Spam Filtering Project

[Jonathan Hutchins]

An article in New Scientist on spam filtering:

 http://www.newscientist.com/news/news.jsp?id=ns99992141

And speaking of spam, our published addresses in the open archive, and such
things, has the rest of the list been getting large volumes of email in
unprintable characters?  Opening one or two of these has shown that they're
full of image files that display Asian text characters.  Not only are they
annoying as spam, they also tend to fill up my mailbox due to the attached
images.  I have not yet taken the time to analyze them well enough to
develop any common feature other than the unprintable characters in address
and subject.  They all seem to come through standard .com domains so you
can't just filter .kr and have it fixed.

If somebody else is having this fun, but has come up with a working filter
I'd love to hear from you.