Relay-Version: version B 2.10.2 9/5/84; site titcca.UUCP
Posting-Version: version B 2.10.2 9/5/84; site titisa.UUCP
Path: titcca!titisa!kuno
From: kuno@titisa.UUCP (Yasusi Kuno)
Newsgroups: fj.kanji
Subject: Re: Is there Kanji-> Romaji conversion program
Message-ID: < 504@titisa.UUCP> 
Date: 8 Jul 85 04:02:21 GMT
Date-Received: 8 Jul 85 04:07:54 GMT
References: 
Organization: Tokyo Institute of Tech., Japan
Lines: 36

>    Has anybody ever written a filter program that converts KANJI
>  file to Romaji?

I agree that such a tool will be valuable for current Unix
environment.

>    What is most difficult is, I suppose, that how to do
>  WAKACHIGAKI, to separate words from words.  What do you think ?

One way to do this is -- to scan kanji text and insert space
where character type changes from Hiragana to Kanji.  This
method comes from the obserbation that, most words which contains
Kanji start from Kanji-character.

>    Or, we may write a program that a person who writes a kanji
>  message runs immediately after his composition.

Isn't it better to capture users input stroke (they must be
romaji or kana text!!) directry?

>    Is there such a program ?  Is there any volunteer ?

I think we need common basis for our kanji text processing,
on which we can share our effort for building tools.  First
of all, we need a text editor (such as ed?), which is written
in C (or some common language -- but not Fortran, I hope!),
runs on most unix systems, and can be distributed freely.
Of course we can use better tools at our will, but such an
editer can serve as standerd for kanji text exchange.  Romaji
text generater discussed above can be included in it.

What do you think?

						Kuno
                         Junet: kuno@titisa.titan.junet
                         Uucp:  titcca!kuno