Logical and Computational Aspects of Programming with Sets/Bags/Lists

We study issues that arise in programming with primitive recursion over non-free datatypes such as lists, bags and sets. Programs written in this style can lack a meaning in the sense that their outputs may be sensitive to the choice of input expression. We are, thus, naturally led to a set-theoretic denotational semantics with partial functions. We set up a logic for reasoning about the definedness of terms and a deterministic and terminating evaluator. The logic is shown to be sound in the model, and its recursion free fragment is shown to be complete for proving definedness of recursion free programs. The logic is then shown to be as strong as the evaluator, and this implies that the evaluator is compatible with the provable equivalence between different set (or bag, or list) expressions. Oftentimes, the same non-free datatype may have different presentations, and it is not clear a priori whether programming and reasoning with the two presentations are equivalent. We formulate these questions, precisely, in the context of alternative presentations of the list, bag, and set datatypes and study some aspects of these questions. In particular, we establish back-and-forth translations between the two presentations, from which it follows that they are equally expressive, and prove results relating proofs of program properties, in the two presentations. Comments University of Pennsylvania Department of Computer and Information Science Technical Report No. MSCIS-91-29. This technical report is available at ScholarlyCommons: http://repository.upenn.edu/cis_reports/382 Logical and Computational Aspects Of Programming With Sets/Bags/Lists MS-CIS-91-29 LOGIC & COMPUTATION 31 Val Breazu-Tannen Ramesh Subrahmanyam Department of Computer and Information Science School of Engineering and Applied Science University of Pennsylvania Philadelphia, PA 19104-6389 Revised October 1991 This paper has appeared in the Proceedings of ICALP '91 Logical and Computational Aspects of Programming with Sets/Bags/Lists Val Breazu-Tannen Ramesh Subrahmanyam Department of Computer and Information Science University of Pennsylvania 200 South 33rd St., Philadelphia, PA 19104, USA E-mail: [email protected], [email protected] Abstract. We study issues that arise in programming with primitive recursion over non-free datatypes such as lists, bags and sets. Programs written in this style can lack a meaning in the sense that their outputs may be sensitive to the choice of input expression. We are, thus, naturally lead to a set-theoretic denotational semantics with partial functions. We set up a logic for reasoning about the definedness of terms and a deterministic and terminating evaluator. The logic is shown to be sound in the model, and its recursion free fragment is shown to be complete for proving definedness of recursion free programs. The logic is then shown to be as strong as the evaluator, and this implies that the evaluator is compatible with the provable equivalence between different set (or bag, or list) expressions . Oftentimes, the same non-free datatype may have different presentations, and it is not clear a priori whether programming and reasoning with the two presentations are equivalent. We formulate these questions, precisely, in the context of alternative presentations of the list, bag, and set datatypes and study some aspects of these questions. In particular, we establish back-and-forth translations between the two presentations, from which it follows that they are equally expressive, and prove results relating proofs of program properties, in the two presentations. We study issues that arise in programming with primitive recursion over non-free datatypes such as lists, bags and sets. Programs written in this style can lack a meaning in the sense that their outputs may be sensitive to the choice of input expression. We are, thus, naturally lead to a set-theoretic denotational semantics with partial functions. We set up a logic for reasoning about the definedness of terms and a deterministic and terminating evaluator. The logic is shown to be sound in the model, and its recursion free fragment is shown to be complete for proving definedness of recursion free programs. The logic is then shown to be as strong as the evaluator, and this implies that the evaluator is compatible with the provable equivalence between different set (or bag, or list) expressions . Oftentimes, the same non-free datatype may have different presentations, and it is not clear a priori whether programming and reasoning with the two presentations are equivalent. We formulate these questions, precisely, in the context of alternative presentations of the list, bag, and set datatypes and study some aspects of these questions. In particular, we establish back-and-forth translations between the two presentations, from which it follows that they are equally expressive, and prove results relating proofs of program properties, in the two presentations.