On Wed, Dec 06, 2006 at 01:14:07PM +0000, Martin A. Brooks wrote:
> On Wed, 2006-12-06 at 10:38 +0000, Chris Lightfoot wrote:
> > On Wed, Dec 06, 2006 at 10:26:35AM +0000, Philip Hazel wrote:
[...]
> > > In our case, if our server sends a 550 for an overquota error, we want
> >
> > If you're going to start announcing temporary errors as
> > permanent, it's hardly a surprise that people are going to
> > start treating announced permanent errors as temporary!
>
> The recieving server has defined the problem as permanent. End of
> discussion. Trying to add "ahah! but if..." or "what if" to this is
> meaningless.
In the real world one frequently encounters sites which
issue permanent failures for over-quota conditions *even
though* retries shortly afterwards would succeed. There is
no point in propagating these back to the user since
presumably the user will just retry the mail themselves;
treating these conditions as permanent just wastes
everybody's time and risks confusing the user.
Yes, it would be nice if a 5xx message could always be
taken to mean ``there is no point in retrying
automatically'', but that's not a description of the world
as it currently is.
--
``Do you have any French colloquial dictionaries?''
``No, we only have French-to-English and English-to-French.''
(actual conversation with a salesperson in a Waterstones' book shop)
