Re: [gtk-i18n-list] Unicode PUA supporting issue in gtk+/pango

[Chia-I Wu <b90201047 ntu edu tw>]

> Please let me know the situation of update cycle of HKSCS and
> HKSCS-compliant fonts. Reading the inclusion policy of HKSCS
> at
> 
> http://www.info.gov.hk/digital21/eng/hkscs/applicn.html#application
> 
> I think HKSCS will be used as a sum of temporal charset of
> the characters that are not included in Unicode (at that time)
> and pushed for future Unicode (at that time). So it may be very
> often revised. I heard of GCCS-1998, HKSCS-1999, HKSCS-2001,
> and HKSCS-2004. Nothing to say, even if the charset/mapping
> table of new HKSCS is published, the release of the font for
> new HKSCS will be delayed (especially high-quality scalable
> fonts). Therefore, even if we ignore the possibility that
> some fonts have all characters of HKSCS-2001 and a part of
> HKSCS-2004, we have to use HKSCS-2001 mapping table for the
> font that is compliant to HKSCS-2001 and not to HKSCS-2004.
> If we use HKSCS-2004 mapping table and HKSCS-2001 fonts,
> some PUA codepoints are mapped to undisplayable public codepoints,
> coversion scripts generate undisplayable string. Therefore,
> I think, still we have to care which standards the font is
> compliant...?
Yes, of course.  Let's quote some vendor's description of its product:

	All the fonts are  in Unicode encoding standard and supports the
	Hong Kong SAR Government's Chinese Common Interface and latest
	version of Hong Kong Supplementary Characters Set (HKSCS-2001). 

You never want to upgrade your system to HKSCS-2004 unless some or all
of your fonts have been upgraded to support HKSCS-2004.
> >Some characters in (Taiwan-)BIG5 are also mapped to PUA.  The `P' here
> >means Taiwan.  A codepoint in PUA may be assigned different glyphs in
> >different areas.
> >
> >Therefore, even if HKSCS contains all characters defined by
> >(Taiwan-)BIG5, a font to be used in HK can _not_ be used in Taiwan.
> >
> >(Well, it _can_ actually, but not under pango.)
> Interesting. Could you tell me the framework to use the font
> in Taiwan?
Well, if the font comes with big5 cmap (platform 3, encoding 4), and you
use it, not the unicode one, to look up glyph index, then you can use it
in Taiwan.  But this is just in theory.  I don't know if there are HKSCS
fonts having such cmap nor if it really works :)

pango can not do this because it uses unicode cmap to look up glyph
index.

-- 
Regards,
olv