I don't know if this has been discussed before, but a lot of the SPAM that I get now is "spammy" in the text/html part, and "bayes-unfriendly" in the text/plain part.
It seems to me that if you could strip the HTML out of the text/html part, and some how compare it with the text/plain part (reuse the Bayes engine perhaps?), that if they are grossly *different*, this would imply SPAM - as all valid use of multipart/alternative tends to be that the text/html and text/plain are just different marked-up versions of the same content... Has this already been discussed to death - and I just missed it? -- Cheers Jason Haar Information Security Manager, Trimble Navigation Ltd. Phone: +64 3 9635 377 Fax: +64 3 9635 417 PGP Fingerprint: 7A2E 0407 C9A6 CAF6 2B9F 8422 C063 5EBB FE1D 66D1
