Vernon Schryver wrote:
From: Andrzej Filip <anfi(_at_)Box43(_dot_)pl>
...
This year I had to solve the following email delivery problem:
* the receiving site send to /dev/null messages classified as spam
* the sending site was some time ago (4 month+) an open relay for a
short period
* the receiving site used RBL with "no retests" policy, so the sending
site was listed "for ever"
[It took a few days to be detected by sender and recipient and a few
days to fix]
Are you ready to defend RBL/anti-spam filtering use in such way ?
...
YES!
When you agree that the recipient is the only party with standing to
determine what is spam and what to do about it, you necessarily also
agree that the recipient can be an idiot and use foolish filtering
policies.
If the intended recipient of your mail does not want it, there is
no problem unless you continue to try to send it. Whether your
intended targets are rejecting your mail because it is spam or
because they are idiots, it is all their decision.
It is not that simple. Some subtle consequences are hard to predict even
for non idiots. Postmasters do make mistakes too.
I like self-healing systems. Even if the anti-spam system seldom makes
mistakes then we should give people a chance to correct it. Redirection
to /dev/null is like a trial without the right to appeal with instant
death penalty [happily no humans are killed].
BTW Have you heard about "friendly fire" ?
Do you suggest that in every case an idiot had to be involved ?
It is none of your business whether the targets of your mail pick
kooky and silly DNS blacklists (there are plenty of those) or they
pick ISPs that choose unreliable hardware and software and loses lots
of incoming mail (there are also plenty of those). If your message
absolutely must get through, then hire a courier or force the targets
of your email to hire ISPs that will accept it.
Vermon, I am not a spam lover. Spam/email ratio in the mailboxes I have
used for usenet and mailing lists is 10 or more. I know how badly anti
spam filters are needed. I know that "good old days" of the best effort
delivery of EVERY piece of email are over.
I am simply not ready to sacrifice email reliability for spam fighting
especially when it is not necessary.
I do not deny you right to refuse to take over responsibility (5??
reject) for delivery of any message even because "the phase of the moon
is wrong". Once your server takes over the responsibility for delivering
the message (250 "after the final dot") it MUST deliver it or hand over
the responsibility to another server or send back the bounce message
even if it decides that it made a mistake and accepted message later
classified as spam.
You can reject the messages after "RCPT TO:" so even the blacklisted
"spammers" can contact postmaster and abuse. You can reject the messages
in reply to "the final dot" so when the content is wrong you do not take
over the responsibility. In both cases server sending you the message is
responsible for generating the bounce. For me 250 after the final dot is
like a signing yet another implied contract described by the RFC.
Please do not fight me, try to fix erosion of email reliability problem.
I do not say that what I propose (always notify the sender) is the only
possible solution but it is the most obvious choice. If you do not like
it then propose something else e.g. daily report to the intended
recipient (easy if you reject in reply to "RCPT TO:") - I do not like
it but is better than no notification at all.
For me /dev/null redirection of messages "classified as spam" breaks RFC
(at least spirit of it).
If we decide that email reliability may be sacrificed without very good
reason than sooner or later email as we know it will face the fate of
mechanical calculators. Loosing war against spam will bring the same fate.
--
Andrzej [pl>en: Andrew] Adam Filip http://www.polbox.com/a/anfi/
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg