[84810] in North American Network Operators' Group
Re: 209.68.1.140 (209.68.1.0 /24) blocked by bellsouth.net for SMTP
daemon@ATHENA.MIT.EDU (Suresh Ramasubramanian)
Mon Sep 26 00:08:41 2005
Date: Mon, 26 Sep 2005 09:36:57 +0530
From: Suresh Ramasubramanian <ops.lists@gmail.com>
Reply-To: Suresh Ramasubramanian <ops.lists@gmail.com>
To: Sean Figgins <sean@labrats.us>
Cc: nanog@nanog.org
In-Reply-To: <20050925213127.I84280@mail.celicas.org>
Errors-To: owner-nanog@merit.edu
On 26/09/05, Sean Figgins <sean@labrats.us> wrote:
> And if the customer specifically requests that YOU do not filter his
> email, or set up a system that allows him to see ALL email, even if ti is
> tagged as spam?
sell the customer a colo box or a virtual private server and have him
do whatever he wants with it
commodity / customer mailserver operations do involve filtering
> provider, and they may all be very true, but it does not change the truth
> that a single false positive can ruin a business.
If you filter spam, return clear bounce messages that show why the
filtering was done. Ideally return a url in the bounce message that
links to a clear explanation + tells you what to do about it. And a
response mechanism to handle false positive reports, that addresses
these ASAP.
Just for example (and never mind the content .. 127.0.0.2 is a generic
address thats inserted in most blocklists) -=20
http://spamblock.outblaze.com/127.0.0.2
Bad spam filtering is what gives all filtering a bad name ... what I
would call the Wile E Coyote school of spam filtering. Like a trained
mining engineer can use a fused bundle of dynamite to blow a hole in a
rock - he'll blow up just what he wants to blow up, nothing else.=20
Give Wile E Coyote that dynamite and ask him to blow up the roadrunner
.. you know what happens next.
-srs
