[XeTeX] Tibetan fonts and XeTex

Jonathan Kew jonathan_kew at sil.org
Thu Feb 24 12:01:26 CET 2005

On 24 Feb 2005, at 10:41 am, Yves Codet wrote:

> Le 24 févr. 05, à 10:43, Jonathan Kew a écrit :
> Hello.
>> An AAT font *ought* to work, however, as the AAT architecture doesn't 
>> require any script-specific knowledge in the application; the 
>> behavior is all defined by the font tables. So I would have expected 
>> XenoType's AAT font to work, but apparently that's not the case. I 
>> don't have a copy of this font myself, however, so it's difficult to 
>> investigate further--sorry! I'll see if I can find out anything more.
> If you typeset the attached test file you'll notice that superscripts 
> and subscripts aren't correctly shaped nor placed, though it uses an 
> AAT font (STFangsong).

The Apple Chinese fonts such as STFangsong do not contain AAT tables 
for Tibetan rendering; the only AAT features supported by these fonts 
are a few Roman ligatures, fractions, etc.

>  Curiously the display in TeXShop is almost perfect: only superscript 
> "ra" is wrong. Does TeXShop use another (and better) font? I was 
> unable to check that.

Fonts like this that lack full AAT support may look better in TeXShop 
(or TextEdit, etc), ironically, because the Cocoa text view system 
provides some default positioning behavior for diacritics. It won't be 
necessarily be as "perfect" as the font designer would have intended, 
as it simply "stacks" the diacritics, centering them on the base 
character. But that's often good enough to look acceptable.

The plain ATSUI APIs that XeTeX uses, however, don't do anything like 
this; they render the font strictly in accordance with the AAT tables, 
with no additional default positioning of diacritics.


More information about the XeTeX mailing list