs in a global machine.
> They seem quite content to have their own machines limited to supporting
> their own language(s), and nothing more. I think that's short-sighted.
> There's a lot to like about Linux. The price is right, especially for
> developing countries where Unicode might be especially useful. Its open
> model would seem to offer the hope that, with the right Unicode-based APIs,
> the implementations could be gradually fleshed out by developers around the
> world to handle most of the world's languages/script systems. Unfortunately,
> single-byte text APIs won't provide the necessary foundation to enable such
> a scenario, and it doesn't look likely that Linux's support for Unicode will
> rival Windows' any time soon.
> Glen Perkins
> ==========================================================
> GP wrote:
> [snip]
> >I understand the problem of dragging legacy balls-and-chains, but I
> would think that any *new* "alternative to Windows" would start out
> with a pure Unicode-based GUI API.
> It's good to ask about Unicode, but bear in mind that the issues go
> well beyond what encoding models are supported by text APIs. Correct
> rendering of the various scripts included in Unicode is far from
> trivial, even if only for the writing systems of the major languages
> using those scripts, and not even MS has fully crossed that bridge.
> And rendering is only one of several issues. If you're looking for a
> GUI for Linux that really does what is needed to handle the way the
> world writes, make sure all the issues are addressed!

tague griffith (
client internationalization

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:43 EDT