Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ukma!husc6!bunny!slzr
From: slzr@GTE.COM (Suzanne Sluizer)
Newsgroups: comp.lang.prolog
Subject: Re: retractall and backtracking
Keywords: Quintus, retract, backtracking
Message-ID: <7056@bunny.GTE.COM>
Date: 9 Jun 89 19:51:20 GMT
References: <575@hfserver.hfnet.bt.co.uk>
Reply-To: slzr@bunny.UUCP (Suzanne Sluizer)
Followup-To: comp.lang.prolog
Distribution: comp
Organization: GTE Laboratories, Waltham, MA
Lines: 30

In article <575@hfserver.hfnet.bt.co.uk> nigel@hfserver.hfnet.bt.co.uk (Nigel Cliffe) writes:
>
>I have a problem with backtracking, and retracts that don't appear to
>happen - here is a simplified example:

          [deleted example of retracting something that you
           are backtracking over]
 
>Can anyone explain why this happens.

I would imagine that you will get much more technical explanations than
I could give you, from various net people.  However, I can give you a
very simple idea of why Prolog isn't handling this.

Your problem comes from the fact that by retracting something that you
are backtracking over, Prolog gets "confused".  Basically, Prolog
loses track of "foo(X)" when you retract it, which makes it impossible
to properly resatisfy that goal.

You are lucky that you're using a robust Prolog -- it is possible to
get "bus errors" and "core dumps" if Prolog doesn't handle this
situation gracefully.



-- 
Suzanne Sluizer                CSNET:  slzr%gte.com@RELAY.CS.NET
GTE Laboratories               UUCP:   ...!harvard!bunny!slzr
617-466-2882
"Truth is a pathless land." -- Krishnamurti