Re: FW: Question about Shift_JIS mapping table

From: Martin J. Duerst (
Date: Thu Jun 24 1999 - 19:07:45 EDT

Hello Magda,

What you say is basically true, but it's fairly limited.
The differences appear in the mapping of symbols, some
vendors seem to prefer mapping 'full-width' symbols to
the official (but half-width) symbols in Unicode, others
prefer to map them to the corresponding compatibility

As far as I understand, the problem can be avoided by
refraining from using the full-width variants of these
symbols. (By the way, this is actually what JIS X 0208-1997

It's very important to note that neither Katakana nor
Hiragana nor Kanji are affected.

As for the problem with one-to-one, this is also true.
The problem here is the backslash/Yen symbol. It takes
a programmer (or some AI or what) to e.g. correctly
convert a C program. This problem has been perpetuated
into Unicode by Microsoft (and maybe others) by providing
separate Unicode fonts for the Japanese market. The fonts
differ in only one place: They use a Yen mark for the
backslash. That way, the Japanese never get out of their
confusion between backslash and Yen.

So in summary: The problem is there, but it is fairly local.

Regards, Martin.

At 09:03 99/06/24 -0700, Magda Danish (Unicode) wrote:
> -----Original Message-----
> From: Johan Troedsson []
> Sent: Thursday, June 24, 1999 7:31 AM
> To: ''
> Subject: Question about Shift_JIS mapping table
> Hello,
> just want to verify whether a rumour is true or not. I've heard that there
> in practice exist a number of different mapping tables (four?) between
> shift_jis and unicode. Secondly, I've heard that not all the characters used
> in Japan in the shift_jis character encoding can be mapped on unicode, or
> there is at least no one-on-one mapping. Could you please help me on this
> matter?
> /regards Johan Troedsson
> * * *
> Johan Troedsson
> AU-System Radio AB
> S-223 70 Lund, SWEDEN
> Phone: +46-46 28 656 82
> Mobile: +46-70 66 656 82
> Fax: +46-46 28 656 20
> * * *

#-#-# Martin J. Du"rst, World Wide Web Consortium

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:47 EDT