I recently ran across a Bayesian Filtering POP3 proxy written in Perl called
PopFile. Good documentation aimed at the common user and easy to use. It might be hard for me to say objectively , but I think most computer "users" could get it going very easily.
After feeding it my saved spam and other boxes, I so far have had 100% accuracy on SPAM. Other subjects (client vs cron/server message/etc) is after a couple of days of training 94%. I'm quite impressed with it's overall look, feel and performance. You can manage it via a local web server.
I've seen many references to
Paul Graham's article but not this so I thought I would bring it up.
A good product, easy to use, that solves an enormous problem most of the computing world has to deal with on a daily basis. I wonder if this might not be the best form of Perl advocacy?
-Lee
"To be civilized is to deny one's nature."