Re: indexing
From: Stephan Lehmke (Stephan.Lehmke_at_ls1.cs.uni-dortmund.de)
Date: 08/10/04
- Next message: rhapsody: "Help I'm stuck!"
- Previous message: Thomas Gagne: "Re: Prolog Implementation / Trace on Backtracking"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 10 Aug 2004 11:50:08 GMT
In article <1089485089.394555@seven.kulnet.kuleuven.ac.be>,
bart demoen <bmd@cs.kuleuven.ac.be> writes:
> This shows one of the biggest scandals in Prolog implementation: all the
> theory and technology for doing multiple argument indexing has been
> around since about two decades. But AFAIK, there are very few who do
> it - I know of only two Prolog systems: ilProlog (now hipP) and
> BIM-Prolog (now MasterProlog).
> Similarly, deep indexing (looking deeper inside the terms than just
> the principal functor) is not done by most Prolog systems; XSB can do
> it through unification factoring.
AFAIK SWI has a declaration like
:- index(p(0,1))
which would make p second-argument-indexed.
ECLiPSe does a lot of indexing things I only vaguely remember,
including looking whether the first literal of a clause is some
sort of comparison which could be used to verify cases are
exclusive. At least I remember giving Prolog course once
using ECLiPSe, saying a lot of clever things about first
argument indexing and then having an extremely hard time to
make up at least one case where a redundant choice point was
created.
regards
Stephan
-- Stephan Lehmke Stephan.Lehmke@cs.uni-dortmund.de Fachbereich Informatik, LS I Tel. +49 231 755 6434 Universitaet Dortmund FAX 6555 D-44221 Dortmund, Germany http://ls1-www.cs.uni-dortmund.de
- Next message: rhapsody: "Help I'm stuck!"
- Previous message: Thomas Gagne: "Re: Prolog Implementation / Trace on Backtracking"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|