I’ve been having a spam problem for a LONG time, considering I use catch-alls on my domains doesn’t help much, but I usually get in excess of 1500 spam messages a day. Since I only get around 10-15 legitimate mail, that makes me go insane.
TMDA stops most of the spam messages, but alot of spammers now use email addresses at your domain or your own email address to get through spam blockers, I’ve had spamassassin running with TMDA, but it just hasn’t cut the muster yet.
The other day I was stumbling along looking for ways to mount an sftp host on a drive in windows (as I have yet to write an article about my fun with curlftpfs and sshfs on linux) and stumbled across POPfile, which is a basic bayesian filter for mail.
POPfile looks nice, but I use exchange for my email, and things that work with exchange can usually be counted on one hand …
Here is where Outclass appears, Outclass uses POPfile as a plugin for outlook, and I have to agree, after going through 2000+ messages last night, it can even tell the difference between Yahoo Groups posts and spam, it does have difficulty with those stock spam messages, but hey, it’s so far cut down the 100-200 or so spam I get in my inbox to 2-3.
If I had one complaint, it was that processing can stall on LONG lists of emails, when I first loaded it, Outclass tried to parse every message in my inbox (4000+) which froze outlook, and it never came back (I gave it about an hour to try), after I logged back in and moved the messages out, I classified them manually then moved them to Deleted Items or Inbox as appropriate, worked like a charm. (Also good fodder for training POPfile’s filters)