Paul Douglas Franklin of Union Gospel Mission wrote: > > SA is marvelous! I installed Kolab last summer and have been using SA > to greatly cut down on spam. (It is filtering out about 5,500 e-mails > per week for our organization.) I'm not sure which version of SA > Kolab uses, but I believe it is 3.0.x > Lately I have been writing quite a few rules to try to keep up with > the spammers. Bayes wasn't doing the trick. I find it a little > disappointing that Bayes within Thunderbird catches on much faster > than Bayes in Spamassassin, and that makes me wonder whether something > is malconfigured. > I'm the only one in the organization whose system is configured to > input into the SA bayesian filter, and I faithfully send those > messages which I have marked in Thunderbird as junk to the spam > learning folder for SA. Nonetheless, SA continues to allow some > e-mails thru which Thunderbird then automatically marks as junk. > Any suggestions? > --Paul
Are you also sending your non-spam (ham) emails to be learned by Bayes? Bayes needs to learn both ham and spam in order to work properly. -- Bowie
