[XeTeX] First experiences with xetex and some bugs

Jonathan Kew jonathan_kew at sil.org
Mon Jul 5 15:55:45 CEST 2004

On 4 Jul 2004, at 3:39 pm, Bruce D'Arcus wrote:

> On Jul 4, 2004, at 9:20 AM, William Clocksin wrote:
>> Xetex should appear to add to TeX, not diverge from it.  Lots of 
>> potential users have legacy files containing escape sequences and 
>> also need to share bibtex files. There is no reason why xetex should 
>> not support escape sequences.
> Remember that the author of XeTeX doesn't use LaTeX even, so the 
> design decisions are not surprising.  It was never intended to be 
> about compatibility really.

There has been discussion on this list as to how the standard LaTeX 
accent macros can be redefined to work with Unicode fonts; in my 
opinion, "supporting escape sequences" at this level is entirely a 
matter of (La)TeX macros. It does not (and should not) involve any 
changes to the xetex processor itself.

Note that all such access to "special characters" is inherently 
dependent on the font encoding. The "standard" TeX/LaTeX commands work 
ONLY with fonts that match the expected encoding. (Note that if you use 
legacy TeX fonts such as CM in XeTeX, all this stuff should work 
unchanged.) But as they're simply macros, they can be redefined to work 
with Unicode fonts if you prefer not to use the Unicode characters 
directly in source.

> Still, I think this points out why open sourcing xetex would be 
> valuable, as it would presumably better allow interested coders to add 
> to and enhance xetex without Jonathan having to do it himself (though 
> I guess Ross has managed to do some related TeX stuff without access 
> to the xetex source).
> Bruce
> PS - Would rather see margin-kerning and other such features as well.

If the fonts have 'opbd' tables, this ought to work (I think; don't 
have an example though!).

Margin kerning is inherently font-dependent and the necessary metrics 
should be provided by the fonts. In the TeX world, we're used to 
specifying lots of font-specific details through TeX source. But as 
fonts become richer, this should be less necessary, and may indeed 
conflict with the desire to actually use the full capabilities of 
"smart" fonts, rather than simply regarding them as collections of 
glyphs with no inherent behavior.


More information about the XeTeX mailing list