On 11/21/06, Clay Fandre <clay at fandre.com> wrote:
>
> How about everyone else? Anything new out there we should take a look at?
>

I had been using some form of Bayesian filtering (bogofilter,
spamassassin, etc) for a long time. It was working great until I
started getting the "poison" spams.  These are spams that just load it
up with 50% keywords to throw the stats off.  It more or less made my
training useless.  I started using greylisting (postgrey) and it cut
my spam down by at least 80%.  Once I get bogofilter re-trained I
think I'll have a good setup again for a while.

Info on greylisting: http://greylisting.org/

-- 
Jay Kline
http://www.slushpupie.com/