Open Source One

Google Tech Talks February, 25 2008 ABSTRACT We all love to hate large software systems. They are hard to build, hard to evolve, and don’t work very well. Why is this? A lot of reasons, some social, some technical, and some socio-technical. We believe that OSS provides an excellent source of data to test hypotheses about the factors that affect important phenomena/outcomes in software projects. Our group at UC Davis, comprising bio-informaticians, organizational behaviourists, physicists, and software engineers, is using a range of different approaches to the analyze the veritable torrents of data pouring out of open source projects to understand how things work in OSS, and what tools and techniques can help. One important issue is IMMIGRATION: how do new people join projects, and how can we help the difficult intellectual and social challenges they face. We present two results: 1.a) Can we build “recommender” tools that help programmers deal with “information overload” by helping them focus their attention? 1.b) Such tools have always been evaluated with user studies. Can we do something more quantitative? 2) What are the factors influencing immigration of new developers in Open source projects? Joint work with: V. Filkov, A. Swaminathan, G. Hsu, and students C. Bird, Z. Saul, and A. Gourley We gratefully acknowledge support from NSF (Science of Design and Human and Social Dynamics Programs), the IBM Faculty Fellowship Program, and the grammatech and scitools …