On Thu, 2004-12-09 at 06:05, Ow Mun Heng wrote: > On Thu, 2004-12-09 at 01:07, Scot L. Harris wrote: > > > > > What you could do is implement spamassassin, the new version 3.0 has > > SURBL support which I have heard is very very good. But you will still > > be processing each message. > > Now, this would be interesting. I'm on Spamassassin 2.64, and it uses > bayes. What's this SURBL you talk about? I would head over to > spamassassin's website but I don't have I-Net access. SURBL is a realtime block list which spamassassin uses to check the URLs that are normally embedded in the spam messages. Reports I read awhile back indicated great success in tagging spam. SURBL replaced a number of the rule sets that were previously maintained on the rulesemporium site such as bigevil and another one that I can't remember. Those extra rulesets were getting so big that they were consuming tremendous amounts of memory. Bayes is very good and is used in combination with SURBL and the other rules. -- Scot L. Harris webid@xxxxxxxxxx Depart not from the path which fate has assigned you.
