Hypersequential programming is a new method of concurrent-program development in which the original concurrent program is first serialized, then tested and debugged as a set of sequential programs (scenarios), and finally restored into the target concurrent program by par-allelization. Both high productivity and reliability are achieved by hyperse-quential programming because testing and debugging are done for the serialized versions and the correctness of the serialized programs is preserved during the subsequent parallelization. This paper proposes scenario-based hypersequential programming for reactive multitasking systems that have not only concurrency and nondeterminacy, but also interruption and priority. Petri nets with priority are used to model reactive systems featuring interruption and priority-based scheduling. How reactive systems are made highly reliable by this approach is explained and the effectiveness of the approach is demonstrated through the example of a telephone terminal control program.
展开▼
