Re: Prolog and AION

On May 24, 8:21 am, Peter Van Weert <Peter.VanWe...@xxxxxxxxxxxxxx>
wrote:
rupertlssm...@xxxxxxxxxxxxxx schreef:

Peter,

Why do you say "backward chaining", is not really backward chaining?
Is it because in these systems you have to explicitly backward chain
rather than automatically determine the sub-goals as in Prolog?

Yes, it's been a while since I thought this through, but it was
something like that.

Or is it because when multiple rules are fired in a forward chaining
system, if it processes both rules, it does not undo the additions to
the "working memory" from one rule to the next, or just takes the
first rule, or uses some other heuristic (such as rule priority) to
decide which rule to use? Whereas, in prolog it will take one path,
then back up, undoing choice points, and take a different path. Yes,
I'm a bit unclear as to whether these rule systems can undo things,
which is definitely true backtracking.

Rupert

.