Re: tlug: Li18nux-related meeting 11/30

[Adrian Havill <havill@example.com>]

On Tue, 30 Nov 1999, Stephen J. Turnbull wrote:
>     Scott> Is this really an i18n initiative or is it a j12n
>     Scott> initiative? :) -- ie,what other languages/countries are
>     Scott> represented here?

Most of the "blue-chip" languages  are represented, esp. in the
kanji/hanxi/hanja dept (all the major Chinese/Japanese (not as much
_active_ Korean participation groups) and the German/French gang. I
_would_ like to see more representation in the middle east dept., however,
is the display tech for that makes displaying kanji seem trivial. Almost
all languages are present "in spirit" (endorsement), but the "actively
participating" parties primarily have the above languages as their "killer
langs" to support.

> Developers and Japanese are represented.  They had a bitch of a time
> getting a European co-chair and the U.S. co-chair is native Japanese.

The European co-chair is a SuSE. The N. America (US/CA/MX) rep is
a Sun guy, and one of the primary architects of the original X
XIM/XOM/FontSet API/protocols.

> It is probably a positive sign though.  Turbolinux is there and AFAIK
> Laser5 is not.  ;-)  Seriously, now....

It's not a closed group. Anyone can join. Laser5 is most certainly
welcome. The mail CC headers I get show that the other guys (in Japan) are
present "in spirit."

> I attended one meeting of the Linux Research Society/National Language 
> Systems Section, which seems to be the core of the Li18nux Japan
> chapter.  Most of the people there seemed to be corporate developers
> waiting for Linux I18N to happen so they can take advantage of it.

The general gist that they seem to give off (IBM, Sun) is that they've
developed solutions for I18N (or at least better L10N support than what
GNU/Linux systems have) on their native proprietary platforms (AIX and
Solaris being the best examples... and now they want to write apps for
Linux, but they're finding that without giving a helping "push", it's not
going to happen as fast as they would like.

And to be fair, some of those corporate guys are actively testing glibc
2.2 _and_ submitting patches to Ulrich Drepper... and their
participation is far exceeding that of the non-corporate
entities. Non-corporate entities are more than welcome to pull the source
off of CVS and chip in. :) 

I'd love to see more active participation by the TLUGers in NLS. If you
have an empty partition you too can contribute to getting glibc 2.2 out
the door-- you don't need to hack it-- even just pulling the source off
CVS, grabbing a C manual, and randomly trying out anything in the
"wchar.h", "locale.h", or "wctype.h" headers in small demo programs and
reporting "it works" or "it doesn't work" would be EXTREMELY
valuable. (There's a lot of functions in there, and in those functions, a
lot of combinations that need to be tried to test the robustness) 

> There was a small core of I18N developers, but they have a strong
> agenda.  The fact that the first major activity is to have a "system
> architecture" subgroup is a bad sign on that front; they're pushing
> certain corporate subsystems (Sun's IIIMF and IBM's Unicode library)
> hard.  OTOH, Adrian Havill didn't seem to have objections to it, so
> maybe it's sufficiently plausible to be a big step forward.

Let's put a conditional clause on that. :)

I'd gladly accept and use technologies out there that already exist rather
than re-invent the wheel if and only if they're OSS (as defined by
www.opensource.org)-- preferably with licenses a little more GPLish than
Xish (see below). And the technologies can't suck, of course.

The track record of Sun/IBM/HP (SGI's L10N/I18N support in IRIX never
quite worked right at the clib level. ;) ), which took X11, fixed the
(among other things) multi-byte/wide support, and kept those changes
proprietary without releasing the code so that the Linux community that
cares about things other than ASCII couldn't benefit, shouldn't be
forgotten. We also shouldn't forget the lessons learned from what
Motif/CDE did to the freeness of a standard X platform.

If and if IBM's Unicode Classes is open source (haven't checked on it's
license status recently), I believe if deserves a serious look. It's
extremely portable (fully Std. C, it compiles on non-GNU platforms as
well) and fairly robust.

IIIMF has had Hiura's (one of the architects of the X11R6.4 
XIM/XOM/FontSet stuff) hand in it, and was designed to address many of the
prior X system's definciences (too complicated to use even for trivial
programs), and is independent of X at the API level (originally designed
for Java, which we all know runs on non-X systems such as Win32 and Mac.

I've yet to see IBM's classes in action in software, but you can see IIIMF
in action on Java 2 systems ironically with a piece of software mentioned
on this thread: ARK for Java. You can smoothly enter
Chinese/Japanese/Whatever (those are the only two IMEs *I* have installed
at least) and it works like a charm.

The Unicode Classes are really too new to comment on how sound the API
is-- there's nothing quite like it anywhere else (the closest possible
thing would be Java's API relating to Unicode/character
classification). The IIIMF seems to have been designed from a viewpoint of
"lessons learned from IMEs with X", and adds some cool new paradigms for
working with IME server/clients.

-------------------------------------------------------------------
Next Nomikai: December 17 (Fri), 19:00 Tengu TokyoEkiMae 03-3275-3691
Next Technical Meeting: January, 2000
* Topic: "glibc - current status and future developments"
* Guest Speaker: Ulrich Drepper (Cygnus Solutions)
* Place: Oracle Japan HQ 12F Seminar Room (New Otani Garden Court)
-------------------------------------------------------------------
more info: http://www.tlug.gr.jp        Sponsor: Global Online Japan