Mailing List Archive

Support open source code!


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: tlug: kinput2 vs. egg [was: Yatta! XEmacs Japanese input with kinput2 via XIM]



--------------------------------------------------------
tlug note from "Stephen J. Turnbull" <turnbull@example.com>
--------------------------------------------------------
>>>>> "C" == C Oda <craig@example.com> writes:

    C> On Tue, 20 May 1997, Dennis McMurchy wrote:

    denismcm> Way to go Steve!  You did it!

    C> Yes, congratulations Stephen.  Another notch in the old
    C> computer case.  :-) That looked like it required a bit of
    C> effort.

Actually, the hardest part was getting gdb and learning enough about
it to get the stack trace.  Just before the thing bombed there was an
unknown function in a weird place; this almost invariably means that
something got passed a null pointer (one good reason to use Java; no
pointers means no null pointers).  This told me where the relevant
code was in Emacs.

Then when I figured that out, I knew from previous experience that you 
have to be careful about passing the areas to XOpenIC, and once I
found those things were zero, it was that obvious thing to try.  Next
I would give up :-)

    denismcm> No, let me ask the really basic (read: possibly very
    denismcm> dumb) question here: why on earth would anyone who could

(1) Craig's reason:  uniform interface out of Mule and in
(2) Even simpler reason:  to get any kind of interface out of Mule
(3) Somewhat uniform interface across back ends

    denismcm> have Wnn + tamago(Egg) input (that actually seems to
    denismcm> make a pretty good job of converting long strings of
    denismcm> input) want to use Kinput2, which is horrible to use on
    denismcm> anything but the shortest input strings (i.e. single
    denismcm> words) and even

    C> Ahh yes, you noticed this too.  I thought that it was just my
    C> settings.  I found this pretty hard to understand since kinput2
    C> is set to use wnn on my system, so I thought that it would
    C> parse the Japanese bunsho correctly and work the same as

As Dennis points out, if you've got the same back end, you get the
same bunsho and the same trial henkans.

What is probably horrible about "kinput2 + your favorite back end" is
the key-strokes.  Configuring keystrokes in Mule is relatively easy.
You make a mode map and hang call-backs on the keystroke sequences.
People use control and meta keys, not funky-shun keys and
"control-kanji" (what's that on my US keyboard?)  Of course, Mule
hackers are always playing with the keymaps.  So it's not surprising
that tamago has a well-configured keyboard interface.

With X, on the other hand, there are at least 3 (4 in Motif) ways I
can remember off-hand to configure the keyboard.  So my guess is that
it's hard to move around in renbutsu and select new henkan in kinput2, 
mostly because since nobody uses it (yet, for non-Japanese hackers---I 
suppose people with well-set-up proprietary multilingual systems do)
it isn't very well configured for us non-native-speakers.  This will
improve.  But probably more slowly, since there are several ways to do 
the config, unlike Mule, where it's extremely standard.

Also, kinput2 used to look a lot like uum, which I've always hated.
And with the over-the-spot method, you can concentrate on the screen.

    C> egg-wnn.  Which brings up the question, where does the
    C> responsibility of the FEP stop and the JIM take over?  I'm not
    C> even sure I have the terminology correct here.  I had thought
    C> that kinput2 was just a front-end processor and Wnn was the
    C> Japanese input method.  I know that kinput2 can be used with
    C> Canna, so I see the distinction there.

The terminology is not at all stable on this.  I think your
distinction is fine.  According to Nishikimi, et al, _Realization of a
Multilingual Environment_, input methods come in three types:
1-keystroke-per-character, multi-keystroke-per-character, and
multi-keystroke-plus-conversion-server.  In general these correspond
to national language keyboards, compose keys and/or romaji, and
ideographic input respectively.  But kanji can be input by "Tcode" in
which two arbitrary keystrokes are assigned to each kanji.  And you
can think of ku-ten or JIS code input as multi-stroke-without-henkan
input, too.

The XIM standard describes the conversion server as a "back-end".  But 
this is because XIM only describes the interface between Xlib and the
front-end; it doesn't care about the back-end.

    C> Here is what I thought:

    C> Front End Processors: kinput2, egg, uum, canuum

Also xwnmo, specific to Wnn.

    C> Japanese Input Method: canna, wnn

    C> both FEP and JIM: skk (I've never tried this at all)

    denismcm> I must be missing something (admittedly XIM is beyond my
    denismcm> ken - does it really make the difference?).

It doesn't make a difference as long as you want to stick to one
Oriental language, or do everything in Mule.  But what are you going
to do when Netscape really gives you Japanese/Chinese/etc in form and
search input?  Are you going to blow off Netscape and use w3.el
anyway?  Netscape will get those capabilities just as soon as XIM
matures, you can bet on that.  But until there's a standard, I don't
think there's enough market advantage in having a truly Japanese-
capable browser in the Orient for Netscape to care.  I don't know
anybody who uses MSIE for Japanese.  That doesn't mean it's worse than
Netscape; it does mean it doesn't have a significant advantage.
Japanese aren't doing that much form-filling yet, I guess.  Once
there's a standard, though, everybody's gonna have it, and Netscape is 
going to want to be first.

    C> Re: List Admin

    C> There is a 40,000 character size limit on any single post.

Aha!  I'll check that in the future.  Pity, it's "only" 50kB, it
almost fit :-)

    C> I can either take the limit off (which I feel is a little
    C> dangerous), or we can adjust.  Let me know if anyone has
    C> preferences.

I think you should stick to the limit.

    C> Re: Meeting this weekend?

    C> Anyone up for a meeting this weekend?  Saturday or Sunday?  JWT

I can't make it :-(

-- 
                            Stephen J. Turnbull
Institute of Policy and Planning Sciences                    Yaseppochi-Gumi
University of Tsukuba                      http://turnbull.sk.tsukuba.ac.jp/
Tel: +81 (298) 53-5091;  Fax: 55-3849              turnbull@example.com
-----------------------------------------------------------------
a word from the sponsor will appear below
-----------------------------------------------------------------
The TLUG mailing list is proudly sponsored by TWICS - Japan's First
Public-Access Internet System.  Now offering 20,000 yen/year flat
rate Internet access with no time charges.  Full line of corporate
Internet and intranet products are available.   info@example.com
Tel: 03-3351-5977   Fax: 03-3353-6096


Home | Main Index | Thread Index

Home Page Mailing List Linux and Japan TLUG Members Links