The allocation of collaborative efforts in open-source software

TitleThe allocation of collaborative efforts in open-source software
Publication TypeJournal Article
Year of Publication2008
Authorsden Besten, M, Dalle, J-M, Galia, F
Secondary TitleInformation Economics and Policy
Volume20
Number4
Pagination316 - 322
ISSN Number0167-6245
Keywordsage, apache, complexity, cvs, division of labor, functions, gaim, gcc, ghostscript, lines of code, loc, log files, mozilla, netbsd, openssh, postgresql, python, revision control, scm, size, source code, Stigmergy, version control
Abstract

The article investigates the allocation of collaborative efforts among core developers (maintainers) of open-source software by analyzing on-line development traces (logs) for a set of 10 large projects. Specifically, we investigate whether the division of labor within open-source projects is influenced by characteristics of software code. We suggest that the collaboration among maintainers tends to be influenced by different measures of code complexity. We interpret these findings by providing preliminary evidence that the organization of open-source software development would self-adapt to characteristics of the code base, in a 'stigmergic' manner.

Notes

"we have selected a set of 10 large open-source projects"
apache, cvs, gaim, gcc, ghostscript, mozilla, netbsd, openssh, postgresql, python
"Our data were extracted from logs of development activity generated by software version control systems. For each project in the selection, we extracted CVS development logs"
"We notably computed for each file in the sample, and for each month in its history, the number of distinct maintainers that had committed a change during that month, and the number of commits, the blocks of code addition, each file had received during that month."
"other variables used in the regressions are proxies for the size, age, and granularity of files; the size of a file is represented as its number of lines of code (LOCs), its age by its creation date (Youth), and its granularity by the number of functions it contains."

URLhttp://www.sciencedirect.com/science/article/B6V8J-4SSG4PN-1/2/88b3824c30a31c18929d8a5ca6d64f62
DOI10.1016/j.infoecopol.2008.06.003
Full Text