Re: devnull relayed spam?

On Nov 10, 2004, at 13:09, Rick Moen wrote:

> This decade's argument of "Please don't punish my badly administered 
> mailing-list box's MTA" seems to strongly resemble last decade's 
> argument of "Please don't punish my badly administered open relay". I 
> wonder how many people will accept the reasoning, this time?

Closing open relays is a simple, well defined, unambigous task.  Either 
you are relaying mail between third parties, or you are not.  Simply 
put, you do not want to deal with sites that do relay - hence the 
various RBLs that specifically list such hosts.

Stopping spam relaying is a very different ballgame.  There is no 
"black" and "white" mechanism to identify spam - spam 
classifications/criteria vary from one filtering program/mechanism to 
another - something classified as ham by the receiving MX may in turn 
be classified as spam by the primary/final MX.  Unless everyone in the 
world use the same mechanism to classify spam (at least for SMTP 
rejection purposes), you will end up punishing your friends, not the 
enemy - and you will end up (indirectly) generating collateral spam.

(Of course, if everyone in the world _did_ use the same criteria, it 
would sort defeat the purpose of filtering in the first place).

Finally, you ought to have a little chat with the Debian organization's 
postmaster - this is a guy who deals with hundreds of frozen messages 
(due to mail rejected by mailing list subscribers) each day.  Although 
most of these rejected messages are spam (and you could argue that 
Debian ought to change its stance on spam filtering on their servers) 
you find plenty of cases where the mail rejected by list subscribers is 
not spam at all.

-tor