Author: Tim Jackson Date: To: exim-users CC: Jonathan E. Brickman Subject: Re: [Exim] Re: To smarthost on direct failure -- how?
Hi Jonathan, on Wed, 03 Mar 2004 06:35:34 -0600 you wrote:
> > No. As you should have seen in the archives before you posted to the
> > list, there is strong resistance to the idea that Exim should retry
> > sending a mail (even to a different target) after being given a 5xx
> > refusal to a direct attempt.
> Very well. I had already learned that due to this resistance, others
> have been forced to abandon Exim in favor of other products; apparently,
> I am now one of them.
Jonathan, maybe I'm interpreting your tone incorrectly, but please
consider that the reason that there is strong resistance to this idea
isn't because nobody sympathises with your problem, or that anyone wants
Exim to be crap, but because it is the WRONG THING TO DO. Just because you
*can* do it with other software doesn't mean you *should*, or make it any
less of a bad idea. If you don't like Exim, that's fine - you're free to
use whatever software you choose. However, *irrespective* of what software
you use, before ignoring well-grounded advice, you should ask yourself
whether you really should be doing the thing in question against strong
advice to the contrary, particularly given issues like interoperability
and the impact on other people's systems, which you cannot control.
Saying "oh well, screw you, I know it's a bad idea but I don't care about
the impact on anyone else so I'm going to use some software that lets me
do what I want" seems to me to be rather arrogant not to mention a bit
silly (choosing to pass over a piece of software you would have otherwise
chosen because it doesn't do something *wrong* doesn't seem to be me to be
a very good way to make good decisions, particularly in the case of Free
software like Exim where you can easily add the relevant functionality if
you so desire whilst still keeping all the other benefits). You wouldn't
be able to use a mail server in the first place if everyone put their
specific needs above the need to interoperate, because nothing would ever
talk to each other properly. That's not to say that there aren't
limitations in the standards we're working with, nor that there are never
circumstances in which measured deviations from standards are a necessary
evil, but adopting an aggressive attitude of "I *WILL* do what I want, and
to hell with the consequences" isn't very constructive.
(but hey, I'm sure you'll scoff - after all, you have a
business/organisation/life to run and don't have time for the boring
advice of some stupid techno- stick-in-the-muds - who needs their approval
anyway, eh?)