An exceptionally user-friendly approach to computer-aided validation / verification of concurrent and reactive systems is presented. In it, the user needs not express his verification questions formally in detail. Instead, he specifies a point of view to the system by choosing a subset of its externally observable actions. An automaton abstracts and reduces the behaviour of the system according to the choice, and shows the result graphically on a computer screen. The resulting picture represents all executions of the system, as seen from the chosen point of view. Thus the information in it is as comprehensive as that obtained by ordinary verification. On the other hand, like ordinary testing, the method makes it possible for a system designer to get rapid feedback with ease, to "just try the system and see how it behaves". The article concentrates on practical and philosophical issues regarding the method and contains a detailed example.

This article presents a user-fn'endly f o m l method for obtaining comprehensive feedback from the behaviour of a design of a distributed system. The method consists of abstracting and reducing the behaviour and showing the result in a graphical form on a computer screen. Deadlocks, livelocks, incorrect actions and many other kinds of design errors can often be easily discovered from the graphical representation. The method differs from animation, simulation and testing in that it analyses all possible behaviours simultaneously. For instance, i f the graphical representation does not contain a deadlock, then the system is guaranteed to be free from deadlocks. The method is illustrated by using it in the development of a wellknown communication protocol. This article concentrates on the use of the method; the underlying theory and computer tools have been discussed elsewhere.

No abstract

No abstract

No abstract