🚀 go-pugleaf

In article <1649@gould.doc.ic.ac.uk>, cdsm@sappho (Chris Moss) writes:
>In article <8205@pt.cs.cmu.edu> fp@PROOF.ERGO.CS.CMU.EDU (Frank Pfenning) writes:
>>I am looking for papers/reports/code/information on "legitimate" uses
>>of "assert" and "retract" in logic programming, that is, uses which are
>>considered to be in good taste by the connoisseurs of the field.
>
>Many of the best uses are for assert WITHOUT retract. They generally come into
>the area of "lemmas": asserting things you've found to be true so you don't
>have to prove them all over again. This is mentioned in Sterling and Shapiro
>(The Art of Prolog) p.181 - section entitled "Memo Functions" with an example
>of Towers of Hanoi. ...

Slightly more general than this is allowing some rules to be forward chaining
rules which always add their conclusions to the database.  If you do this and
want to support default rules or a dynamic database, then retractions are
needed, although they should be done in a principled way using a truth
maintenance system.  The result is a very useful component for building certain
kinds of systems in Prolog.  It supports the use of a blackboard type
architecture and also provides a very easy and potentially efficient way to do
certain computations in which the deductive closure of a set of clauses needs to
be computed.  Our own work in this area is described in

     Finin, Tim, Rich Fritzson and Dave Matuzsek, ``Adding Forward Chaining 
     and Truth Maintenance to Prolog'', IEEE Artificial Intelligence
     Applications Conference, (pp. 123 - 130), Miami, March 1989.
-- 
 Tim Finin                                   finin@prc.unisys.com
 Center for Advanced Information Technology  215-648-2840, 215-648-2288 (fax)
 Unisys, PO Box 517, Paoli, PA 19301         215-386-1749 (home)