On Fri, 10 May 2002 12:41:23 -0400, Pete Maclean wrote:
> By the way, while thinking about this, a plausible reason did occur to
> me.  That is that it could be very difficult to implement a multi-message
> MOVE in such a way that it always either completely succeeded or completely
> failed.

That is a reason, but I was holding that in reserve to kill MOVE in case the
more fundamental arguments did not work.

> Of course that obstacle could be alleviated either by limiting a
> MOVE command to moving a single message or by relaxing the all-or-nothing
> requirement and providing feedback on each message moved.

I seriously doubt that those who advocate MOVE would be pleased by the former
solution.  It's the type of solution I would propose as a way of killing MOVE;
by providing a MOVE that is of such limited use that it won't be used.

The latter solution is an abomination.  We've been there, done that, and got
the T-shirt with feedback on each message.  The result was the all-or-nothing
requirement.

MOVE is even worse than COPY in this respect in that it is not undoable.
There is no way to revert things to the original state with a MOVE that
partially succeeded (think about it carefully).

Reply via email to