On Sun, Feb 22, 2004 at 03:39:48PM +1100, Jeff Waugh wrote: > So, who's been having fun with their anti-spam tools recently? I'm still > using spamassassin and bogofilter [1] these days, but finding more and more > crap in my real inbox, thanks to all this random-text crap. Gar.
I've been using bogofilter too. I have about 15,000 messages of junk mail divided up into spam/scam/virus/bogus delivery messages. I trained bogofilter on my spam and ham (ham being all my archived mailing list mail and personal mail archives) The results were ok. I got far too many false positives. The main false positives were mail from myself at work with a single web site link -- I mail myself interesting links to read. So these look pretty spammy, although I obviously need a whitelist. I really dislike false positives, and I don't see how bayesian analysis can avoid them. I'm always going to have the occasionally valid but spammy look mail. As unfortunate as it is, I think the only way to combat spam is to make it expensive to send by default. Expensive here means more expensive than a good return return by the spammers. Say the spammers get a hit rate of selling $100 worth of stuff for every million messages, then the cost of sending mail should be at least double that rate -- it should cost them about $200 to send a million messages. Is there any mail client that integrates tightly with spam detectors? I'd like to click on a column to sort messages by spamicity. Matt -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
