Pattern Driven Task Model Refinement

Task models have been used as a tool to elicit requirements in the early development stages. Moreover, they have recently proved to be a suitable starting point for modeling of interactive processes. During the different development stages several corresponding task models are built. Although every model is just a refined version from the previous one, this refinement process is not a trivial operation as a lot of rules and restrictions have to be respected in order to successfully infer the suitable task model concerning the current level of abstraction. In this paper we aim to assist the developer by giving him the opportunity to move with a given model from one abstraction level to another one in an easier and more seamless way. Thus, we present an approach consisting of some guideline patterns which help the developer to transform a given task model between the different development stages in a more performing and less error-prone manner.