The Optimal Approach to Recursive Programs

Abstract

The classical fixedpoint approach toward recursive programs suggests choosing the “least defined fixedpoint” as the most appropriate solution to a recursive program. A new approach is described which introduces an “optimal fixedpoint,” which, in contrast to the least defined fixedpoint, embodies the maximal amount of valuable information embedded in the program. The practical implications of this approach are discussed and techniques for proving properties of optimal fixedpoints are given. The presentation is informal, with emphasis on examples. © 1977, ACM. All rights reserved.