Howdy folks!

I'm interested in using spamassassin for some experimental purposes... or any bayesian system for that matter.

I'm wondering how tightly spamassassin is tied into email formatting, and if I can use it for other things. Iike feeding it quotes from Shakespeare and sun tzu and having it determine who said what, or whatever. That's a bad example... my purpose is to take blobs of text and classify them as spam, ignoring all aspects of email headers (because my blob of text is not an email).

Maybe I can remove all header rules and stick to body rules? Who knows. Maybe something else does that better.

The ability to train the filter is critical here. Are there any spamassassin fanatics here that can help me on my way?