If the spam storm gets too heavy on haskell-cafe we can switch to moderator approval as on the main Haskell list. We have the spam assassin available at haskell.org but I've been reluctant to turn it loose on the lists. Maybe the time has come ... John

On Tue, 25 Mar 2003 14:59:20 +0100, Peter Ross wrote:

On Tue, Mar 25, 2003 at 02:10:56PM +0100, Jerzy Karczmarczuk wrote:

...

Or tell me how to *effectively* filter all that away. Or, I'll go mad and I strangle some of my students. Anyway they deserve it.

I use bogofilter (bogofilter.sf.net). I am currently getting 98% accuracy in filtering my email after about two months training and as far as I know only a few false positives early in the training.

If you are on a windows box, then I would suggest looking at Popfilter (popfilter.sf.net). ^^^^^^^^^^^ This url is just a blank directory.

On Tue, 25 Mar 2003, Jerzy Karczmarczuk wrote: (snip)

Anyway, I am willing to spend a part of this money on your behalf. If somebody has any idea how to empoison, strangle, shoot, electrocute or burn alive this annoying bastard who proposes regularly to everybody on Internet all that financial transactions with Nigeria, Congo, etc., please contact me. You don't even need to do the dirty job. Just show (snip)

Rich Kulawiec, on Dave Farber's Interesting People list, recently said interesting things at http://tinyurl.com/84dm Personally, I like SAUCE at http://tinyurl.com/84dp but it's not for everyone, or even most people. (Needs to be attached to exim.) ObHaskell: I have bought my own domain and have got to the point where soon there will not be a public permanent e-mail address for me. I plan to write software to help keep public addresses rolling over, keeping track of which "private" addresses have been given to whom, what is whitelisted and blacklisted, etc. I worry a bit about character IO being slow, but maybe Haskell might be a good implementation language for this system. Perhaps I'll have to look out for a library for parsing e-mail messages. -- Mark

Peter Simons writes:

Mark Carroll writes:

...

Perhaps I'll have to look out for a library for parsing e-mail messages.

I have written a set of parser functions for RFC 2822 messages, which should do exactly that. It's not finished yet, but if you're interested in using the code (and in providing feedback), I'll gladly give you -- and anybody else, for that matter -- a copy.

..and if anybody cares, I've cooked up functions that build word frequency tables, and compare them using a Bayesian approach (more or less like Paul Graham's "A Plan for Spam"). Haskell, of course. Need a bit of tweaking, but works reasonably well, last time I looked. -kzm -- If I haven't seen further, it is by standing in the footprints of giants