[XeTeX] Re: [MacTeX] Re: A Zapfish request (was Re: [] XeTeX)

[Jonathan Kew]

On 14 Apr 2004, at 3:57 pm, Ross Moore wrote:

> Now that Jonathan is on-board, he can extend \char experimentally,
> just for XeTeX --- unless there is already an accepted way
> used by Omega, which then should be employed here too.
> Something like  \char U+xxxx for a specific Unicode code-point ?

xetex should accept \char with values up to 65535 / "FFFF.

>
> But this wouldn't cover alternates.

Indeed. The character/glyph distinction needs to be very clear in 
people's minds here.

> Can a font like Zapfino be relied upon to have a natural ordering
> for its glyphs; e.g. an encoding vector covering the whole font?

No. Well, there is a "natural ordering" in the sense that the glyphs 
are stored in a defined order, and indexed from 0..n-1. But that order 
is essentially arbitrary, and may not bear any relation to any ordering 
that would make sense to a user.

> Or are the glyphs identified only by name, with encoding vectors
> just a derived concept for use with embedded subsets ?

In TrueType fonts, they're identified primarily by glyph ID, which is a 
number from 0..n-1, for a font with n glyphs. They normally also have 
names (which would be used in building encoding vectors, if the font is 
translated to PS).

The character map (think "encoding vector") in the TTF maps Unicode 
characters to glyph IDs, but there may be (and in fonts like Zapfino 
there certainly are) many glyphs that are not directly accessible this 
way, but only via the AAT tables that operate on the sequence of glyphs 
*after* the initial character-to-glyph mapping. This is all primarily 
under the control of the font developer, who chooses what options to 
create and how to expose them to users.

Jonathan

_______________________________________________
MacTeX mailing list
http://tug.org/mailman/listinfo/mactex