I'm developing an OpenType engine for the JDK, and have been working with a
developer who's developing an Arabic font for rendering Koranic Arabic.
We've discovered a problem with the way my engine renders Alef Maksura. My
shaping tables say that this letter should only join with letters on the
right. (i.e. it has no initial or medial form) This seems to correct for
modern Arabic, but not for the Koran. (Also, the Unicode standard contains
presentation forms for the initial and medial forms of Alef Maksura
(U+FBE8, U+FBE9) which it says are for Uighur, Kazakh and Kirghiz.
The obvious way to fix this is to change my shaping tables to say that Alef
Maksura also joins on the left (and thus has initial and medial forms)
However, when I do this, Microsoft's "Traditional Arabic," "Tahoma" and
"Arial Uncode MS" fonts render the initial and medial forms of Alef Maksura
as the isolated form. (even though "Arial Unicode MS" does have the initial
and medial presentation forms)
So, I don't know what's the right way to fix this. No matter what I do, it
seems that some text will be rendered incorrectly with some fonts... Can
someone please provide me with some guidance?
IBM Cupertino Center
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:17 EDT