Community-cation: A Look Back at Week 38, 2008

The big news at the Linux Foundation this week was not one but two--count 'em!--two events in the lovely Portland, Oregon. Let OSCON pack up and prove it knows the way to San Jose--we like the Rose City, just fine, thank you.

The first event was our own Linux Kernel Summit, where 80 kernel developers, including Linus Torvalds, got together to hash out ongoing issues with the Linux kernel, and plan out the future direction of the kernel's development. It was an invitation-only event (apparently my invite was lost in the company e-mail), so there wasn't a lot of press coverage of what happened. If you have a subscription to Linux Weekly News, though, Jonathan Corbet provides an excellent round of articles covering the two-day event. If you don't have a subscription, the article will be available on September 25.

Right after that was the Linux Plumbers Conference, a community-organized event that the Foundation underwrote. I had the opportunity to meet the show organizers this summer while I was attending OSCON, and they really followed through on putting together a primo event geared for system developers. There wasn't much press coverage of this event either, but thanks to the blogging efforts of keynote speaker and kernel developer Greg Kroah-Hartman, there was some news generated from the event anyway.

Greg K-H, who has criticized the kernel contribution efforts of Ubuntu maker Canonical in the past, used his keynote speech to toss some hard numbers out there on who among the commercial Linux vendors contributes the most to the Linux kernel. According to Greg's findings, among commercial vendors (with Debian and Gentoo thrown in as well), Canonical ranks 81st in number of kernel contributions between 2.6.15 and 2.6.27-rc6:

With only 100 patches in these kernels, it sure does appear that Greg makes a good case for calling Canonical out as a non-contributor to Linux. Canonical's CTO Matt Zimmerman certainly begs to differ, though, as one might expect. Matt indicates that Greg's data is wrong and even if it's close, Canonical has never claimed to be a big kernel contributor anyway.

But here's the thing that neither one of them seem to have pointed out. Where is it written that you have to be a prolific kernel contributor to be a valued member of the community? If you go by that logic, then Greg's own company, Novell, is a much less valuable member of the Linux ecosystem than Red Hat. Plus, this line of reasoning completely negates the other contributions a company can make to Linux as a whole. Desktop environments, applications, packaging systems, funding, documentation, driver development... there are a ton of ways someone or some company can contribute to the Linux ecosystem, which is not comprised of just the kernel.

Looking at Canonical's other contributions to the ecosystem, I would be hard pressed to call them a bunch of slackers. Could they contribute more to the kernel and other systems software? Sure, that would be great. But given the strides they are making on the desktop and getting Linux out on OEM-delivered machines, I don't think it's necessary to call them out and try to shame them into it.