[XeTeX] Microtypography?

[Jonathan Kew]

On 10 May 2006, at 7:44 pm, Geoffrey Alan Washburn wrote:

> Jonathan Kew wrote:
>> On 24 Apr 2006, at 4:31 pm, Geoffrey Alan Washburn wrote:
>>
>>> 	Are there any plans to implement something like pdfTeX's
>>> microtypographic features in some future release of XeTeX?  Thanks.
>>
>> I think I failed to reply to this, being otherwise occupied at the
>> time....
>
> 	No problem.
>
>> It's certainly something that would be interesting to explore, but
>> would be a significant amount of work... I don't think pdfTeX's
>> implementation would mesh easily with XeTeX's world of Unicode, the
>> character/glyph rendering model, OpenType "smart fonts", etc.
>
> 	I think protrusion might not be too difficult,

Possibly, but it'll be very different from pdfTeX's implementation.  
In the non-Roman world, there's no assurance that the first  
*character* on a line (which is the "addressable unit" in terms of  
TeX's processing of the text) will actually fall at the visual  
edge.... glyph reordering in Indic scripts, for example, break this  
assumption. So when considering line-break positions, it may not be

Of course, we also can't express amounts of protrusion in terms of  
characters, as a given character may be rendered by very different  
glyphs in different contexts. So the whole paragraphing algorithm  
will have to be much more aware of the two-level character/glyph  
model than it is at present; right now, XeTeX works in terms of  
characters, and the details of individual glyphs are largely hidden  
by the rendering technology (ATSUI or ICU/OpenType).

> but support for
> expansion/shrinking of glyphs could be quite complex depending upon  
> the
> rendering API you have available.

Actually, that's probably the easier part, as it can be done by  
manipulating the transform matrix through which the text is rendered;  
at least, for uniform expansion/shrinking of all the glyphs in a run  
of text.

>   I can't really comment, not knowing
> all that much about the internals of either.  I have actually been
> thinking some about developing a new outline font format that would
> provide hooks for doing these and other things in a much more
> programmatic fashion,

Before thinking of anything new, be sure to look closely at not only  
Multiple Masters (which Adobe has largely abandoned) but also Apple's  
'glyph variation' technology and other aspects of AAT. There's a lot  
of cool stuff in there, at least in the design and the low-level  
implementation, but it hasn't been widely adopted either by font  
vendors or application developers.

JK