Software on the edge

Abstract

As our society becomes more technologically complex, computers (and the software that they run) are being used in a potentially alarming number of high consequence safety-critical applications. When these systems fail, the outcome can be devastating. Formal methods provide what, by a growing number of experts, is considered to be the best approach to making the software construction process more reliable. What makes formal methods so attractive, from a reliability standpoint, is that one has the ability to conclude, with mathematical certainty, that a software component is correct. In this context, when we say a software component is correct, we mean it satisfies its formal specification. The paper discusses the problems of software reliability and the use of formal methods.