Robert G. Brown wrote: > On Thu, 6 Nov 2003, Hedemark, Magnus wrote: > > >>This raises an issue with the mailing list... if it is going to be >>configured to relay mail from non-subscribers without moderator approval, >>can it at least be set up with SpamAssassin to nuke stuff like this before >>it hits the list server? > > > X-Spam-Status: No, hits=4.7 required=5.0 > tests=AWL,BIG_FONT,CALL_FREE,EXCUSE_3,EXCUSE_7, > HTML_COMMENT_SAVED_URL,NO_REAL_NAME,SPAM_PHRASE_08_13 > version=2.44 > X-Spam-Level: **** > > Alas, didn't quite make the standard SA threshold anyway. > > A fair bit is getting through to me even with SA. It's the eternal > paradox -- how to set a filter threshold to eliminate noise without > risking signal. SA tends to be conservative but it is still saving me > about 500K worth of SPAM-clearing a week. I can live with a few 'd's a > day... > > rgb You'll do a lot better with SA > 2.55. Add Bayesian to the mix and you'll have much better tagging results. RPMs are available at spamassassin.org. The AWL is probably what drops it below the threshold. (my SA 2.60 didn't catch it either, but I think I've got a whitelist set up for dulug.duke.edu) My problem is that this is the 7th message from 'alvin@xxxxxxxxx' and it has not yet been added to the "automatically reject all mail from these addresses" list. Joseph
