[XeTeX] Syriac Script

[Jonathan Kew]

On 19 Oct 2005, at 10:07 am, Malte Rosenau wrote:

>
> <xetex at tug.org> schrieb am 18.10.05 23:12:37:
>
>
>>> Would this work out of the box with XeTeX if those AAT tables were
>>> available?
>>>
>>
>> Yes, it would, as AAT uses a model where *all* the contextual
>> rendering behavior is built into the font tables and executed by a
>> generic engine, rather than relying on separate engines for each
>> complex script, as OpenType does.
>>
>
> So my imaginary Syriac AAT font would work in other OS X  
> applications too?

Yes, it should work in those that properly support right-to-left  
text, just like the AAT Arabic fonts that Apple ships (or those we  
offer at <http://scripts.sil.org/arabicfonts>).

> Sounds promising. I will ask George Kiraz, co-author of the Meltho  
> fonts, then
> about an AAT version. According to his website
>
>   http://www.bethmardutho.org/support/meltho/faq/
>
> they were expecting syriac support for OTFs to become part of OS X  
> at some
> point, but this is obviously a misconception: AAT is the way to go,  
> or each and
> every application has to come up with its own syriac fontengine.
>

Yes and no. I think the AAT model is superior in many ways,  
particularly in that it is more general and does not require script- 
specific knowledge built in to the software, only a general-purpose  
engine. However, it does have shortcomings too.

Anyhow, it's clear that OpenType is here to stay, and with Tiger,  
Apple has begun to implement OT layout in the ATSUI engine alongside  
AAT. Right now, they only support some Latin-script features, but if/ 
when they implement shaping for the various complex scripts, then  
such OT fonts will also begin to work system-wide. So I think  
George's expectation is reasonable; it just hasn't happened yet.

Of course, I have no idea what Apple's timescale for delivering non- 
Roman OpenType layout might be. In the meantime, AAT fonts would  
provide a way forward, if you can get them built.

Regards,

Jonathan