Re: Unicode end-users

From: Christopher Hogan (
Date: Sun Aug 03 1997 - 15:31:22 EDT

Martin J. Duerst writes:

> There is a nice keyboard resource editor, surely easier to use than quail,
> on the Mac. Just drag-and-drop your letters onto the keys you want them.
> I'm sure something similar exists somewhere for Windows.
> Quail is a little bit more powerful in terms of letter combinations
> (for examlpe to input Korean Hangul), but it doesn't go as far as
> covering Japanese or Chinese ideographic input. At some point, you
> need to be able to specify keyboard behaviour programmatically, and
> that's not something for the end user.

While I allow that specifying an IPE for ideograms is more complicated
than specifying one for an alphabetic language, I do not think that it
requires the full power of your standard programming languages. I
think there are some generalities to be captured about entering
ideographic scripts (and about alphabetic and syllabic ones), which
could be made available as functionality in an IPE toolkit.

I do not know very much about entering ideographic characters, but one
way to do so that I have seen is to permit the user to type in an
appropriate phoneticization (say pinyin or katakana/hiragana/romaji)
and convert into ideographs at the end of words by presenting the most
common alternate and allowing the user to select others. I do not
know if this is the most common way of entering Chinese and Japanese
characters, but I have used at least one product that did this for
Japanese, and it seemed to have worked fairly well (of course my
meagre amount of Japanese is no where near native).

If the above is an acceptable system for entering ideographic
characters, then perhaps a wizard built into an IPE toolkit would
permit something like the above to be built more easily.

The main problem I see with the above kind of solution is generating a
mapping file from the phoneticization to the ideographs. I agree that
this is a time-consuming task, especially for languages where such an
animal does not already exist, certainly it is better than no input

p.s. is it true then, that the Japanese and Chinese IPEs in MULE are
not written using quail?


christopher m. hogan			language technologies institute			     carnegie mellon university
computational linguistics		                 pittsburgh, pa

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:36 EDT