Howard Goldstein <hg(_at_)penny(_dot_)n2wx(_dot_)ampr(_dot_)org> wrote:
Before we start in on another round of "I'm calling the FBI" is there some
way to use these constant spammish reminders in a good way, that is to
force us to look seriously at how mailing list design eases the door open
for the garbage and makes tracing it difficult? A few quick thoughts (I'm
not married to either of them):
o - outbound remailings ought to include the Received: headers (CERN <?>
lisproc does this)
No, they should not. Received: fields are being used to avoid mail loops.
If you include Received fields that were on the mail already, it is not
unlikely that the mail will not reach some of the subscribers.
Besides, it would be a waste of bandwidth, and perhaps giving away
sensitive information (in some applications).
In order to easy tracking spammers and to resolve mail problems SmartList
(just a random mailinglist manager which happens to run this list :-),
from the start, provided the ability to query the archive server for
recent postings. The archive *always* contains the mail messages as
they arrived at the mailinglist before *any* munging took place.
o - postings containing certain 'trigger words' (yoshio, "Magazine
Sub","natvan") maybe ought to be gated over to the list owner for review
before passing them along?
After this list had been hit a second time, I installed a rudimentary
filter, but I picked the wrong trigger words. So this third spam came
As a matter of fact, since the second spam, I've been sufficiently annoyed
by e-mail spams in general that I'm now actively gathering past and future
e-mail spams (preferably in a pristine state, the way they were *before*
entering the mailinglist, i.e. like from a SmartList archive), in an effort
to create some procmail filtering recipe(s) that can reliably detect most
existing and future spams.
Stephen R. van den Berg (AKA BuGless).
Climate is what you expect. Weather is what you get.