There was a write-up on the front page of the weekly version of LWN this
week (http://www.lwn.net) that compared the new bogofilter with
spamassassin. In general, bogofilter (by ESR) is every bit as good as
spamassassin, which is considered about the best spam filter out there.
Spamassassin doesn't rely on the same algorithm that junkfilter and
spambouncer do. Junkfilter/spambouncer rely on experience with known
spammers, and must be constantly updated as spammers change their
tactics. This is not so much true of spamassassin.
However, spamassassin is written in perl. And in my experience,
junkfilter, spambouncer and spamassassin are slow as snails when
processing your emails for spam. I've had to wait as much as half an
hour for it to process 400 pieces of mail.
Bogofilter, on the other hand, is coded in C, and uses statistical
analysis of the mail you feed it to determine the spamness of mail. This
makes is adapt to the particular types of spam you're getting. And it's
nearly as fast as having no filter at all. You must train bogofilter,
but it doesn't take much. I've been using for a few days, and I'm
suitably impressed by its accuracy and speed. If a piece of mail has
been miscategorized by bogofilter, you can tell it that it made a
mistake. Kind of like voice recognition software, I suppose, only much
better.
Anyway, for those of you who hadn't caught this, I thought you might be
interested. Good article on LWN. Read that first.
Paul
P.S. There's a mis-doc in the man file and the README for this package
that I've emailed ESR about. If you don't catch it, ask me about it and
I'll point it out.
This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 19:39:49 EDT