Re: [PLUG] Hi

[sean finney <seanius@seanius.net>]

On Fri, Apr 30, 2004 at 07:57:04AM -0400, gabriel rosenkoetter wrote:
> RBLs don't actually work, especially when you let other people
> who you don't know manage them. And doing the checks is
> indescribably expensive, relative to just accepting the mail,
> especially at SMTP time, and if you take too long then, legitimate
> clients may hang up on you.

i'd disagree with that.  rbl's aren't perfect, sure, but they're a
good effort and do more good then harm if done in a well thought out
way.  by this i mean not heavy-handedly rejecting the tcp/ip
connection or refusing mail service.  i also mean not relying on
a single rbl, and not letting any single rbl be able to authoritatively
tell you what to do.  

on the mail servers that i administer, these checks are done from inside
spamassassin, and against at least a half dozen different rbl lists.
not a single one of these lists if tested positive will score enough to
identify a message as spam, but two or three probably would, and almost
certainly any other spammish characteristics that spamassassin looks
for would tip the scale if it were a close call.  i haven't been
compiling stats on the accuracy, but i have been monitoring the
aggregate amount of spam caught, and after enabling rbl checks the
total spam caught jumped by about 500%.


	sean

Attachment: signature.asc
Description: Digital signature