James Howison (chair), Diane Bailey, Unmil Karadkar, Wenhong Chen (RTF)
Do open projects "break the mirror" between organizational and product structures?: Re-conceptualization of the Mirroring Hypothesis in Free/Libre Open Source Software (FLOSS) development
Location: Location: UTA 5.522 (1616 Guadalupe St., 5th Floor)
The “mirroring hypothesis” holds that when products and the organizations that create them have structures that mirror each other coordination costs will be reduced and projects will be more successful. This idea is used in both the design of products and organizations. Yet research suggests that FLOSS projects can be successful even when violating this principle, “breaking the mirror” when presumably loosely-coupled FLOSS contributors develop tightly-coupled systems. This dissertation investigates this idea through a multiple case studies using an explanatory sequential mixed methods design. I studied how work was organized in three FLOSS projects by combining analysis of trace data and interviews with participants and using a configurational approach that employs fuzzy set qualitative comparative analysis (fs/QCA). First, I examined product structure over time, through an automated study of the structure of source code, identifying periods of change for further study. Second, I conducted a fine-grained analysis of these periods, breaking down the project level into increasingly concrete levels: successive inter-release periods and then individual episodes of work. I find that there is indeed alignment between two structures at those concrete levels in three successful FLOSS projects: GNU grep, IPython, and Scikit-image. Further, this dissertation identifies the coordination recipes for tightly-coupled software work: commonly occurring patterns of organizational configuration dimensions leading to increases in software coupling. This dissertation makes contributions to the study of open collaborations in organization science and empirical software engineering by reconceptualizating, reconfirming, and refining the continued relevance of the mirroring hypothesis. Practically, my results suggest that organizations seeking to benefit from aligning product and organizational structure must grapple with the organization of periods and episodes of work rather than primarily rely on the abstract level of software and organizational architectures in order to take advantage of the coordination benefits of mirroring relationships between products and the organizations that produce them.
12:30pm to 3:00pm