--On 10 January 2007 14:38:14 +0000 Phill Harvey-Smith
<phill@???> wrote:
> Peter Bowyer wrote:
>> On 09/01/07, Phill Harvey-Smith <phill@???> wrote:
>>
>> <snip>
>> You can configure the behaviour at this point - whether you blackhole
>> them or reject them is up to you - rejecting is generally better
>> because a non-zero percentage will be genuine mis-typed addresses
>> which should be notified to the sender.
>
> However, rejecting however will still send a copy of the email back to
> the supposed sender, which for the most part will be spam, rather than a
> genuine email, so the overall effect is that someone still gets spammed.
The word "reject" isn't clearly defined in rfc2821, but in the exim-users
mailing list context it usually means issuing a 5xx response to either the
HELO, MAIL, RCPT, or DATA commands, or post data.
That leave the sending MTA with the responsibility of deciding whether to
send a reply. If the sending MTA is, in fact, a spambot, then it won't send
a reply. Of course, if another MTA (say another institution forwarding a
user's email to yours) has accepted and forwarded a message that you
reject, then it might generate a bounce-but that's their problem.
> Dunno If there's a good solution to that problem tho :)
>
> Cheers.
>
> Phill.
>
> --
> Phill Harvey-Smith, Computer Technician,
> Department of Biological Sciences, Warwick University.
> Email: phill@???, Phone: 024 7652 8385.
> "The Days' as dark as the night is long" -- U2, Ultraviolet.
--
Ian Eiloart
IT Services, University of Sussex
