Re: [tlug] List of Japanese Open Source Projects?

["Stephen J. Turnbull" <stephen@example.com>]

Gen Kanai writes:

 > I'd love to hear other examples of OSS from Japan.

Emacs/MULE (XEmacs/MULE was done by Sun in San Jose).  Many Emacs
    applications (dictionary lookup: edict, lookup; MUAs: mew,
    wanderlust, tm, SEMI, WEMI; IRC clients: liece, riece; encryption:
    easy-pg).
The M17N project at the Electro-Technical Laboratory.
XIM.
IIIMF.
Mosaic-L10N.
"The Japanese patch" for file utils such as jgrep and editors like jvim.
    (There's really only one. :-)
The Wadalab outline fonts.  Several other Japanese fonts.
FreeWnn, Canna, and friends.
Several Linux distros (including Green Frog Linux).
Several terminal emulators (kon, kterm, m17nterm or st like that).
Winnie (and other P2P software).
Lots of copy-protection crackers.

 > > Are there any good online resources covering Japanese open source
 > > projects?

As a class?  I don't know of any.  Most Japanese are either FLOSS fans
not really associated with any project, or they feel a much stronger
association with their particular projects than with the FLOSS
movement as a whole.  Nothing wrong with either tendency, but neither
is likely to produce a good survey site.

You might try starting at Hiroo Yamagata's site (cruel.org).  Accept
no imitations: there is an infamous "producer of depthless paintings"
by the same name, but that's not our Hiroo.

You could also try the O'Reilly book LINUX日本語環境 (disclaimer: if
you buy it, Hiroo, Craig Oda, Rob Bickel, and I collect royalties, so
I'm not disinterested ;-).  While it's about building a Japanese
environment, it's likely to mention Japanese-only tools that were
built in Japan. 

 > > Ruby, of course, is famous.

Ruby is really the only one that was well done, IMO.  Most of the rest
(as you can see from the list above) are very much specialized to
Japanese language support, and typically (even if labeled
"multilingualization" or "internationalization") were not terribly
useful even to other Hanzi-using folks.  Even IPv6 (where WIDE made
major, major contributions) looks more like an ISO standard (beautiful
structure, not used in the real world) than a real IETF RFC.

The only software I'm aware of where the "Japanese patch" made a
successful transition to internationalized software was Emacs/MULE
(which started out as NEmacs = Nihongo Emacs).  In most other cases,
the internationalized software currently in use (often superseding a
Japanese patch) grew out of software internationalized across ISO 8859
(ie, unibyte) encodings using the locale model and gettext, and now
supporting multibyte encodings via Unicode (and iconv at the
perimeter).

And there are some fascinating Japanese paradoxes.  For example, at
NEC's Linux lab in Kobe in the mid-2000s, except for Steve Baur's
short stint there, the only Linux box was the test host.  All the
developers used Windows NT as a host for the development environment
to write Linux code.  "Only in Japan ...."