On Fri, Jun 20, 2008 at 09:34:40AM +0200, Hans Hagen wrote:
> Idris Samawi Hamid wrote:
> > On Thu, 19 Jun 2008 18:23:05 -0600, Khaled Hosny <khaledhosny@eglug.org>  
> > wrote:
> > 
> >> Arabic index entries are all listed under "unknown" instead of its
> >> respective Arabic letters. I'm not sure if this is a bug or a
> >> misconfiguration from my side. See the attached example.
> > 
> > We need to include arabic-farsi-urdu etc. databases in the distro. If Hans  
> > can tell us what file to emulate/edit etc....
> 
> first we need to discuss the logic ... say that we have a sequence of 
> chars ... do we need to erase the vowels? etc

Erase vowels as in not counting them? Then yes we should only respect
full letters. We might need also need to strip the Arabic definite
article "ال", but this will be tricky since there are words that start
with it. May be we better have syntax like \index[a]{entry} where this
entry will be under "a", or we already have this?

Regards,
 Khaled

-- 
 Khaled Hosny
 Arabic localizer and member of Arabeyes.org team