Generalized semantics and abstract interpretation for constraint logic programs

By: R. Giacobazzi, S. K. Debray and G. Levi

Roberto Giacobazzi
Dip. di Informatica
Univ. di Pisa
Corso Italia 40, 56125 Pisa (Italy)


We present a simple and powerful generalized algebraic semantics for constraint logic programs that is parameterized with respect to the underlying constraint system. The idea is to abstract away from standard semantic objects by focusing on the general properties of any--possibly non-standard--semantic definition. In constraint logic programming, this corresponds to a suitable definition of the constraint system supporting the semantic definition. An algebraic structure is introduced to formalize the notion of a constraint system, thus making classical mathematical results applicable. Both top-down and bottom-up semantics are considered. Non-standard semantics for constraint logic programs can then be formally specified using the same techniques used to define standard semantics. Different non-standard semantics for constraint logic languages can be specified in this framework. In particular abstract interpretation of constraint logic programs can be viewed as an instance of the constraint logic programming framework itself.

Available: DVI, PostScript, BibTeX Entry.