zs@mulga.oz (Zoltan Somogyi) (04/28/88)
voda@ubc-cs (Paul Voda) writes: > [In Trilogy] the system checks that all calls to 'preds', i.e. to the > true backtracking predicates are enclosed in an all (min, one) solution > construct, that there are no values returned through an "or" and > that there are no symbolic (logic) variables with constraints used > (except enclosed in all,min one constructs). I have a different view: determinacy should be a property of the predicate and its mode[s]. I have a paper on a closely related subject which dwells on this issue at length. It is Zoltan Somogyi: "Stability of logic programs: how to connect don't-know nondeterministic logic programs to the outside world", Technical Report 87/11, Department of Computer Science, University of Melbourne, Australia. If you want a copy, drop me a line. Zoltan.
debray@arizona.edu (Saumya Debray) (04/29/88)
Several people have looked at the connection between predicate determinacy
and modes. Some references that come to mind:
- U. S. Reddy, "Transformation of Logic Programs into Functional Programs",
Proc. SLP-84, Atlantic City, NJ, Feb. 1984.
- C. S. Mellish, "Some Global Optimizations for a Prolog Compiler",
J. Logic Programming vol 2 no. 1, Apr 1985.
- S. K. Debray and D. S. Warren, "Detection and Optimization of Functional
Computations in Prolog", Proc. 3d ICLP, London, Jul 1986.
Also, a paper by Gert Smolka:
- G. Smolka, "Making Control and Data Flow in Logic Programs Explicit",
Proc. 1984 Symp. on Lisp and Func. Prog., Austin, TX, Aug. 1984.
may be relevant (it's been so long since I read the paper I've forgotten
if he treats this issue explicitly, but it's a good paper that people ought
to read anyway).
--
Saumya Debray CS Department, University of Arizona, Tucson
internet: debray@arizona.edu
uucp: {allegra, cmcl2, ihnp4} !arizona!debray