Unearthing open source decision-making processes: A case study of python enhancement proposals

Abstract

Good governance practices are pivotal to the success of Open Source Software (OSS) projects. However, the decision-making processes that are made available to stakeholders are at times incomplete and may remain buried and hidden in large amounts of software repository data. This work bridges this gap by unearthing enacted decision-making processes available for Python Enhancement Proposals (PEPs) from 1.54 million email messages that embody decisions made during the evolution of the Python language. This work employs a design science approach in operationalizing a framework called DeMaP miner that is used to discover hidden processes using information retrieval and information extraction techniques. It also uses process mining techniques to visualize the processes, and comparative structural analysis techniques to compare different decision processes. The work identifies a richer set of decision-making activities than those reported on the Python website and in prior research work (48 new decision activities, 199 new pathways and 6 new stages). The extracted decision process has been positively evaluated by a prominent member of the Python steering council. The extracted process can be used for process compliance checking and process improvement in OSS communities. Additionally, the DeMaP Miner framework can be extended and customized to suit other OSS projects, such as the OpenJDK project.