QUOTE(Mehdi Hasan @ Jul 12 2007, 12:36 AM)

Alamgir Bhai,
Why don't you announce this in the announcement section also? I know more people visit that section than here to find resources outside OmicronLab.
Google would pick omicron instead of me! (A checky way to gain!)
QUOTE
Anyway, I appreciate the good work. As the Linux version is free and run without limitations, I can place it on the third party tools section of our main site, if you need more exposure more OmicronLab. For downloading, people will be forwarded to your site/to the direct link you provide.
Let me know by PM/Mail if you need this, with a brief description.
Please wait until I fix the loading time issue.
QUOTE
I'm also working on a spell checker for Windows. The optimization is a tough issue here. I guess Alamgir bhai is using metaphone/double metaphone/phonetic based algorithms in the dictionary. One suggestion here to Alamgir bhai (if I'm not wrong in guessing), instead of building metaphonic codes (you might be calling it suggestion list) on the fly at startup, port it with the dictionary and just load with the word list. This may save some loading time.
The dictionary is a hashmap. Suggestion database is a different data structure. The algorithm of generating suggestions is very crucial. I dont build metaphonic code on loading. They are made during suggestions generation. Most of the time is spent on checking the correctness of words (as most words in a document are correct). The data structures/algorigthms are the opmial one in my findings. The problem is related with specific issues of Linux platform.
I'll have a Mac version in next week or so (if I can manage a copy of OSX) and will see.
alamgir