The incremental development of correct specifications for distributed systems

Abstract

Provably correct software can only be achieved by basing the development process on formal methods. For most industrial applications such a development never terminates because requirements change and new functionality has to be added to the system. Therefore a formal method that supports an incremental development of complex systems is required. The project CoCoN (Provably Correct Communication Networks) that is carried out jointly between Philips Research Laboratories Aachen and the University of Oldenburg takes results from the ESPRIT Basic Research Action ProCoS to show the applicability of a more formal approach to the development of correct telecommunications software. These ProCoS-methods have been adapted to support the development of extensible specifications for distributed systems. Throughout this paper our approach is exemplified by a case study how call handling software for telecommunication switching systems should be developed.