Semi-automatic improvement of software development methods: Doctoral consortium paper

Empirical studies show that the use and adoption of software development methods (SDMs) in practice is still low. Not just that developers often avoid to follow what is prescribed in their organisations, very often the organisations don´t even have prescribed and documented methods. This situation is recognized as one of the main reasons for failures in IT projects and a contributor to the low quality of software. In this paper we describe a PhD thesis that addresses this problem. We introduce an innovative approach and supporting tool that will semiautomatically monitor in-action methods during project performance, and capture the gained knowledge with only negligible involvement of the developers. Our expectation is that since the prescribed methods will be in this way based on the practice that is accepted in a certain organisation, its users will not perceive it as something imposed by their managers and will follow it much more rigorously. Furthermore, we intend to show that it is possible to establish a continuous software process improvement lifecycle with just a marginal user involvement.