Jeff Breidenbach wrote:
Running a serious spam filter on incoming mail is too computationally expensive (there is a lot of mail!) but it is an otherwise good suggestion. Unfortunately I know from my own inbox that a lot of spam these days is specifically designed to get around spam filters.
This seems like a fairly important problem, and one which could possibly make intelligent use of donated cpu cycles in a distributed fashion.
Potentially, you could even have volunteers to watch over the spam filter and help train it...
Personally, I use Mozilla's build-in bayesian anti-spam filters to filter out spam. Given a decent number of spam and non-spam messages for learning, they produce fairly good results.
So I have a personal smap collectin of about 25.000 letters (150 Mb) and growing, which can be used for such training. Also, there are stand-alone bayesian filters implementations (can't provide exact URLs not but I believe it's quite easy to find it).
-- == kir_at_asplinux.ru == 7551596_at_ICQ == 6722750_at_sms.beemail.ru == There are two ways of constructing a software design. One way is to make it so simple that there are obviously no deficiencies and the other is to make it so complicated that there are no obvious deficiencies. - Charles Anthony Richard Hoare
_______________________________________________ Gossip mailing list [EMAIL PROTECTED] http://www.mail-archive.com/cgi-bin/mailman/listinfo/gossip