A Longitudinal Study on Collaboration Networks and Decision to Participate in a FLOSS Community
|Title||A Longitudinal Study on Collaboration Networks and Decision to Participate in a FLOSS Community|
|Publication Type||Conference Paper|
|Year of Publication||2010|
|Authors||Conaldi, Guido, and Tonellato Marco|
|Secondary Title||5th Workshop on Public Data about Software Development (WoPDaSD 2010)|
|Keywords||bicho, bug fixing, bug reports, bugzilla, COLLABORATION, developers, epiphany, flossmetrics, gnome, social network analysis|
In this paper we conjecture that individual decisions of FLOSS (Free/Libre Open Source Software) developers to take on a task are influenced by network relations generated by collaboration among project members. In order to explore our conjecture we collected data on a FLOSS project team consisting of 227 developers committed since 2002 to the development of a web browser. We reconstructed 2-mode co- collaboration networks (software developer by bug) in which a tie represents an action taken by a developer in order to solve a specific bug. Co-collaboration networks were collected at five points in time during a six-month development cycle of the software. We report and discuss results of longitudinal actor-based modeling that we specify to test for the influence of local network structures on developer’s decision to take action on a specific bug. The study controls for bug-specific and developer-specific characteristics that may also affect developers’ decisions exogenously. We also control for priority and severity levels assigned by the team to bugs in an attempt to manage voluntary contribution.
"we chose as a case of study Epiphany, which is the default web browser of the GNOME graphical desktop environment"
"We collected all relevant data by parsing all the bug reports in GNOME Bugzilla repository relative to Epiphany. The data collection and storing was done using Bicho (v. 0.4 rev. 7198), a software part of the FLOSSMetric project ."