Help: More speedup needed

[Miki Tebeka]

Hello All,

I'm writing an M.Sc. thesis in NLP. My goal is to implement Eric Brill's
unsupervised algorithm for Hebrew (see http://www.cs.bgu.ac.il/~tebeka) 
for more details.

However on a very small set of data my application runs very slow. I did
my best to improve the speed (minimal data types, psyco ...) but it is still too
slow.

If you have the time and can't find something more interesting to do, grab
http://www.cs.bgu.ac.il/~tebeka/University/tpost.tar.bz2 and try to help me.

Read the HELP.ME file for more details.

TIA.

Miki