Re: Tanween variants and Unicode

[Gregg Reynolds <gar at arabink dot com>]

Meor Ridzuan Meor Yahaya wrote:
> What I suspect is Fontlab's implementation of the GSUB table that have
> problems. I've seen Fontlab under windows (Demo version) reading
> arabic fonts, displaying some wierd GSUB table structure. I'm not sure
> how the Mac version works.  Maybe you can show me some of your work
> and I'll try to find out what is the problem.

That would be great.  I'm relatively new at making fonts, although I 
studied the TT and OT specs quite a bit.  I know that FontLab uses 
Adobe's OT dev kit, so it should work properly, but who knows.  I 
probably won't be able to send you anything till the weekend or even 
later.

Ideally, I think it would be good to package GPL fonts with textual OT 
and TT hinting files.  I expect to look into using the tools from 
http://home.kabelfoon.nl/~slam/fonts/.  In principle one should be able 
to add OT tables to any font, without needing a full font/glyph editor, 
if I understand things correctly.
> > So it looks like we really need to write an OT Service provider to work
> > with Freetype.  Anybody game?  I've started looking at the Freetype
> > code; it's very clean and well-organized, and they have a bunch of OT
> > stuff that they're migrating out of FT, since it is outside of the scope
> > of FT.
>
>
> I don't think we need to create one, Pango and ICU is out there. We
> just need to improve it to get it all right.

I'll have to look into those further, I guess.  I suspect I'm thinking 
of something much simpler, though.  I understand Pango and ICU try to 
provide a complete text mgmt system, whereas I'm just thinking of 
something like an enhancement to Freetype - take a string of chars, and 
return a string of glyphs after performing OT stuff.  Maybe return some 
datastructure that indicates the mapping from char to glyph.  If Pango 
implements something like this as a clean component then that would be 
the place to start, I guess.

-gregg