A framework for the integration of partial evaluation and abstract interpretation of logic programs

  • Authors:

  • Michael Leuschel

  • Affiliations:

  • University of Southampton, Southampton, U.K.

  • Venue:
  • ACM Transactions on Programming Languages and Systems (TOPLAS)
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Recently the relationship between abstract interpretation and program specialization has received a lot of scrutiny, and the need has been identified to extend program specialization techniques so as to make use of more refined abstract domains and operators. This article clarifies this relationship in the context of logic programming, by expressing program specialization in terms of abstract interpretation. Based on this, a novel specialization framework, along with generic correctness results for computed answers and finite failure under SLD-resolution, is developed.This framework can be used to extend existing logic program specialization methods, such as partial deduction and conjunctive partial deduction, to make use of more refined abstract domains. It is also shown how this opens up the way for new optimizations. Finally, as shown in the paper, the framework also enables one to prove correctness of new or existing specialization techniques in a simpler manner.The framework has already been applied in the literature to develop and prove correct specialization algorithms using regular types, which in turn have been applied to the verification of infinite state process algebras.