Hello Frank,
> Thus, in case of I/O-based programs, we could not be sure that we always
> avoid runtime errors due to non-determinism. In my mind, this would be
> a major drawback. The fact that we might not see a good example at the
I agree that this is a problem (and in fact Rafa and I already have been
trapped by that during our work on the debugger). But we have to face the
fact that we cannot have both sharing and encapsulation of all
non-determinism.
> moment is not a big point. The situation might occur very fast when
> one starts considering a special/new class of problems in Curry. From
> my point of view, a solution where non-determinism is encapsulated
> in *every* situation was preferrable. However, I don't see an easy solution
Yes, and so would be a solution where trying to bind a non-local variable
would not cause unexpected deadlocks in your program. However, the problem is
the same in both cases; there is no sound way to do this.
> at the moment.
Wolfgang
--
Wolfgang Lux Phone: +49-251-83-38263
Institut fuer Wirtschaftinformatik FAX: +49-251-83-38259
Universitaet Muenster Email: wlux_at_uni-muenster.de
Received on Di Jan 15 2002 - 06:46:19 CET