Like, I suspect, many of you, I get a lot of spam. On weekdays, I average around 400 e-mails per day, many of them consolidated digests containing numerous postings to the discussion lists to which I subscribe. The 75 to 100 spam e-mails out of that total that I'm receiving daily (the number seems to grow each month) are serious impediments to my ability to efficiently process the flood of incoming information. On weekends, the total amount of e-mail I receive decreases, but the volume of spam doesn't. Roughly three-quarters of the 100 e-mails I receive on an average weekend day are spam. This ever-growing spam deluge is the impetus behind my recent evaluation of various antispam software packages and the detection algorithms that they run. The good news is that I've found a spam-swatting technology that works well for me, and, even better, it's free. The bad news is that I had to sort through a lot of poorly implemented products before I stumbled across a mention of my eventual winner in a magazine article. By first understanding the techniques and shortcomings of these also-rans, the superiority of the Bayesian-based heir ascendant becomes easier to comprehend in comparison.
展开▼