MATISSEProject reference: IST-1999-11435
Funded under: FP5-IST
Methodologies and technologies for industrial strength systems engineering
Total cost:EUR 3 171 498
EU contribution:EUR 1 844 437
Coordinated in:United Kingdom
Funding scheme:CSC - Cost-sharing contracts
The core objective is the development of industrial strength methodologies and associated technologies for the engineering of software-based critical systems. These methodologies and technologies will support industry in providing essential services for the information society of the Third Millennium that are highly dependable and therefore lead to increased public confidence and trust in the services.
The project will make significant use of rigorous, mathematically-based software engineering techniques, so-called formal methods, that support validation throughout the development life-cycle by providing rigorous specification and design notations as well as proof techniques, model-checking techniques and simulation techniques.
The driver for the research and development in MATISSE will be three major industrial case studies representing a spectrum of the essential services for the information society.
MATISSE will provide:
1) guidelines that allow formal methods to be integrated into existing systems engineering lifecycles so that higher degrees of safety and reliability can be achieved;
2) methodologies that expand the range of application of these formal methods; and
3) enhancements to technologies associated with these formal methods.
The approach of the MATISSE project will be to exploit and enhance existing generic methodologies and associated technologies that support the correct construction of software-based systems. In particular, a strong emphasis will be placed on the use of the B Method, invented by Jean-Raymond Abrial, and its associated technology, as well as the CSP formal method, invented by Tony Hoare. Between them, the industrial and academic partners, have considerable expertise in the development and application of these methods.
The work programme will be based on three major industrial case studies representing a spectrum of the essential services for the information society.
An embedded verifier for a smartcard system.
A railway signalling and control system.
A diagnostic system for healthcare clinicians and researchers.
These case studies have been provided by three of the industrial partners and represent real products and services that the industrial partners currently have under development. The industrial partners developing the smart card system and the railway system respectively already have successful experience of applying the B Method to system development over several years. The objective for these partners is to increase the benefit of formal methods further by introducing them earlier in the life cycle and by applying them to entire systems rather than just the software parts of systems. The industrial partner developing the healthcare system have very little experience of the use of formal methods but are keen to introduce them to their organisation and understand the benefits. The MATISSE project will allow them to develop their products and services to a higher standard of safety and dependability and will provide benefits beyond the lifetime of the project by providing an improved engineering methodology
The objectives of the project have been broadly met. The Case Studies outputs are of great interest because of their diversity and levels of participant maturity in using formal methods. Their contributions to the MATISSE Methodology make it more likely that the Methodology will be applicable to other industrial critical systems design projects. Hence MATISSE has achieved its first objective. The use of heterogeneous methods has not been fully developed in MATISSE, partly due to the sound advice and direction given by the reviewers. There is some evidence of the necessity to complement the B method largely used in MATISSE Case Studies with other appropriate formal and semi-formal methods when designing a complex system. This should not be seen as contradicting the Methodological approach developed by the project but instead as an extension of this Methodology. Within this context the MATISSE project has adequately met the second objective. Four tools have been developed to support the MATISSE Methodology: an event B to B translator, by ClearSy; an Xemacs Interactive Prover Interface, by ClearSy; a translator from UML to B, by Southampton University; and a set of tools to generate distributed programs following the CORBA paradigm from B, developed by CNRS.
These tools are either supported commercially or are available for download from the Web. These tools demonstrate that MATISSE has achieved its third stated objective. The MATISSE project has made significant use of rigorous, mathematically-based software engineering techniques to support validation throughout the development life-cycle by providing rigorous specification and design notations as well as proof techniques, model-checking techniques as required by the specific case studies. The overall aim achieved by the project was the development of industrial strength methodologies and associated technologies for the engineering of software-based critical systems. These methodologies and technologies will support industry in providing essential services for the information society of the Third Millennium that are highly dependable and therefore lead to increased public confidence and trust in the services.
SW1E 6PD LONDON
75794 PARIS CEDEX 16
13855 AIX EN PROVENCE LES MILLES
SO17 1BJ SOUTHAMPTON