Greg K-H gave a talk at the Linux Plumber's Conference ostensibly about the Linux Ecosystem, but appears to have been primarily about Ubuntu's leading commercial sponsor, Canonical. I wasn't in attendence; as a mere amateur, I'm satisfied to read reports and watch videos from conferences like LPC. It's a lot cheaper on the wallet, certainly.
But I do wonder about his accounting. It's simple enough to measure the rate of change of the kernel, the speed of releases and so on, but it's much harder to referee the score. I haven't seen any published scripts to automatically attribute his tally, and I imagine publishing such a thing might be an invasion of privacy. Still there are questions in my mind. If someone chooses to contract a kernel hacker to write a network driver, who does Greg attribute? Does his account of X.org include any work done by Daniel Stone?
Greg is right though, that it takes a lot of work just to tread water with the kernel. The rate of change will bowl you over, which I'm told is one reason Redhat hires a small army of kernel developers to backport patches to RHEL, and I'm sure their customers love them for it.
The Good News
There is, I think, an unprecedented opportunity coming up to collaborate. Fedora, openSuSE and Ubuntu will all be shipping a .27 kernel in a stable release. It occurs to me that this would be a perfect time for one of those extended stable kernel cycles that Greg mentioned in that Google Tech Talk some time ago. I hope this idea is brought up during a conference about solving problems and the "kernel ecosystem".