Last night I worked for 5 hours for basic RnD about my upcoming product, a unicode based phonetic parser for Bangla language. This parser will parse keystrokes phonetically (or you better say translitterally), for example when we will type “amar” you will see bangla “AMAR”.

I was stuck with a strange problem. But I solved that. Now conjunctions, Vowels and Consonants, Preceding Kars and Following Kars are working pretty fine. Though I need to spend some time with “Khanda To”, “Ref”, “Hasanth” and “Jo Fola”. Also “Oi kar”,”Ou kar” and “II Kar” needs to be processed with special care.

Thanks to Paul Nelson from Microsoft (Well, ya its Microsoft) and developers of Unicode Renderers for *nix distros for their unicode rendering engine.

I am releasing my script under CC/LGPL dual license from Ekushey.org, if everything goes fine.