Evidences in the evolution of OS projects through Changelog Analyses

Most empirical studies about Open Source (OS) projects
or products are vertical and usually deal with the flagship,
successful projects. There is a substantial lack of
horizontal studies to shed light on the whole population of
projects, including failures. This paper presents a horizontal
study aimed at characterizing OS projects.
We analyze a sample of around 400 projects from a popular
OS project repository. Each project is characterized
by a number of attributes. We analyze these attributes
statically and over time.
The main results show that few projects are capable of
attracting a meaningful community of developers. The
majority of projects is made by few (in many cases one)
person with a very slow pace of evolution.
We then try to observe how many projects count on a substantial
number of developers, and analyze those projects
more deeply. The goal is to achieve a better insight in the
dynamics of open source development.
The initial results of this analysis, especially growth in
code size and tendency to stability in modularity, seem to
be in line with traditional close source development.

Citation:
Capiluppi, A., Lago, P., Morisio, M. (2003). ‘Evidences in the evolution of OS projects through Changelog Analyses.’ in Feller, P., Fitzgerald, B., Hissam, B. Lakhani, K. (eds.) Taking Stock of the Bazaar: Proceedings of the 3rd Workshop on Open Source Software Engineering ICSE’03 International Conference on Software Engineering Portland, Oregon May 3-11, 2003. pp.19-24..