Re: [exim] Re: bounce messages and their potential misuse

On 31/03/06, Jeremy Harris <jgh@???> wrote:
> Peter Bowyer wrote:
> >>>We're talking about an outbound relay sending to arbitrary
> >>>destinations, with verified senders. Callouts are a waste of time,
> >>>because it can deliver a bounce to the known sender if it's unable to
> >>>deliver a message.
> >>
> >>Which known sender would this be?
> >
> >
> > The one which it authenticated, or which an upstream trusted MTA
> > authenticated. It's an outbound relay for a known community of users.
>
> Ah, sorry; first time I've seen authentication mentioned.
> The original just said "accepting anything from a list of
> known IPs".
>
> With auth, yes, accept-then-bounce is permissable (but still
> suboptimal, I think. I prefer, as a user, an instant error
> to my mistyping a destination address. As a networking engineer
> I prefer the fewer number of connections).

Unfortunately, not all MUAs are able to deliver a good user experience
when a recipient is rejected by the MSA, especially when a message has
multiple recipients. For that reason, I prefer accept-then-bounce in
these circumstances.

It also gives the user something persistant to read, keep, and to
forward to the helpdesk for support. A transient dialogue box
containing a reject message may be suitable for technical users but is
less so for the non-technical.

Peter

--
Peter Bowyer
Email: peter@???