Hello,

In the thread on vertical CJK, I mentioned that I would like to
continue helping with CJK LaTeX development. Of course, I am also
looking at Omega, XeTeX and pTeX for ideas.

What I am interested in most is getting UTF-8 support for things which
at present are only available in a country-specific encoding (in this
case SJIS and EUC-JP for Japanese).

This may mean, among other things:

 - making changes to CJK macros, and/or

 - modifying font definition and encoding files, /or

 - obtaining more fonts if desired glyphs are not available.

At present, I want to use half-width katakana, and although Werner
Lemberg has noted that they are typographically "useless" they are
nevertheless quite commonly used in official documents, even, for
example, in CVs.

To continue from the previous discussion:

Gernot Hassenpflug wrote:
>Werner LEMBERG <[EMAIL PROTECTED]> writes:

>>> I have been looking for about 6 months for a way to obtain
>>> half-width katakana in LaTeX while using the package CJKutf8.  I've
>>> read through the CJK documentation files on CTAN, and some TUGBoat
>>> articles on the structure of the CJK package.
>>
>> There are two cases.
>>
>>   . You have a font which you can access in SJIS encoding (for a
>
>OK, I checked, it works in SJIS just fine.
>
>>   . Unicode fonts for Japanese normally have half-width katakana in
>>     the range U+FF65 to U+FF9F, as defined in the Unicode standard.
>>     Then proceed as usual!  Again, this works only if you actually
>>     have the proper glyphs in the font.
>
>All right, this is what I actually want (not SJIS, since I want to mix
>more languages).

Werner, you mentioned that Wadalab fonts do not include half-width
katakana. I am afraid I had no idea of that.

I've spent some time on H. Okumura's wiki at

http://oku.edu.mie-u.ac.jp/~okumura/texwiki/

and found literally dozens of Japanese fonts, and some notes on
half-width and full-width characters. I do not even know where to
begin yet, the information is so much. It'll take me some days to go
over all the interesting links.

Incidentally, there is an interesting and very recent interview with
him here, dating from 4 June 2007:

http://www.tug.org/interviews/interview-files/haruhiko-okumura.html

I wonder the following:

1. How can one get a font test output that shows all the font's glyphs
   in at least one typeface? (nfssfont.tex outputs some tables and
   various other interesting test cases, but no way does it give a
   table of all the font's characters). I see there is a font test
   program on Okumura's wiki, but I have not yet tried to see if it
   works with LaTeX (TeXLive) rather than pTeX. Or is it more sensible
   to look at the font with other tools instead (off hand I don't know
   the best tool, but for fonts installed under X, I have a "character
   map" program here on my Ubuntu system which allows me to at least
   see all the characters in the font, very useful provided the glyphs
   are actually shown under X, some just appear as "boxes" with the
   unicode index shown).

2. If wadalab does not include all characters (AFAIK, mixed styles in
   Japanese in LaTeX are already done out of need, since support for
   Level 3 kanji is not sufficient in all fonts, so other fonts need
   to be substituted sometimes), are there alternative preferred fonts
   that might be better? Or at least used to complete gaps? As I wrote
   above, there appear to be many many Japanese fonts, which I will
   look at slowly.

3. I am not capable of doing much coding with latex macros yet, so I
   wonder what useful font manipulation could at present NOT be
   provided in CJK with UTF-8 encoding even if decent fonts are
   available?

4. I have several high-quality Epson TTFs and TTCs, which I would like
   to try out under LaTeX with CJK. I have heard that the metrics can
   be distributed freely so that people who have these fonts (on their
   MS Windows system, for example) can easily make use of them. Would
   it be OK to incorporate such metrics into CJK?

>>     Note, however, that the CJK package has no mechanism to
>>     automatically map full-width katakana to the half-width forms
>>     (except for SJIS encoding).  With a good Japanese OpenType font,
>>     this can be done by activating thee `hwid' GSUB feature for the
>>     `kana' script.  My CJK macros can't do that, for obvious reasons;
>>     If you really need this you might try XeTeX which has native
>>     support for OpenType features.
>
>I'll check out the XeTeX way, and do more reading/experimentation.

If UIM is being used to input characters in Emacs, and the font
supports half-width katakana (glyphs), would that be sufficient, since
that would then be a direct input of the half-width code rather than a
mapping of full-width to half-width?

Regards,
        Gernot
-- 
Gernot Hassenpflug, NICT, Tokyo


_______________________________________________
Cjk maillist  -  [email protected]
https://lists.ffii.org/mailman/listinfo/cjk

Reply via email to