2008 Poster Sessions : An Empirical Study of Developers and Bugs in Linux

Student Name : Philip Guo
Advisor : Dawson Engler
Research Areas: Computer Systems
Many common wisdoms about open-source development exist but have not been rigorously quantified by research studies. For instance, open-source proponents often point to extreme democratization ('anyone can contribute code') and source code transparency ('more eyeballs able to scan the code means more bugs found') as reasons why open-source software can be robust and secure. But can everyone contribute code equally? What is the value of expert developers? Do experts and professional developers write higher-quality code than marginal contributors? Is it true that having more people work on a piece of code means that it will be less buggy, or will it be even more buggy? We attempt to answer questions such as these by performing data mining and statistical analyses on one of the largest and most well-known open source software projects: the Linux kernel.

Philip J. Guo is a 2nd-year Ph.D. student in the Department of Computer Science. His research interests are in performing empirical studies of software systems and programmer productivity. He is currently working with Prof. Dawson Engler on quantifying development patterns and finding factors that correlate with bugs in large open-source software projects.