The ASM Method for System Design and Analysis. A Tutorial Introduction

We introduce into and survey the ASM method for highlevel system design and analysis. We explain the three notions|Abstract State Machine [37], ASM ground model (system blueprint) [7] and ASM re nement [8]|that characterize the method, which integrates also current validation and veri cation techniques. We illustrate how the method allows the system engineer to rigorously capture requirements by ASM ground models and to stepwise re ne these to code in a validatable and veri able way. 1 Scope and Achievements of the ASM Method An outstanding feature of the ASM method is that within a single precise yet simple conceptual framework, it naturally supports and uniformly integrates the following activities and techniques, as illustrated by Fig. 1 (taken from [24]): { the major software life cycle activities, linking in a controllable way the two ends of the development of complex software systems: requirements capture by constructing rigorous ground models, i.e. accurate concise high-level system blueprints (system contracts), formulated in domain-speci c terms, using an application-oriented language which can be understood by all stakeholders [7], architectural and component design bridging the gap between speci cation and code by piecemeal, systematically documented detailing of abstract models via stepwise re ned models to code [8], validation of models by their tool-supported simulation, veri cation of model properties by tool-supported proof techniques, documentation for inspection, reuse and maintenance by providing, through the intermediate models and their analysis, explicit descriptions of the software structure and of the major design decisions, { the principalmodeling and analysis techniques, on the basis of a systematic separation of di erent concerns (e.g. design from analysis, orthogonal design decisions, multiple levels of de nitional or proof detail, etc.): integrating dynamic (operational) and static (declarative) descriptions, combining validation (simulation) and veri cation (proof) methods at any desired level of detail.