Proposal for a Flexible Software Process Model : Extended Abstract

Current process simulation models do not allow us to easily and cost-effectively compare different kinds of software development process (for example, Waterfall and XP), or easily choose between viable process alternatives. There are two main reasons for this. The first is that the models tend to be process-specific, meaning that they cannot be easily applied to different processes. The second is that most models are produced for specific organisations, and use historical data from the target organisation in the formulation of the model. This limits the scope of the model, since it is only applicable within the target environment, and constrains simulation efforts to relatively mature groups that can supply the required historical data. There is interest in the selection and customisation of processes for individual projects. This interest comes from three directions: from the topical area of ‘traditional’ versus ‘light’ discussions [4]; from economics and the idea that value creation requires the use of a process that is no more costly than necessary [5]; and from a need for a stronger software development foundation [2, 10]. Customisation of processes is not new [3, 7] and a number of articles relating to this theme are available [9, 13, 16, 17]. However, these works involve making small changes to a conventional base process. This paper gives a proposal for a model that can be applied to all software development processes. This model can then be used as the basis for process simulation to provide comparisons between the processes and to explore the effects of variations on a process.