On Fri, Sep 05, 2003 at 02:17:51PM -0500, Kirk Strauser wrote: | At 2003-09-05T15:56:15Z, "Derrick 'dman' Hudson" <[EMAIL PROTECTED]> writes: | | > IMHO there isn't really a need for whitelisting not to mention it is prone | > to false positives particularly when the spammers have direct access to | > the whitelist. | | SpamAssassin has blocked mail from: | | The EFF | Bruce Schneier, ie the periodic Cryptogram | My ISP (my quarterly invoices) | A computer store I was buying parts from | | SA is great, but certainly not perfect. That's where whitelisting comes in.
Instead of whitelisting, which as noted above is prone to being fooled, try the bayesian classification. Recent versions of SA include an implementation of the algorithm as described at http://spambayes.sourceforge.net/background.html. FWIW I've been using the bayesian classification for a while now and haven't had any "ham" marked as "spam", and very rarely see "ham" marked as "unsure" (and those are usually from a mailing list and not terribly important at that). I also haven't been seeing "spam" marked as "ham" in quite a while either. -D -- Your mouse has moved. You must restart Windows for your changes to take effect. http://dman13.dyndns.org/~dman/
pgp00000.pgp
Description: PGP signature
