Lies, Damn Lies and Statistics
This is sort of “better late than never”, but between traveling to JavaOne and our Foundation staff member meeting last week, I missed Kirill Grouchnikov’s blog posting on who fixed what bugs in Eclipse 3.1.
With regard to the Eclipse projects, IBM still has over half of the committers working on them. (I’ve covered this before here.) So it is neither a secret nor a surprise that they are still carrying the load on many of the projects that were started before the creation of the Foundation. But the interesting data is the trend line on new projects such as BIRT or ECF, where 0% of the committers work for IBM. Kirill’s list of projects was more than a little self serving. It reminds me of the old line about “lies, damn lies and statistics”. For every project within Eclipse that someone can point to which has a lot of IBMers on it, I can point to another which has few or none. What does this prove?
What I don’t understand is why anyone thinks it matters. Eclipse started its life as an IBM project and is evolving into an independent open source community with a diverse group of committers. These are non-controversial facts as far as I’m concerned.
When I talk about Eclipse independence, I am usually thinking of the governance model, where IBM truly set Eclipse free to evolve in the future as an independent entity. The fact that fierce IBM competitors such as BEA, Borland and Computer Associates feel comfortable with the Eclipse governance model to join the board speaks for itself IMO.
Yes, it will take time to evolve the older projects to have more diverse committer populations. But there is no way the Eclipse community wants IBM to slow down its investment in Eclipse. The more the merrier, we say. The goal is to grow the committer community and over time lower IBM’s percentage of a much bigger pie. And new projects like GMF (led by Borland), DSDP (led by Wind River) and DTP (led by Sybase) are moving us quickly in that direction.