Process mining for process conformance checking in an OSS project: An empirical research

Abstract

With almost 20 years of research, Process Mining can now be considered to be in a mature phase allowing its application to a variety of sectors. In this article, the bug closure process that is followed by a community of an open source software project is investigated in order to perform process conformance checking. Actual data that reveal the process steps have been extracted from the project’s Bugzilla database and have been used as input in Disco process mining tool. The data includes extracted information for more than 19,000 bugs for the past 15 years in a csv form, formatted appropriately to construct an event log suitable for process mining. The extracted models have been compared to the process described in the project’s blogs and wikis by the community. The same models are also compared to the bug closure process that Bugzilla suggests to be used by the projects using this software for bug tracking purposes. The findings reveal that indeed the process followed in the OSS project is very similar to the declared one but variations do occur under specific circumstances. However, the process is not identical to the one proposed by Bugzilla suggesting that each OSS project can customize its processes in order to better address the needs of the project and the community. This empirical research highlights the importance of process mining in OSS projects in order to investigate the processes followed and identify outliers helping to standardize and improve the processes and enhance the collaboration among the members of the communities.