FYI...
(strange! the msg didn't get to the dev list.)
Howdy everyone!
Its been quiet on here for sometime...Anyway, I thought I'd chip in make
some noise :-), post some updates and an action plan, and see what
others think.
I have had some play around with:
a) UIMA (
http://uima-framework.sourceforge.net/)
b) GATE (
http://gate.ac.uk/) (not much with this!)
c) ProgramD
(a) and (b) provide for a fairly advanced text analysis and are complex.
While the scope of a project like FAQBot could only be limited by
imagination, IMHO we should start with something relatively simpler and
see how things evolve. I believe it would be a good learning exercise
and also help make rational decisions on future directions for FAQBot.
I therefore +1 for ProgramD - what do others on this list think?
A heads up on project organization:
I have set up the SVN trunk for FAQBot development. I have classified it
into two parts: sandbox and main. Some main projects are set up in the
main area but I imagine the development of new stuff to take place in
the sandbox and then graduate to the main area (thoughts?). We also need
and appreciate some help setting up stuff with Maven2 - any volunteers
who can review what is there currently and make suggestions?
Todo:
a) We need to chalk a scope for the first iteration. Please chip in with
ideas here.
b) Subsequent to (a) , have some mulling around design ideas.
c) Get started with code construction for the first iteration.
From top my head, programD uses XML-based AIML sets, so the challenge
is going to be able to create specific AIML sets quickly and intuitively
rather than have to hand-feed everything - any ideas on how we could do
this?
Cheers,
Rahul
