In: Proceedings of Applications and Theory of Petri Nets 2004: 25th International Conference, ICATPN 2004, Bologna, Italy, June 21-25, 2004, pages 391-410. Volume 3099 of Lecture Notes in Computer Science / Cortadella, Reisig (Eds.) --- Springer-Verlag, September 2004.
Abstract: One of the key issues of object-oriented modeling is inheritance. It allows for the definition of a subclass that inherits features from some superclass. When considering the dynamic behavior of objects, as captured by their life cycles, there is no general agreement on the meaning of inheritance. Basten and Van der Aalst introduced the notion of life-cycle inheritance for this purpose. Unfortunately, the search tree needed for deciding life-cycle inheritance is in general prohibitively large. This paper presents a comparative study between two possible algorithms. The first algorithm uses structural properties of both the base life cycle and the potential sub life cycle to prune the search tree, while the second is a plain exhaustive search algorithm. Test cases show that the computation times of the second algorithm can indeed be prohibitively expensive (weeks), while the computation times of the first algorithm are all within acceptable limits (seconds). An unexpected result of this case study is that it shows that we need tools for checking life-cycle inheritance.