Thomas T. Veldhouse wrote:
kashani wrote:
1. Block mail up front.
Use greylisting as it stops spam before it enters the MTA's queue.
This keeps 90% of my spam from even entering the more resounce
intensive filtering processes.
This is a very effective filter. However, it does greatly slow down
delivery of legitimate email. I found it a bit of a pain. Further,
there are those servers out there that respond to greylisting as a
bounce, so you need to specifically configure accordingly.
I set mine with a time of one minute. Hardly any spam retries so the
time really isn't important. However hotmail and the like often retry
once every minute for the first three minutes and then attempt again
fifteen minutes later. With the one minute time most people don't notice
any problems.
2. Don't use blacklists
30% false positive rate. Comapared to 1-2% for Bayesian or
Markovian filtering.
I use both. As far as false positive goes, I have had very few false
positives ... in fact, i can not think of any. But, for a corporate
setting, I would not use it, but instead leave it all to software like
DSPAM or Spam Assassin.
How do you know if you've had false positives? On a personal server you
might be able to tell, but in an office of fifty people you can't be
sure. And according to the math for every email that ends up in your
junk folder in your mail client thirty are getting bounced by your
blacklist.
The last straw for me was when some jackass listed a few hotmail
servers. So 90% of the tests worked unless you came in from a particular
set of servers. I've got better things to do than deal with someone
else's spam jihad nonsense.
kashani
--
gentoo-user@gentoo.org mailing list
