[Tfug] Suggestions for an email gateway
Sam Hart
hart at physics.arizona.edu
Tue Nov 25 09:12:17 MST 2003
* On 03-11-24, Mike Morrell wrote:
> Thanks Sam. It may be overkill for my pitiful little network but it
> should be fun putting it all together.
There is no overkill when it comes to spam and viri ;-)
But seriously, the components from my setup that I've found work the
/least/ (i.e., if you want to eliminate them, you probably wont notice)
are the following:
* Vipul's Razor : The idea behind it is great and all
(distributiuve P2P SPAM checking network), but the problem I've found from
snooping the checksums is that typically the SPAM distributes faster than
the checksums do on the P2P network. In other words, you get the checksum
too late to do any good. Plus, since I've set this up on a couple of
machines, I've found you get some really wierd gaps in the P2P network,
and sometimes a checksum will just fall off before it gets to a particular
machine.
* Bayesian Filters : This can work /really/ good. The problem is,
you have to be fairly technically minded to use them (you have to bounce
the message /exactly/ to the filter, not forward it or whatever-
basically, you can't mung the headers or message). Most users of the mail
systems I see people set these up for dont know how to do this (how do you
explain to a secretary how to bounce messages in Outlook?)
Also, one of the absolute best ways to eliminate SPAM is to just get rid
of all references to your email addy on the web (or just obfuscate them,
like I do here: http://www.geekcomix.com/big/contact.shtml). Of course,
that's easier said than done. Just do a search on Google for your email
to see (for example, here's mine:
http://www.google.com/search?q=hart%40physics.arizona.edu )
Plus, it doesn't help to have your email addy(s) in your sig ;-)
|
\./
--
Sam Hart
University/Work addr. <hart at physics.arizona.edu>
Personal addr. <criswell at geekcomix.com>
Alternative <criswell at tux4kids.net>
end
More information about the tfug
mailing list