i just finished this challenge ( and it was kinda fun ;D ).
I used python for everything ( with xml.dom for xml parsing and cairo for the drawing part + a small class i wrote for login/downloading file/sending result )
But in the end i had to look at the picture and manualy enter all the letters so i was wondering if any of you guys did this with OCR (or something) and automated this last part too ?
Any ideas are appreciated

cheers,
easY
P.S. PM is nice too, wouldn't wanna spoil the fun for all the others
