Are formal methods useful for software development?

Abstract

The relevance of formal methods for practical software system design is discussed. Prominent representatives of formal approaches present their findings and experience about the use and the usefulness of formal methods. It has been proposed that all programmers would be more productive and produce higher quality products if they would learn two things: predicate calculus; and program correctness (including formal program development). It is argued that the complexity, pervasiveness, and critical nature of modern and future computer systems makes it imperative that such systems be engineered for reliability and maintainability. Formal methods constitute an extremely promising approach to the design of reliable systems. The schedulability aspect of real-time system development is discussed. In general, formal methods should be preferred over other less formal methods since they can provide much better and stronger guarantees on real-time system performance.<>