Context-aware process management for the software engineering domain

Abstract

Historically, software development projects are challenged with problems concerning budgets, deadlines and the quality of the produced software. Such problems have various causes like the high number of unplanned activities and the operational dynamics present in this domain. Most activities are knowledge-intensive and require collaboration of various actors. Additionally, the produced software is intangible and therefore difficult to measure. Thus, software producers are often insufficiently aware of the state of their source code, while suitable software quality measures are often applied too late in the project lifecycle, if at all. Software development processes are used by the majority of software companies to ensure the quality and reproducibility of their development endeavors. Typically, these processes are abstractly defined utilizing process models. However, they still need to be interpreted by individuals and be manually executed, resulting in governance and compliance issues. The environment is sufficiently dynamic that unforeseen situations can occur due to various events, leading to potential aberrations and process governance issues. Furthermore, as process models are implemented manually without automation support, they impose additional work for the executing humans. Their advantages often remain hidden as aligning the planned process with reality is cumbersome. In response to these problems, this thesis contributes the Context-aware Process Management (CPM) framework. The latter enables holistic and automated support for software engineering projects and their processes. In particular, it provides concepts for extending process management technology to support software engineering process models in their entirety. Furthermore, CPM contributes an approach to integrate the enactment of the process models better with the real-world process by introducing a set of contextual extensions. Various events occurring in the course of the projects can be utilized to improve process support and activities outside the realm of the process models can be covered. That way, the continuously growing divide between the plan and reality that often occurs in software engineering projects can be avoided. Finally, the CPM framework comprises facilities to better connect the software engineering process with other important aspects and areas of software engineering projects. This includes automated process-oriented support for software quality management or software engineering knowledge management. The CPM framework has been validated by a prototypical implementation, various sophisticated scenarios, and its practical application at two software companies.