Static Analysis

interpretation serves as a powerful theoretical tool for developing and justifying program-analysis algorithms. It provides a way to establish that information extracted from a program by a program-analysis algorithm is a meaningful characterization of what can occur when the program is executed. Typically, however, it is not an easy task to obtain the appropriate abstract statetransformation functions and to show that they are correct. On the contrary, papers on program analysis often contain exhaustive (and exhausting) proofs to demonstrate that a given abstract semantics provides answers that are safe with respect to a given concrete semantics. In contrast, the parametric framework for program analysis proposed by Sagiv, Reps, and Wilhelm [POPL ’99, TOPLAS (in press)] holds out the promise of allowing different program-analysis tools to be generated automatically from high-level descriptions of what an analysis designer desires. Analyses generated in this fashion would be correct by construction. This framework has two parts: – A language (3-valued logic) for specifying various properties that an entity may or may not possess, and how these properties are affected by the execution of the different kinds of statements in the programming language. – A method for generating a static-analysis algorithm from such a description. The talk will review the principles behind the paradigm of “abstract interpretation via 3-valued logic,” discuss recent work to extend the approach, and summarize on-going research aimed at overcoming remaining limitations on the ability to create program-analysis algorithms fully automatically. (Joint work with M. Sagiv, R. Wilhelm, F. DiMaio, N. Dor, T. Lev-Ami, A. Loginov, A. Mulhern, N. Rinetzky, and E. Yahav.) M. Hermenegildo and G. Puebla (Eds.): SAS 2002, LNCS 2477, p. 1, 2002. c © Springer-Verlag Berlin Heidelberg 2002 Open Modeling in Multi-stakeholder Distributed Systems: Research and Tool Challenges