Path: utzoo!attcan!uunet!clyde.concordia.ca!mcgill-vision!bloom-beacon!snorkelwacker!tut.cis.ohio-state.edu!ucbvax!decwrl!shelby!eos!ptolemy!raymond
From: raymond@ptolemy.arc.nasa.gov (Eric A. Raymond)
Newsgroups: comp.ai
Subject: Re: Are there extensions to Rete to allow arbitrary matching of lists?
Message-ID: <3947@bacon.ptolemy.arc.nasa.gov>
Date: 8 Feb 90 00:38:07 GMT
References: <1990Feb7.225451.16224@hri.uucp>
Organization: NASA Ames Research Center
Lines: 15

mull@hri.uucp writes:
>The normal Rete pattern matching algorithm is meant to match a flat
>(list of atoms) pattern against flat working memory elements.  We would
>like to know if the Rete algorithm can be extended to handle arbitrary
>lists for both wme's and patterns where the pattern could have match
>variables at arbitrary places in its list structure.  We would like
>to do this without resorting to unification.

ART (Inference) does this to a limited degree.  I would expect Mercury
(AI Technologies ???) to do the same (it is like ART on top of CLOS).


-- 
Eric A. Raymond  (raymond@ptolemy.arc.nasa.gov)
G7 C7 G7 G#7 G7 G+13 C7 GM7 Am7 Bm7 Bd7 Am7 C7 Do13 G7 C7 G7 D+13: Elmore James