DEV Community

Joseph Castle
Joseph Castle

Posted on • Updated on


Open Sourcing the People's Code

This is a summary of my PhD dissertation completed Spring 2020 at Virginia Tech and supports my entry into $ git remote graduation.


Computer software is essential for daily operations of industry and government. Business and government are being run on software for entertainment, for solving public health crises, and for supporting national security.

The federal government spends $6 billion on software each year but very little of the software is open sourced. This is costly, allows little to no reuse, results in duplication, and often yields poor quality.

In 2016, the White House published the Federal Source Code Policy directing major federal agencies to create an internal source code policy, update acquisition language to capture new custom code, and inventory source code while publishing a portion as open source software (OSS).

Even with the policy, agencies in 2019 were publishing OSS with mixed results.


Given each agency has an IT budget relative to its size and employees with IT skills, yet little publishing of OSS, it was believed there were systemic organizational factors with code publication. Therefore, the premise of the study centered on broader organizational factors related to organizational cultural beliefs, public engagement, structural dimensions, and organization location either hindering or aiding agency OSS publication.


The qualitative study included a metadata analysis, interviews, and artifact collection (e.g., SOPs, organization charts). The sample contained the following attributes.

  • 25 participants were from SW development units in 20 of 24 CFO Act agencies.
  • Some agencies had multiple participants, but no participant was from the same SW development unit.
  • Most units conducted SW development and data science activities.
  • Units often consumed more OSS than they published.
  • More units were located outside the agency Chief Information Officer (CIO) office.


The study findings pertain to the four organizational factors of the study - cultural beliefs, public engagement, structural dimensions, and organizational location.

  • Cultural beliefs - Beliefs are a key component of organizational culture and support organizational performance. Beliefs allow organizations to make predictions, select actions, and experience new phenomenon. Cultural beliefs in this study were both advantageous and cautionary.

    • Advantageous beliefs - These beliefs were based on perceived or realized outcomes. Offices that published OSS minimally to frequently held these beliefs.
    • Cautionary beliefs - Participants whose units held cautionary beliefs were reticent in their response to new technology and technology policy. Approximately half of the sample held cautionary beliefs and minimally published OSS.
  • Public engagement - Public engagement is part of a larger public participation construct. Public engagement pertained to actions taken with individuals inside and outside the software development shop. Offices publishing more OSS tended to work with those outside their office through various activities including in-person events with multiple electronic tools.

  • Structural dimensions - Related to technology-structure research and pertains to decision-making, rules, skills, and communication of an office. All units, whether publishing minimally to frequently, required formalized policy allowing them to do so.

  • Structural location - Similar to structural design research, focuses on hierarchy or where the office is located in the broader organization (i.e., on an organizational chart). Offices with more autonomy, less hierarchical layers and close to an approving executive were able to more frequently publish OSS.


The "tl,dr" nature of this post prevents adding all the details of the study. I encourage you to read my dissertation as there is much more detail of the study and implications for public administrators, policy makers, and organizations in general.

Top comments (0)