Configuration management support for software process models

Abstract

The urgent need for configuration management was recognized in applying process management in an industrial context. We considered it an marginal problem as long as we were working with a process management prototype which focused very much on the basic process management support, i.e. modeling, analysis, enaction. Based on this experience, we strongly believe that support for further marginal activities, e.g. application data import and export, evolution of application data, migration of legacy system data to process management solutions, easily can turn out to be key success factors (or, if not available, major roadblocks). Configuration management is one area where tool support is urgently needed. Only if appropriate support is provided, the promised benefits of process management will be exploitable in industrial software development.

The discussed problems of configuration management for process models have been experienced in the scope of a business process management project [5]. In this project a set of interconnected process models has been delivered to five customers. Each of these customers modified parts of these process models, three of them nontheless wanted to benefit from later process model releases. In this project we learned that manual solutions to the configuration management problem are error-prone and labour-intensive. For the time being, a set of tools supporting rudimentary configuration management for process models have been implemented. Due to the enormous problems with static process model modifications, we did not tackle configuration management problems of dynamic modifications.