And frankly, you have no idea how relieved I am to type that subject heading. I have, this hour, finished the last of my graphs. Here's my general conclusion: There is no definite conclusion. There are, however, two surviving possibilities: * Bazaars break Brooks' Law. Never before seen anywhere that the Law of Diminishing Returns is taken to apply. * Bazaars obey Brooks' Law, but their marginal costs rise very, very slowly. I suspect this is the case. Directions proposed: * The study's method be repeated, on a larger project. The GNOME project's set of data - although large - was not nearly large enough to derive solid conclusions. * Some sort of software to automate the gathering of project metrics would make future research infinitely easier, as well as assisting projects in auditing their own 'bottlenecks'. The graphs were highly oscillatory. The Marginal Output, averaged across the 211 GNOME subprojects, was highly volatile. The Average Total Output averaged across the subprojects was predictable, with a sharp "kink" halfway along the curve, which manifests as a huge downward spike in the MO graph. You'll all get to see them tommorrow. What I am going to send is essentially the final draft. It will have the main body as I hope to submit it, including the discussion and conclusions. It will also include the completed appendices, most importantly Appendix III (Caveats) and Appendix IV (Economics in Free Software). It will be missing a few diagrams that will be scanned in for the dead-trees version. These are unessential for the logic in the essay. I'd really like to thank everyone who has helped me so far. In particular, I'd like to thank Michael Zucchi, who helped me extract the raw data I needed from the GNOME CVS, and I'd especially like to thank Richard M. Stallman. Whilst Richard has not provided the bulk of economic commentary, he has politely and promptly answered all of my pesterings, and pointed out errors I have made on various Free Software items. And, of course, I'd like to apologise to license-discuss. In retrospect, it was an inappropriate forum to post to; and, further, it led to a mild flamewar that continues to smoulder. It has, however, produced some excellent postings, some of which have helped *me* to better understand the community license-discuss represents. Thank you all again; JC.
