[XeTeX] paper available

[Jonathan Kew]

On 17 Apr 2005, at 5:36 pm, Will Robertson wrote:

> On 11 Apr 2005, at 5:57 PM, Jonathan Kew wrote:
>
>> This is a paper about the creation of XeTeX, aimed at a Unicode 
>> technical audience, rather than a TeX audience, so it's probably of 
>> limited interest to many users.
>
> Hey Jonathan,
>
> In here you say that if a font supports both OpenType and AAT 
> features, OpenType is used by default but it is possible to choose. I 
> haven't seen anywhere how to actually do this, and I believe that the 
> Mac OS X-bundled Japanese fonts indeed are such fonts (and I'm trying 
> to investigate them). I can think of three possibilities:
>   1.  I am blind;
>   2.  you haven't documented it; or
>   3.  you haven't implemented it yet
> I would guess #1 on a pessimistic day, #2 after a moment or two's 
> thought :)
> Can you enlighten me?

Well, without looking back to check, I'd say #2 is the likeliest! 
Though I think if you read the paper closely, it comes up in one of the 
examples. Or maybe that's in the slides I used.

Anyhow, the answer is: if you add the qualifier /AAT to the font name, 
the ATSUI renderer will be used (regardless of what tables are actually 
present in the font); and if you add /ICU then the ICU OpenType-based 
renderer will be used. (Some day, of course, there might be other 
options, especially on other platforms where a different native text 
renderer may be available.)

These modifiers are like /B, /I, and /S=##, in that you use them on the 
end of the font name, *before* any colon-delimited list of feature 
settings. The difference is that slash-separated modifiers can change 
which real outline font (or rendering engine) will be loaded, so they 
are required before the renderer can actually be instantiated; whereas 
the features after the colon are options that are passed to the 
renderer/font once we've decided exactly which to load.

JK