Re: Font lookup ranges [was Re: Notes on Pango Xft backend]

[Steve Underwood <steveu coppice org>]

Keith Packard wrote:

> As for fonts without an OS/2 table, perhaps we could generate a heuristic 
> that could guess the tag.  I'm guessing that we could probe the fonts 
> Unicode coverage and guess which languages it was designed to support, 
> either within the Han range or outside -- find Kana and guess Japenese, 
Probably not reliable. The Taiwanese like supporting Japanese, and I 
think I have seen traditional Chinese fonts that include Kana.

> find Hangul and guess Korean.
That one probably is reliable, as far as this non-Koren reader understands.

>  Differentiating between traditional and 
> simplified chinese might be possible if we could get some help from
> someone more familiar with the differences.  Back that up with config file 
Earlier versions of Unicode had a lot of messy stuff and actual bugs 
related to simplified versus traditional. As far as I know this has been 
pretty well cleaned up in 3.2. Chinese which looks different, in other 
than purely font style ways, pretty much always has its own code point 
in 3.2. I'm sure there are still exceptions/errors, but its pretty 
clean. Of course, you still have the problem that traditional and 
simplified fonts don't generally merge well, unless they are from the 
same vendor. You still need to figure out the predominant dialect when 
choosing a font, and anything inserted from a font of the other dialect 
is not guaranteed to blend nicely. Information that explicitly says this 
traditional font and that simplified font are actually from the same 
vendor's superset font and blend perfectly could help here.

As far as spotting if a font is traditional or simplified is concerned, 
that should be easy. Although many fonts only cover a rather limited and 
variable subset of Chinese, the majority of the characters that got 
simplified are common ones, so they are always present. Spotting one 
simplified only character does not a simplified font make. Spotting a 
substantial number would. Of course, the font may have broad 
simplified+traditional coverage, so a similar probe of a list of 
traditional only characters should give an effective conclusion as to 
whether the font is simplified, traditional or combined.

The growing popularity of traditional Chinese in the PRC (yes, its 
making a comeback), and a more relaxed political environment, probably 
means more combined fonts will appear, and could become the norm.

Regards,
Steve