Mark wrote: > Still, with SA present, which also does Bayes, and to which hundreds of > volunteers contribute 'big evil' rule lists and such on a fairly regular > basis, for free, I am but left to wonder: why I would use your paid > product?
I don't want to get into too much advertising on the list, so I'm not going to talk about CanIt (I did that off-list). I just want to clarify what we're thinking of offering: A Bayes implementation is only as good as the data you feed it. While our Bayes implementation is somewhat better than SA's (we consider word pairs as well as individual words), the real benefit we provide is an up-to-date database of millions of tokens, voted on by actual people. It's the quality of our token database that really counts. It lets you use Bayes without having to go through a training period. (I was skeptical that it would work, because the conventional wisdom says that Bayes databases need to be customized per user to be effective, but empirical evidence is that conventional wisdom is wrong --- most people seem to agree on what is spam.) Regards, David. _______________________________________________ Visit http://www.mimedefang.org and http://www.roaringpenguin.com MIMEDefang mailing list [email protected] http://lists.roaringpenguin.com/mailman/listinfo/mimedefang
