I do understand the fact that as SIP has undergone lots of changes
since 2008, an understanding of the code underlying is required so
that the spell checker can be integrated into the main trunk.
I was working on a spell checker implementation based on the JOrtho
API, Damain however used JMySpell API.
After looking at both the APIs, I believe JMySpell has a better design
when compared to JOrtho. Moreover I am not sure about multiple locale
support in JOrtho which JMySpell handles pretty well.
Damain in his gsoc blog discussed the feasibility of using both the
APIs. I think he chose JMySpell for obvious reasons.
I had a talk with Emil regarding my approach and he was not sure if it
wont mess formatting. So I believe working on Damain's code would be
the best bet here! His implementation is tested and looks neat.
I would start my SIP quest soon, understanding the
design/architecture/implementation of the core classes in trunk.Being
just a student and a learner I hope the learning curve wont be too
hard on me.
I believe the core spell checking code wont need much of a change as
pointed out by Yana.
Let me clear my SIP design concepts and implement the SpellChecker
locally on my machine. I could then simply write a patch for it. I
think write access to a branch in the repository is not what I need at