I am (unexpectedly) at the MIT/HBS Free/Open Source Software conference at Harvard. The first session included a paper comparing Apache, Mozilla and a commercial software project. The results suggested open source development does indeed deliver higher productivity levels and lower defect counts than closed team development. While I felt intuitively that this was the case, I'm pleased to see a growing research base supporting the fact.
All three papers also suggested that, regardless of the overall size of the community, most of the work was conducted by a core of participants which, as one delegate pointed out, tended to be of similar size to hunter/gatherer groups across human history. Much of the presentation activity is trying to model and understand effects which seem intuitively correct but lack logically coherent discussions.
But the one topic I'm not sure is really being explored is the phenomenon of commercial involvement in open source activities. All the discussions assume that all the participants are volunteers, but in communities like NetBeans.org and OpenOffice.org there are a large number of participants who are employed as professional engineers to participate in the community. I'll be watching today to see if this dimension gets some airtime.