[Mailman-Developers] suggested improvement for Mailman's bounce processing

Brad Knowles brad at stop.mail-abuse.org
Tue Aug 8 03:44:07 CEST 2006


At 4:26 PM -0400 2006-08-07, James Ralston wrote:

>  As a list owner, you shouldn't need to care.  Mailman should just Do
>  The Right Thing.  My argument is that ignoring content-related bounces
>  is the Right Thing.

The problem is determining, in a programmatic and systematic way, 
what really is a content-related bounce and what might mistakenly 
appear to be a content-related bounce, and the converse.

Then look at what happens when you make the guess the wrong way, what 
potential additional "cost" there may be to the system for a false 
positive versus a false negative, and add some weightings to the 
situation so as to try to minimize the overall drawbacks to such a 
technique.


The SpamAssassin people do this kind of analysis on a massive amount 
of spam that they have collected over the years, when re-running 
their complete collection of rule weightings to try to find an 
optimum setting.

Problem is, it takes them something like a month to make a single 
complete run through all the rules with all the input spam, to come 
up with a given set of proposed set of weightings -- and this is on a 
large set of distributed servers, in a manner somewhat akin to 
SETI at Home.  At that point, they're ready to release a new version, 
because more rules and techniques have been introduced since the last 
version they released and the weights have also been updated, and 
they start the whole process all over again.


Now, we're not talking about something quite that intensive, but it 
could still be a pretty big affair to make sure that we're striking 
the proper balance of risking false positives versus false negatives.


As it stands today, it's just some people talking about abstract 
theory.  No one has collected any appreciable amount of bounce 
information to tell us what the real-world picture is at their site.

If you want to move this discussion beyond the theory stage, I'd 
suggest that you start collecting some data.

-- 
Brad Knowles, <brad at stop.mail-abuse.org>

"Those who would give up essential Liberty, to purchase a little
temporary Safety, deserve neither Liberty nor Safety."

     -- Benjamin Franklin (1706-1790), reply of the Pennsylvania
     Assembly to the Governor, November 11, 1755

  Founding Individual Sponsor of LOPSA.  See <http://www.lopsa.org/>.


More information about the Mailman-Developers mailing list