On Sat, 11 May 2013 16:28:16 +0200 Benny Pedersen <m...@junc.eu> wrote:
> i think David missed one more word, no share ? OK, here's my point. In 2004, I was as skeptical as most others on this list that a shared Bayes database would be useful. Then I read the USENIX LISA '04 paper by Blosser and Josephsen that described a very good experience with a shared Bayes database. So we went back to our office and implemented a mechanism for sharing Bayes data and (much to my surprise at the time) it worked very well. You are free to disagree with me, but I would expect you to disagree based on actual evidence and experience rather than just a gut feeling. I think a mechanism for sharing Bayes data would be a very useful addition to SpamAssassin. Having one foot in the free software world with MIMEDefang and another in the proprietary world with CanIt puts me in a somewhat awkward position, but I'd be happy to talk about our design even if I can't really contribute code. Regards, David.
