Adoption of Open Source Software in Software-Intensive Industry

Context: Open source software (OSS) has seen significant adoption, and it has changed not only how software-intensive companies develop and make money on software, but also how they select and acquire software. As a consequence of these changes there has been confusion about how organizations may benefit from OSS. Objective: Through answering the following research questions (RQ1-RQ3), this thesis aims to explore and describe (1) the different ways in which organizations adopt OSS and (2) how these organizations select OSS components. RQ1: How and to what extent are software-intensive organizations currently adopting OSS? RQ2: What is the current status of research on OSS adoption in organizations and how may this research benefit practitioners? RQ3: Which strategies and resources do software developers use to identify, evaluate, and select OSS components? Method: This thesis consists of six related studies, all focused on the adoption of OSS in software-intensive organizations. These studies embody case studies, systematic literature reviews, and surveys using face-to-face interviews, e-mail, and web-based questionnaires. Results: Based on these six studies, this thesis provides the following five contributions (C1-C5) through eight papers (P1-P8): C1 Empirically grounded descriptions of how several organizations adopt OSS. C2 A systematic review of the literature on OSS adoption in organizations. C3 A classification framework presenting six ways of organizational OSS adoption, each with its particular benefits and challenges. The six ways include: deploying OSS products, using OSS CASE tools, integrating OSS components, participating in the development of OSS products, providing OSS products, and using OSS development practices. C4 Descriptions, based on empirical evidence, of the strategies and resources practitioners actually use to identify, evaluate, and select OSS components. C5 A model for situated software selection and its constraints, indicating why formalized selection methods have failed to see significant adoption. Conclusion: Practitioners should observe that they have several possibilities and take this into account when adopting OSS. Research on OSS adoption should focus on a few topics, borrow more support from related fields within software engineering and information systems research, and extend the foundation offered by this thesis. Research on software selection should put stronger emphasis on the situation the selection is conducted in and the rich (text) experience developers benefit from when selecting (OSS) components.