On 22/07/2010 16:29, Eray Aslan wrote:
On Thu, Jul 22, 2010 at 06:52:22PM +04
If the amount of backscatter is small, do not change behaviour. But
accept the fact that (prepare for) you might get blacklisted in the
future.
If that is not acceptable, stop forwarding mail to domains that you do
not control.
In any case, preventing bounces is not a good solution. You will be doing
a disservice to your customers.
Eray makes some good points here. But I think this is worth amplifying
and restating.
"Backscatter" does not mean the same thing as "bounce messages". What it
means is "incorrectly routed bounce messages" - that is, bounces which
go to people who have had their email address forged by spammers, rather
than bounces which go back to the person who actually sent the email.
Some bounce messages are unavoidable, and necessary. They are an
integral part of the way that email works. So, if you find yourself in a
situation where you need to send bounce messages, then you should send
them. That means being prepared to take the risk of sending a small
amount of backscatter occasionally. The important thing is to minimise
the number of times when bounces are necessary, rather than trying to
eliminate even the necessary ones. You can do that by means of several
tried and tested methods, such as recipient verification. It isn't
necessary to prevent every bounce, merely to reduce their numbers to
trivial levels.
In particular, if a bounce is caused by a downstream MTA rejecting an
email because it's too large, then that's an unavoidable bounce. But
it's also unlikely to be backscatter, because most spam is fairly small
and rarely gets trapped by size filters. Size-related rejections are
almost always caused by real, live humans sending their friends and
colleagues unnecessarily large files. So these senders need to see the
bounce, in order to learn the error of their ways :-) It isn't helping
anyone if such bounces are suppressed rather than passed on.
Mark
--
http://mark.goodge.co.uk