Crowdsourcing Scientific Software Documentation: A Case Study of the NumPy Documentation Project

Without good documentation, even the most sophisticated and efficient scientific software is difficult to use and maintain. However, due to lack of time, resources, and incentives, scientists who develop software for other scientists aren´t keen on writing documentation. Scientific software packages that are freely available and shared within scientific communities often suffer from poor documentation, sometimes making them unusable. This article considers documentation crowdsourcing as a way to address the issue. Using the NumPy documentation project as a case study, the article discusses how to leverage the knowledge about software that resides within the user community. The research looks at technical infrastructure, community engagement, and motivation, and considers benefits, such as expanding the community, and challenges, including maintaining commitment level. The conclusion suggests preliminary guidelines for those thinking about documentation crowdsourcing.