Fwd: Arabic dictionary for hunspell
Mohamed Kebdani
بعدما عرضت مشكل الاقتراحات الخاظئة عند التدقيق على مطور برنامج هانسبل،
توصّلت بالإجابة التالية:
---------- Forwarded message ----------
From: Németh László <nem...@openoffice.org>
Date: 12 sept. 2007 08:09
Subject: Re: Arabic dictionary for hunspell
To: Mohamed Kebdani <med.k...@gmail.com>
Cc: Øå taha <taha_z...@gawab.com>
Salam Mister Mohamed Kebdani,
Some Unicode related bug was fixed in Hunspell. I hope, the last
version will work fine with your Arabic dictionary. You can check
it using the source (http://hunspell.sourceforge.net) or send me a
detailed example with the detected problems and I will check it.
There are other efforts to develop an Arabic dictionary for Hunspell.
Mister Taha Zerrouki has developed an extended Hunspell for special
Arabic extensions, too. IGNORE feature is based also on his work, and
he has a deep understanding of the Arabic related problems. I send him
this mail, too. Mister Taha Zerrouki sent me a detailed summary about
the Arabic spell checking dictionary problems that must be solved.
OpenOffice.org in Fedora uses an Hunspell Arabic dictionary based on
Buckwalter's Arabic morphological analyser:
https://bugzilla.redhat.com/show_bug.cgi?id=212984
I think, a dictionary, like your one specialized for spell checking would be
much better for Arabic word processing.
If you have any result for testing, please extend the Arabic section of
Hunspell OpenOffice.org Dictionaries Wiki page
(http://wiki.services.openoffice.org/wiki/Dictionaries).
Thank you for your kind words.
Best regards,
Németh László
2007/9/9, Mohamed Kebdani <med.k...@gmail.com>:
> Salam Mister Németh László
>
> My name is Mohamed kebdani. I am not developer, just advanced user.
> I am working to provide Arabic dictionary for hunspell. The dictionary
> is "almost finished". After 18 months, I did not finish it yet because
> Arabic language is very complex. The dictionary is in beta-2 version:
>
> http://groups.google.com/group/ayaspell-dic/files
>
> Well,
> If I am using OpenOffice.org with Hunspell 1.1.2-1, I have not any
> problems. With hunspell 1.1.5-6 (openoffice 2.3.0, Ubuntu Gutsy) the
> IGNORE feature works fine but I have error messages like:
>
> UTF-8 encoding error. Missing continuation byte in 0. character position:�ي
> This UTF-8 encoding can't convert to UTF-16:��ى
>
> And the words in suggestion list are *wrong* sometimes mixed with good
> words at other time. Can you help me to resolve this problem in OOo?
>
> Thank you for the good work you are doing and sorry for my bad
> English. my second language is French.
>
>
> Mohamed Kebdani
>
Mohamed Kebdani
Good suggestion ;) Thank you very much Mister László!
I tested hunspell 1.1.2 and the result is good :). In the attached
file (bug.odt), you will find the details of the test.
For information, me and taha Zerrouki work together.
On another side, the TIM-Buckwalter's dictionary is heavy , contains
many errors, and it is based on a corpus whereas Ayaspell-dic is based
on the Arabic dictionaries.
Mohamed
2007/9/12, Németh László <nem...@openoffice.org>: