On Mon, 16 Aug 2004, Ryan Malayter wrote:
I've seen that page, but as I mentioned, there's no analysis of the
false-positive ratio. A very high false-postive rate, in my experience,
is what makes IP-based blacklisting unworkable. Yes, a good blacklist
stops a lot of spam, but it also stops a significant amount of ham.
We got a lot of false positives with the SpamCop list, and other BLs are
bound to be worse.
The SpamCop list is one of the worst for false positives. We get very few
complaints from using the MAPS RBL+ and the Spamhaus SBL-XBL, and they are
both quite effective (accounting for about half of our SMTP-time
rejections). The following graph is fairly interesting:
http://canvas.csi.cam.ac.uk/stats/ppsw/img/dnsbl.15.gif
The y axis is in messages per second. A week ago I swapped the order of
the blacklists in my configuration so that the RBL+ is checked first.
Tony.
--
f.a.n.finch <dot(_at_)dotat(_dot_)at> http://dotat.at/
BERWICK ON TWEED TO WHITBY: WEST OR SOUTHWEST 2 OR 3 INCREASING 3 OR 4. FAIR.
GOOD. SLIGHT OR SMOOTH.
