Common Pool Resource in Open-Source Software Projects: The Contributor Energy Framework
Curtis J. Atkisson, Ștefan Stănciulescu, Brenda K. Bushouse, Vladimir Filkov · 2024
Open-source software (OSS) has been intensively studied by researchers to understand predictors of long-term success and sustainability, though less attention has been paid to what leads to consistent contributions to OSS projects. Those who contribute to projects are considered to be part of a Common Pool Resource, and their contributions to OSS projects amalgamate into a public good. Despite an increase in open-source software contributors worldwide, contributors often come and go, and are limited in the time they can provide to projects. In this work, we present the Contributor Energy framework inspired by a theory from Volunteer Management that conceptualizes the Contributor Energy as a limited resource that projects must manage and use wisely or risk fully depleting, leading to several predictions: contributions are cyclical, projects follow periods of intense work with fewer contributions, and projects regularly under stress struggle to get consistent contributions. We test these using a longitudinal dataset of 210 projects from the Apache Software Foundation Incubator. We find that a project using their Contributor Energy intensively one month predicts doing more work in the next month, though there is substantial variation among projects. Projects of different sizes are remarkably similar, though large projects that intensively use Contributor Energy one month are more likely to have no contributions the next month. Those projects that are more likely to have decreased contributions after a month of intensive use of Contributor Energy are those that stress themselves more often. While the ancillary analyses provide some support for Contributor Energy as a limited resource, the main evidence goes in the opposite direction. The Contributor Energy framework we developed opens a new research avenue, allowing us to understand and make predictions on how energy is used and maintained at the project level, which we can test using longitudinal data from OSS projects.