Characterizing the OSS process

TitleCharacterizing the OSS process
Publication TypeConference Paper
Year of Publication2002
AuthorsCapiluppi, A, Lago, P, Morisio, M
Secondary TitleProceedings of the 2nd ICSE Workshop on Open Source
Keywordsbugs, change log, classification, cvs, downloads, freshmeat, metadata, patches, popularity, project success, release history, sourceforge, vitality
Abstract

The Open Source model of software development has gained the attention of both the business, the practitioners’ and the research communities. The Open Source process has been described by the seminal paper by Eric Raymond [4] and [5]. However, sound empirical studies are still very limited [3], [6].
Our goal is to investigate the OS process by empirical means, to analyze, characterize it, and possibly model it with quantitative models. It should be noted that the Open Source process provides open process and product data, and therefore is a rare opportunity for empirical research.
Our initial research focus is on the characterization of the process, starting from the evolution of OS projects. In traditional projects, a significant number of releases in a short time is usually considered an instability factor [7] and [8], while in the OSS community, it is an evidence of vitality, shows the commitment of the authors and the power of attraction of other programmers [9]. Is it possible to characterize the vitality of projects? And, can vitality be traced to some other characteristics of a project?

Notes

"We have considered two well-known Open Source portals (FreshMeat [1] and SourceForge [2] )."
"Using pseudo-random sampling we have selected a sample of 400 projects (mostly from FreshMeat). Each project is described by several variables (programming language, type of license, size of source code, type of documentation available and others). By indirect means (analysis of the Changelog file, or CVS) it is also possible to compute the number of people working on the project, and the number of external contributors. From FreshMeat we get both a vitality index, that considers the number of releases per time period, and a popularity index, which is a first measure of the interest of users to the project (project URL hits, mixed with subscriptions to it)."

Full Text
AttachmentSize
PDF icon CapiluppiLagoMorisio.pdf171.28 KB