[Uim] uim-prime vs. anthy (also, possible help for uim-py)

[yusuke at cherubim.icw.co.jp]

Hi!

> I'm a native Chinese speaker from Hong Kong and I speak both Mandarin
> fluently. I can read traditional characters comfortably, though I'm not
> as good with simplified characters. While I'm no pro when it comes to
> typing Chinese, if any of you have questions regarding Chinese, send it
> to the list, and I'll try my best to reply.
Great! and thanks to explaining Chinese input methods for
Tokunaga-san's question. It was very interesting.

I also hava a long standing question about the difference of
usability between syllable based input methods and word/sentence
based ones.

> I saw the exchange earlier about expanding uim-py. I think the most
> useful will be the addition of compound words to the dictionary (similar
> to what Yukido-san said about 今天, 明天, etc)
To enable this feature, reimplementation of Chinese input method
beside uim's generic im is preferred.

Here's my very preliminary implementation of multi syllable PY input method.
 http://cherubim.icw.co.jp/~yusuke/uim-0.3.3z.tar.gz
You can test by following command
$ ./configure;make
$ GTK_IM_MODULE=uim-z sh ./test.sh
text entry will appear. Please type some sentence and press space key.
Say 'nimenhaoma?<space>'.

Many things are to be implemented. However, could you test and give me
some comments?
If the direction is sane, I will
(1) Integrate current syllable based input method into this code and make it slectable.
(2) Anthy like segment selection and candidate selection.
(3) Turn table (uim/z.h) to more editable text.
(4) Implement multi syllable word facility.
(5) Add history based candidate reordering.
(6) Apply some techniques used in Anthy to get higer precision.
(7) Rename from 'z'

Thanks!
--
 CHAOS AND CHANCE!
  Yusuke TABATA (yusuke at cherubim.icw.co.jp)