Hi Wolfgang,
After your reply I understand more what you are dealing with. I
may have written some incorrect statements in my previous message.
A condition to consider is "functionality". For each input there
is only one (or at most one) output. Another condition is
non-determinism. For each input there can be different outputs,
but at most one per execution. How should these conditions be
considered to give a semantics to findall?
If we consider pure rewriting (or narrowing), all the normal forms
of coin are {0,1}. However, if we consider all the normal forms
that can be obtained with one binding, or during one execution,
then the situation is different. In some sense, 0 and 1 are
alternatives values and considering them together may be wrong.
If we consider them together, we lose the "functionality"
condition.
In Curry the situation may be more complicated because of local
declarations or call-time choice semnatics or something else.
I am afraid that the above paragraph will not help you much,
but solving the subtle points may help with the bigger ones.
Sergio
_______________________________________________
curry mailing list
curry_at_lists.RWTH-Aachen.DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/curry
Received on Do Jan 10 2002 - 09:02:05 CET