Around 0 o'clock on Jun 29, Yao Zhang wrote:

> A GB18030 font (covers CJK Unified Ideographs and its extension A in Unicode
> terms) should really be labeled as
>     Simplified Chinese AND Traditional Chinese
> while fonts with GB2312 coverage should be labeled as
>     Simplified Chinese
> and BIG5 coverage should be labeled as
>     Traditional Chinese

I'm confused by this; my exposure to Chinese fonts says that simplified 
Chinese and traditional Chinese have significant overlap in Unicode 
codepoints, but that the glyphs are quite a bit different in appearance.

I'm not interested in discovering which fonts can display a particular
document; that's easily done with Unicode coverage.  What I'm interested in
is selecting the font best suited for presenting data tagged for a
particular language.

Tagging GB18030 fonts as suitable for traditional chinese seems like a 
mistake; the glyph forms are more likely simplified, and it would be 
preferable to use a traditional chinese font, if any is available.  Of 
course, when no traditional chinese font is present, the system will 
search for *any* font which does cover those codepoints, substituting in 
an available simplified chinese font.

I believe I've found a relatively robust way of distinguishing fonts 
designed for traditional chinese from those designed for simplified 
chinese; traditional chinese fonts cover most of BIG5 while simplified 
chinese fonts don't.  Both cover similar amounts of GB18030; as you say, 
that encoding is enormous.  

What I didn't investigate is whether the simplified chinese fonts cover
*different* parts of GB18030 than the traditional fonts.  That might make
the determination easier; simply use the subset of GB18030 normally needed
to present simplified chinese documents as the touchstone instead of the
whole encoding.  For that to work, I'd need a lot more simplified chinese
fonts from various vendors.

> If you need those fonts for testing, I will send you one typical font
> in each category (They are huge, at lease several MB in size).  For
> example,

Actually, I could really use as many Han fonts as you have, especially if
they are from different vendors and of different ages.  All I really need
is the fonts.cache files generated from these fonts; that holds the unicode
coverage and any OS/2 table information.  That would be a lot smaller, and
also avoid any copyright or trade secret problems.

Keith Packard        XFree86 Core Team        HP Cambridge Research Lab


_______________________________________________
Fonts mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/fonts

Reply via email to