- Info
Bugs, libre software and Wikipedia
Taking advantage of the stay of Prem Devanbu (UC Davis) in Madrid, we're organizing a short open seminar, hopefully with interesting discussions about libre software, Wikipedia, bug reports, and other related and unrelated topics.
When
Thursday, June 25th, 10:00
Where
Seminar 103 (first floor)
Edificio Departamental II
Universidad Rey Juan Carlos, campus de Móstoles
Program (tentative)
10:00 Opening. The GSyC/LibreSoft group, by Jesus M. Gonzalez-Barahona (GSyC/LibreSoft)
10:20 Fair and Balanced? Bias in bug fix reporting in Open Source Software Projects, by Prem Devanbu (UC Davis)
10:40 Wikipedia: From quantity to quality, by Felipe Ortega (GSyC/LibreSoft)
11:00 Closing and final discussion
About the talks
Short presentation about the GSyC/LibreSoft group, and its research activities
- Fair and Balanced? Bias in bug fix reporting in Open Source Software Projects
Bugs are tracked in a bug database, from the time they are reported (opened) to when they are diagnosed, fixed (by a suitable source code change) and closed. Bug fixes and other (e.g., enhancement) code changes are committed to a source code repository (e.g., SVN). Developers are asked to report if the change is to fix a bug. These reports relate specific changes to specific bugs. This <em> linkage </em>, between code change and a specific bug, is critical to bug prediction efforts, which try to fit statistical or machine learning models to predict defect occurrence. These models use source code properties, (e.g.,) complexity, size, rate of change, authorship, dependency structure, etc, as predictors. However, often, there are substantially fewer linked bugs than there are bugs actually reported to have been fixed in the bug-tracking systems. Thus many bugs are being fixed silently, without linkage. This brings up an important question: are linked bug fixes an unbiased sample of the population of bug fixes? Results thus far indicate a systematic bias in defect fix reporting in Eclipse, and other other projects; we study the nature, effects, and potential remedies for this bias.
- Wikipedia: From quantity to quality, by Felipe Ortega (GSyC/LibreSoft).
Wikipedia remains to be the epitome of massive collaborative projects focused on content production. Nevertheless, the traditional concern of the Wikipedia community about the growth of the project (number of articles, editors, contributions, etc.) has been progressively substituted by an increasing concern on improving the quality of its contents. We will examine which parameters have determined the evolution of Wikipedia from its inception until present time, and the key factors that must be consider to effectively face off this forthcoming challenge on quality improvement.
About the speakers
- Jesus M. Gonzalez-Barahona is associate professor at GSyC (Universidad Rey Juan Carlos). He developes his research interests in the area of libre software at the GSyC/LibreSoft group.
- Prem Devanbu is professor of Computer Science at UC Davis. He got his bachelors in Electrical Engineering from IIT Madras, India, and his Ph.D in computer science from Rutgers Unviersity, NJ USA. He is interested in many things, but right now he thinks open source software engineering is pretty cool.
- Felipe Ortega is researcher and project manager at GSyC/LibreSoft. His main research line is the study of open online communities, in particular the Wikipedia project. His Ph.D. thesis "Wikipedia: A quantitative analysis" is the first research work to present a rigurous side-by-side comparison of the top-10 language editions of Wikipedia, following an empirical approach.