[XeTeX] Syriac in Polyglossia

François Charette firmicus at ankabut.net
Thu Jul 24 17:26:17 CEST 2008

Gareth Hughes wrote:
> It's good to know exactly how polyglossia goes about picking its font. 
> How does this work if one has a number of fonts that could be used; 
> does polyglossia just choose the main font if that works, or 
> first/last other to be declared that fits the bill?
It only tries two fonts: first it looks whether \<language>font exists, 
else it tries the currently active font. NB: For some languages and/or 
scripts (but not yet Syriac), it is also possible to define 
\<script>fontsf and \<script>fonttt for sanserif and monospace. This is 
not yet fully implemented, and probably not yet documented.

> No, Syriac abjads are not limited to 499 (taw+sadhe+teth). The largest 
> unmodified numeral, taw, is 400, but the letters can be combined for 
> higher values. Some people would write taw+qoph (400+100) for 500, but 
> the classical system places a dot (possibly U+0741) over nun (50) to 
> increase its value tenfold. This system can then represent numbers to 
> 999 (sadhe+sadhe+teth), then 1000 is represented by alaph with a 
> diagonal stroke (U+0748) below it (and beth for 2000 and so forth). In 
> practice, these dots and strokes are omitted as a small value before a 
> large has to be modified to make sense. Thus, 2008 is often written 
> beth+heth without any modifying marks; anyone knowing the system will 
> know that the beth has to stand for 2000 rather than 2. To complicate 
> things, so very old manuscripts use a variant of the old Aramaic 
> numerals.

Thanks a lot for that information. I have modified gloss-syriac.ldf 
accordingly. Now \abjadsyriac supports numbers up to 9999 (I'll send you 
the file off-list). The update on CTAN should happen this week-end.


More information about the XeTeX mailing list