On Tue, 7 Oct 2003 15:41:29 -0700, Tom <[EMAIL PROTECTED]> penned: > On Tue, Oct 07, 2003 at 09:55:03PM +0000, Monique Y. Herman wrote: >> On Tue, 7 Oct 2003 16:54:49 -0400, Naitik Shah <[EMAIL PROTECTED]> penned: >> > I've got procmail piping mail through spamassassin (about that, is it >> > better to use spamc? ). From what I understand spamassassin learns all >> > the time. Is this automatic and default? Or do I need to give it a >> > folder filled with spam to analyze? >> > >> > Naitik. >> > >> >> My understanding is that you need to run sa-learn to have it learn. >> >> Something like: >> >> sa-learn --spam --mbox mail/my_mailbox >> >> you can also use the --nonspam option to train it to recognize >> *legitemate* mail. >> > > I'd like to ask a question of spamassassin users: > > Do you have to "visually scan" spamassassin logs or output folders, to > "make sure" it's doing the right thing?
I'm not sure what you're asking here. Different people will have different levels of comfort. I have all suspected spam go to a mailbox that I can then check at my leisure. SA has "mis-diagnosed" a few legitemate emails as spam, although I can't say I blame it -- some mailing lists refuse to lay off the html, caps, bright colors, etc. I have an old, old freemail address that gets ~40 messages a day, all spam. I take a brief peek and then tell sa-learn to go at 'em. So far, SA has had one or two false hits for me -- again, obnoxious mailing lists that refuse to offer a plain-text alternative. False hits bother me far more than accidentally letting a spam through here or there. > > [I'm interested in this from a philosophical perspective. Read "The > Illusion of Technique" by William Barrett and the scheduling algorithm > which broke down because people go to the bathroom at random intervals >:-)] > > -- monique Please respond to the group OR to my email, but not both. (Group preferred.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]