Re: Arabization, techniques and problems

[Gregg Reynolds <gar at arabink dot com>]

maysara a wrote:
> Salam all,
>
> Im trying to initiate an open source community in my
> university, and i need to intriduce students to open
> source and to contribute to already existing software
> and documentations and other.

You might want to take a look at kdiff3 
(http://kdiff3.sourceforge.net/).  Late last fall I was looking for a 
diff tool that would work for Arabic, but I couldn't find one.  Some 
tools can do the diff but can't display it properly.  So I contacted the 
developer of kdiff3 and he agreed to add Arabic support, just because he 
found it an interesting challenge.  He knows no Arabic.  In return, I'm 
supposed to test it and also translate the docs to Arabic.

The current status is that he has implemented utf-8 support and RTL 
display support, but not yet shaping support (contact him to find out 
for sure).  He is looking for a chunk of existing code to use for 
shaping support.

Maybe your students could take responsibility for implementing shaping 
logic.  This is actually kind of an interesting project, because you 
have to figure out a good way to show differences involving diacritics. 
 You also need to decide if you want to do anything Arabic-specific; 
for example, "ktb" and "kataba" are different orthographically, but 
equivalent semantically.  Should a diff engine be aware of such 
considerations?  I think it would be useful.  Maybe there should be a 
switch that means "ignore diacritics and show mean only differences in 
huruuf".  It's a design challenge as well as an implementation challenge.

Since the developer knows no Arabic, feel free to have your students 
contact me if they want to bounce around ideas.  They can write me in 
English or Arabic.

As for testing and translating, I'm afraid I've only done a little 
testing and no translating.  If your students would like to help, they 
could create a suite of test cases.  Every combination of 
joining/nonjoining characters with and without diacritics needs to be 
tested to make sure the shaping and display logic is correct.

Help with translating would also be appreciated; in fact since I'm not a 
native speaker of Arabic your students would undoubtedly do a better 
translation than me.

Also consider that a diff tool is an essential part of a developer's 
toolbox, and very useful for ordinary writers.

Hope that helps,

gregg