Here I am thinking out loud while running through my spam log, would a perl interface to a spell checker with some sort of weighting help kill more of this junk?
i.e if there are 100 words in the body, and 75% are misspelled, add it to the score...
This comes up regularly on the SpamAssassin lists. Another approach I'm seeing lately is to use Bayes but assign high spam probabilities to tokens that haven't been seen before. So misspelled words have high spam probabilities, and spam with lots of misspelled words gets a much higher Bayes vote.
_______________________________________________ Visit http://www.mimedefang.org and http://www.canit.ca MIMEDefang mailing list [EMAIL PROTECTED] http://lists.roaringpenguin.com/mailman/listinfo/mimedefang
