Hi Jason,

> I could be having some email problems. It's very frustrating to
> me that I have to delete about 600-800 spam emails a day.
> I have yet to find a spam blocker that does a perfect job.
> Unfortunately I can be a bit sloppy at times and hit delete in
> such a way that the keyboard buffer gets a bit loaded.

> Spam is truly a huge problem for me and I am about tempted to
> convert over to a white-list only email system where if you
> are not on my approved list you will get an email back with
> instructions on how to add yourself to my white-list.

> Of course, I'll need to figure out how I can detect when someone
> is being spoofed too, otherwise, spam can still get
> through that system.

I can recommend PopFile (http://popfile.sourceforge.net). PopFile
is an automatic mail classification tool. AFAIK, it works with
Bayesian filtering. You train it for one or two weeks and then it
works very well. I am using it for mor than a year now and it works
like a charm.

Here are my statistics for 2004:

11.886 messages classified since Jan 10th, 2004
9 classification errors

This results in a accuracy of 99,92%, which IMHO is quite impressive.

The great advantage of PopFile is its flexibility. It works as a proxy
POP3 server, so that you can use it for all your mail accounts and it
does not matter which mail program you use. You can even change your
mail client and the PopFile configuration can remain the same.

If you are interested in PopFile or you need help configuring it, feel
free to contact me. I could perhaps give my classification database to
you so that you do not have to go through the first hard training
phase with your PopFile. My classification database works especially
well in differentiating between technical support messages and
Viagra advertisements and that stuff ;-)


Christian