Re: jisx0213 problems (was: Mule-UCS 0.84 (KOUGETSUDAI) release.)

[川幡 太一 <batta@xxxxxxxxxxxxxxx>( KAWABATA, Taichi / كَاوَابَاتَا تَااِيتشِي / कवाबान=?UTF-8?B?4KS+IOCkqOCkh+CkmuCkvw)]

Hello, Kinoshita-san, 

>>>>> In [mule : No.03301] (jisx0213 problems (was: Mule-UCS 0.84 (KOUGETSUDAI) release.)), 
>>>>> Kinoshita-san wrote:

> > I don't know whether Kawabata-san is reading this, but he said there
> > were a number of problems with the jisx0213 translations and sent a
> > detailed analysis of some other proposed translation tables.  I guess
> > I could pass that on if anyone is interested.

> Please explain the problems in detail.  I hope the problems will
> be resolved.

As far as I know, there are three problems for mapping JIS X 0213 to UCS.


1. Should 2-93-27 be treated as U+9B1D or U+9B1C in UCS?  

  It seems that the original intention of JIS Committee is to map this
  character to U+9B1D, but since they re-used the glyph of U+9B1C used
  in JIS X 0212 standard, 0U+9B1C is probably more appropriate.  (It
  is also convenient to convert JIS X 0212 document to JIS X 0213 thru
  UCS.)  However, as far as I know, there is no amendment of the
  standard concerning to this character.


2. How to convert 1-11-69 and 1-11-70 (or 1-11-64 to 1-11-68) to UCS?

  As described in the UCS standard, the U+02E5 to U+02E9 have the
  nature of ligation, but the corresponding 1-11-64 to 1-11-68 do not.
  Therefore, 1-11-64 to 1-11-68 can not be converted to U+02E5 to
  U+02E9 just one by one, or the different shape would appear when
  these characters appear consecutively.  1-11-69 and 1-11-70 should
  be converted to U+02E5 U+02E9 and U+02E9 U+02E5 respectively
  according to the UCS, but that is beyond the scope of the use of
  1-11-64 to 1-11-68, which is merely cited from the IPA standard.

  One solution may be to convert these character as follows, but this
  is merely my opinion.

    1-11-64     U+02E5 U+200C
    1-11-65     U+02E6 U+200C
    1-11-66     U+02E7 U+200C
    1-11-67     U+02E8 U+200C
    1-11-68     U+02E9 U+200C
    1-11-69     U+02E9 U+02E5 U+200C
    1-11-70     U+02E5 U+02E9 U+200C


3. How to map 1-5-87 and various other characters which are not
   registered to UCS (for now) ?

  These characters do not appear in the UCS, as the people of UCS
  claim that these characters can be "synthesized" from the existing
  characters.  However, since they can be hardly considered to be
  separate character from the sense of other kana characters, JCS
  demanded the independent code points for them.  Anyway, for now,
  they require 1:m mapping feature, which Mule-UCS do not provide for
  now.


-- 
----------------------------------------
KAWABATA, Taichi (batta@xxxxxxxxxxxxxxx)
----------------------------------------