Re: [tlug] Japanese Input on CentOS / KDE

["Lyle (Hiroshi) Saxon" <ronfaxon@example.com>]

Dave Gutteridge wrote:

> The reason that Scott's web page confuses me, and why I'm still 
> generally confused about Japanese input, are that every page, and each 
> piece of advice I get on mailing lists seems to be talking about 
> entirely different systems. There are all these methods, and layers to 
> methods, such as iiimf, canna, wnn, kbx, xim, kterm, kinput... I'm not 
> sure which does what, and each web page, including Scott's, seems to 
> draw on only a few of them. This means that I go to one web page, it 
> gives me a set of instructions, they don't work, so I look for another 
> web page to try and see if I can find what went wrong, and the new web 
> page leads me down an entirely different path. I feel like over the 
> last three weeks of configuring my Linux desktop to use Japanese, I've 
> never touched the same setting twice.
>
> It also means that after fiddling with some setting, I ask on a 
> mailing list about why it didn't work, and someone gives me an answer 
> that probably works for their set up, but because they most likely 
> followed a different route to get there, it doesn't seem to apply in 
> my case. Surely it can't be that everyone has a completely different 
> method of Japanese (or other languages) input, and yet it seems that 
> way now.


I was in the same spot, although my attempts were more spread out - 
squeezed in between other work. I also never did get things rolling in 
Japanese from the advice I got, so I sent anything that needed to be 
sent in Japanese from a SuSE 9.1 box that I'd originally set up as 
Japanese during the install. The SuSE 9.1 box that I'd originally set up 
in English could read Japanese, but I could never enter it.

With the advent of 9.3 though, selecting Japanese as a secondary 
language fixed the problem and now I'm able to write in Japanese - これ 
は日本語なのじゃ！ - Banzai! With one caveat though, when I upgraded my 
wife's computer from 9.1 to 9.3, the input method changed and she's 
none-too-pleased with the new one. There are two e-mail from someone 
about ATOK that I've been meaning to follow up on, but haven't had the 
time to...yet. ATOK you have to buy, but the input method is - 
apparently - more polished. I don't know if it will help at all, but 
here is the second of the two e-mails:

    * Date: Fri, 10 Jun 2005 11:31:03 +0900
    * *From*: *Hiroshi Miura <miura@example.com
      <mailto:miura@DOMAIN.HIDDEN?SUBJECT=Re%3A%20%5Btlug%5D%20ATOK-X>>*
    * *Subject*: *Re: [tlug] ATOK-X*
    * References: <42A1D4C2.2050309@example.com
      <http://www.tlug.jp/ML/0506/msg00009.html>>
    * User-agent: Wanderlust/2.11.30 (Wonderwall) SEMI/1.14.6(丸岡 )
      FLIM/1.14.6(丸太町 ) APEL/10.6 Emacs/21.4(i386-pc-linux-gnu)
      MULE/5.0 (賢木 )

------------------------------------------------------------------------

Hi,

I'm now one of IIIMF developers -
 input method framework for ATOK, Wnn and other OSS language engines
 from this March.

I heard from other developers that ATOK-X is tightly binded
to iiimf rev.10 and troubled with rev.11.

Because SuSE 9.3 have new iiimf rev.12 in it, ATOK-X has trouble.
To solve it, you may want to revert iiimf package version
to the package bundled with ATOK-X.
Or use Wnn8 bundled with SuSE 9.3 Japanese version.

Same thing is reported with ATOK-X2 and SuSE 9.3 at someone's web site.
Because JustSystem modified OSS iiimf source to implement 
their own extention for ATOK-X2.

There is only one solution to use ATOK-(X,X2), 
removing SuSE iiimf package and install one bundled with ATOK.

I hope it helps you.

At Sun, 05 Jun 2005 01:20:18 +0900,
Lyle (Hiroshi) Saxon wrote:
> 
> The wife is *extremely* unhappy with the new kanji input that the
> upgrade to SuSE 9.3 has implemented (she says that a lot of the kanji
> she needs just don't come up, so she ends up using hiragana instead).  I
> need to either get her computer back on the old system or able to use
> the ATOK-X that I bought last year (I've tried several times to get it
> installed, but never succeeded).  After receiving a few spears, I'm
> sitting at the computer with blood dripping on the floor and I promised
> to put some time into seeing what I can do with her computer
> tomorrow....  Not a specific request for help yet, unless there's a
> way to roll back the kanji input of 9.3 to what it was with 9.1.....
> 

I have had the same kind of trouble with Japanese input as Lyle's wife.  The Linux magazine article that i mentioned 1 or 2 months ago really seems to have exaggerated how well Japanese input works with WNN8.  (Sorry about that, Lyle, or anyone else who might have been encouraged to buy SuSE 9.3 because of that).  But i wonder if WNN8 is even installed on my computer.  I followed the instructions that came in the box about how to install WNN8, and it seemed to have installed, but after 1 or 2 months of typing in Japanese with it, i can't believe that WNN8 is supposed to work so poorly.  When i type in Japanese in OpenOffice 2.0 prerelease in SuSE 9.3, i have 6 options, and the 2 that seem to work the best are M17N-ja-anthy and UIM-anthy.  Is WNN8 even installed?  Is there a simple way to find out?  I looked at the Omron web site for WNN8 (http://www.omronsoft.co.jp/SP/pcunix/wnn8/toku.
html), and when i try to type Mori Ogai （森鴎外）i only get one option for the character 
for "O" （鴎）.  What i have is not as fancy as what they describe there as the "characteristics and functions" of WNN8.  Something is not right.  If anyone can help me, i'd really appreciate it.

This is my major complaint with SuSE 9.3, but English input works fine in all the applications, and even with the Japanese input not working great (not as well as Atok 16 on the Mac), it's tolerable.  I have the feeling that i'm doing something wrong.  I haven't had any crashes.  The system seems really stable.  For an old computer like the one i'm using (Fujitsu FMV 6700), things work pretty fast.  Actually the Japanese input, even as it is now, works faster than Atok 16 on the Mac in MS Word.  In Word you have to wait (sometimes as long as 10 seconds!) for it to convert a simple word from hiragana to kanji.  That seems inexcusable to me.  Happens all the time.  I used Word in Windows XP recently because i wanted to compare its ease of use with the Japanese version of OpenOffice (what is it? 1.3 or so?  The "rock solid" version), and i found not only that OpenOffice worked better, for me, than Word, but the conversion from hiragana to kanji seems faster in SuSE Linux (whatev
er Japanese input system i have) than in Windows.  In other words, this experimental period with SuSE 9.3 has given me the impression that i should try to work with it, that not only English input, but Japanese input may eventually work better than on Windows.  Being able to run multiple versions of the same program -- like the old and new versions of OOo -- is really nice.  I.e., many things already work much better in Linux.  I'm probably preaching to the choir here, but just in case anyone is interested how a Linux newbie, coming from the Mac world, is finding SuSE...  i thought i'd tell you my impressions.

If anyone can help me with the Japanese input in SuSE, i'd appreciate it.