On 1/26/99 4:34 AM, Norbert Bollow <[EMAIL PROTECTED]> wrote...
> does anyone know how AOL's spam filters work? (Could it be that
>in some circumstances perfectly legitimate messages might be silently
>discarded, with no non-delivery notification or anything?)
Below is a message I just sent to L-Soft's [EMAIL PROTECTED]
only yesterday. It answers your question.
AOL has VERY strict rules on what it will accept in a message, in an
attempt to weed out UBE[1]. If the headers have even minor problems, they
will be relegated to /dev/null without warning.
In my years as working AOL support, both officially and unofficially,
I've run across a number of common problems:
* Faulty or missing domain in From or Reply-To headers (including
spam
blocks, or just listing the username without a FQDN).
* Mail servers in the Received headers which do not respond to
reverse
DNS lookups properly.
* Messages which pass through a *common* server that doesn't
originate
from the same server's dialup (ie: PSInet).
* Messages which are relayed through a mail server hosted on a common
dialup (ie: the aforementioned PSInet).
Additionally, certain mailing list servers try to deliver to too many
addresses in one connection for AOL's tastes, making AOL believe it is an
attempt to spam. However, I do not believe LISTSERV has ever been known
to trip this particular filter.
That covers 90% of the problems I run across which are actually AOL's
doing. Again, I don't know if it has anything to do with the problem you
are reporting, but it's worth checking into. If you are convinced that
this is AOL's fault, and want to perform further tests, contact me
off-list.
[1] Considering AOL members still get hit with a *massive* amount of UBE,
it frightens me to wonder just how much they'd be getting if it weren't
for these measures.
--
Adam Bailey | Chicago, Illinois
[EMAIL PROTECTED] | "Logic is the art of going wrong with
[EMAIL PROTECTED] | confidence." - George Bernard Shaw
Finger for PGP | http://www.lull.org/adam/
