Download Patterns and Releases in Open Source Software Projects: A Perfect Symbiosis?

TitleDownload Patterns and Releases in Open Source Software Projects: A Perfect Symbiosis?
Publication TypeBook Chapter
Year of Publication2010
AuthorsRossi, Bruno, Russo Barbara, and Succi Giancarlo
Secondary AuthorsÅgerfalk, Pär, Boldyreff Cornelia, González-Barahona Jesús, Madey Gregory, and Noll John
Secondary TitleOpen Source Software: New Horizons
Series TitleIFIP Advances in Information and Communication Technology
PublisherSpringer Boston
Keywordsflossmole, oss2010, sourceforge

Software usage by end-users is one of the factors used to evaluate the success of software projects. In the context of open source software, there is no single and non-controversial measure of usage, though. Still, one of the most used and readily available measure is data about projects downloads. Nevertheless, download counts and averages do not convey as much information as the patterns in the original downloads time series. In this research, we propose a method to increase the expressiveness of mere download rates by considering download patterns against software releases. We apply experimentally our method to the most downloaded projects of SourceForge's history crawled through the FLOSSMole repository. Findings show that projects with similar usage can have indeed different levels of sensitivity to releases, revealing different behaviors of users. Future research will develop further the pattern recognition approach to automatically categorize open source projects according to their download patterns.