> On Mon, 27 Apr 1998, Bruce Bowler wrote:
> > [...] it seems that the default behaviour of 1.90 is to "allow
> > invalid".
On Mon, 27 Apr 1998, Philip Hazel wrote:
> Touche. However, "accepting invalid" isn't quite the same as "generating
> invalid".
I entirely agree with Philip on this one, but we need to make a third
distinction. A system could
(1) Accept invalid (which was what was being discussed)
(2) Generate invalid (a very bad thing)
(3) Reject valid (the ip_literal option is a good example where
one might want to do that).
As for (1), there will be cases for which we might have to accept invalid
messages and communications (such as to accomodate systematic bugs in
other's software).
However, at the risk of being ideological and moralistic, we have a
responsiblity to not do (1). Everyone on a public network as a
responsibility to do (2). But furthermore, we all have a responsibility
to the network community as a whole to insist that others also do (2).
The obvious way to fulfill that responsibility is to do reject invalid
communications, even at the short term expense of our own users.
Sites that behave badly, whether through ignorance, policy or incompetence
should not be accomodated. (It is particularly hypocritical of me to
say this, since two of our mailhubs are still potential bad guys at the
moment, but I am working on it.)
As is true for almost all of my postings, but especially worth saying
here: My views do not represent Cranfield policy, but are my own
opinions. In fact, I suspect that the opinions stated in this message are
even a minority view among our email team.
-j
--
Jeffrey Goldberg +44 (0)1234 750 111 x 2826
Cranfield Computer Centre FAX 751 814
J.Goldberg@??? http://WWW.Cranfield.ac.uk/public/cc/cc047/
Relativism is the triumph of authority over truth, convention over justice.
--
*** Exim information can be found at
http://www.exim.org/ ***
