Free/libre open source software: What we know and what we do not know
Citation: Kevin Crowston, Kangning Wei, James Howison, Andrea Wiggins (2012) Free/libre open source software: What we know and what we do not know. ACM Computing Surveys (RSS)
DOI (original publisher): 10.1145/2089125.2089127
Semantic Scholar (metadata): 10.1145/2089125.2089127
Sci-Hub (fulltext): 10.1145/2089125.2089127
Internet Archive Scholar (search for fulltext): Free/libre open source software: What we know and what we do not know
Download: http://floss.syr.edu/system/files/CrowstonFLOSSReviewPaperPreprint.pdf
Tagged:
Summary
Review of published empirical literature where FLOSS development and use are the main themes: 138 papers from 30 journals and 41 conferences, collected primarily in 2006.
Papers analyzed FLOSS at individual, group, organization, and societal levels, with group analysis comprising 57% of collection.
Case study was the most common form of analysis for all levels, 42% of total. Survey second most common at 24% of total. Case study papers were weak in identifying data collection methods.
52% of studies based on secondary data. Only 35% of sampled papers included references to theory.
Coding of the sample indicated use of the inputs-mediators-outputs-inputs model, drawing on decades of small group research, unsurprising as much FLOSS development occurs in small groups.
Inputs
- Member characteristics
- Project characteristics
- Technology use
Processes
- Development practices
- Social processes
Outputs
- Team performance
- Software implementation
- Software/project evolution
Studies measuring success mostly used some form of code quality.
All areas need more research.
Methodological challenges include use of incomplete archival data, self-reported data, sampling strategies focused on established projects, whether levels of data, analysis and theory match up, and paucity of longitudinal studies.
