the use of unix.wizards

[Guy Harris]

>    a)  Open derision for the person who first brought the issue up -- 
>        why, because a 'user' had the temerity to ask for something
>	natural, rational, and convenient, which only shortsightnedness 
>	in Unix utilities design might have made difficult to implement?

The point raised by several people in the discussion was that it was
*N*O*T* just "shortsightedness in UNIX utility design" that made it
difficult to withdraw mail.  At our site, a large number of users
have their own machines, which means the "unmail" command would have
to work over the network; you can't make this secure without a secure
way of determining whether a request coming from over the network has
been authorized by the person it purports to come from, and this is
non-trivial.

The analogies used to claim that this was "natural, rational, and
convenient" also were flawed; you can't say "well, I can do this with
the US mail" because you can't, in general.

And, finally, the "unmail" command would only work if the mail
message was still sitting unread in the recipient's mailbox.  It's
*not* the best solution to the problem - the best solution to the
problem would be to delay the submission of the message to the mail
delivery system (either by queueing it, or not sending it in the
first place) until you're sure you want it to go out.  (I don't think
"unmail" commands are very common; this may very well be because all
the mail system implementors out there are grues, but it may also be
because the idea isn't very good.)

The point is that not all ideas are created equal; many of them are
to a greater or lesser degree bogus.  If people don't like having
bogosity pointed out to them, that's their problem.  A definite aid
to useful creativity is to weed out bogus ideas early, so you can
spend time on the good ideas.