Measuring the health of the open source economy

One of Michael Meeks' graphs from his fascinating analysis of the health of the OO.o project

Here’s a fascinating thing. A really detailed analysis of a single very important open source project, Open (OO.o). The author, Michael Meeks, a prominent Open Office hacker, has tabulated and analysed the application’s CVS logs (CVS is the system that stores and organises the code contributed to the project by its many developers). What we have here is a quite fine-grained set of direct, numerical indicators of the health of Open – in particular, of the engagement of the contributing developers (Meeks acknowledges in the post various weaknesses in the data). How many developers are involved, how much code is contributed and which organisations are most engaged. Meeks’ conclusion is that engagement is falling and that Open is sick. He thinks the project needs urgent remedial work to get it back on track.

The reason I find this so interesting is because it offers a preview of the way we’ll monitor the health of all sorts of projects in the future. Open source, as a way of working, is spreading to other areas of activity, some quite remote from software development. If it really catches on – in business, education and the media, for instance – we’ll presumably be able to analyse this kind of data for many different kinds of work. A university might select an open source physics course or a manufacturer an open source component design by comparing graphs like Meeks’ for competing projects. And Meeks’ hand-rolled analysis will inevitably mature into a measurement and presentation dashboard for the whole open source economy.

We’re also learning that data, once exposed, is quickly acquired, tabulated, visualised and compared. Open data is much more useful than the closed stuff. Can it be long before I can fly through a 3D virtual world of open source projects floating in space, coloured and shaped to represent their various critical attributes?