Canonical is not interested in the Linux kernel

Tweet
Facebook
Google plus
Send to

LINUX VENDOR Canonical said it has "no interest" in Linux kernel development.

Two weeks ago a Linux Foundation report showed that since version 2.6.32, Microsoft had committed more code to the Linux kernel than Canonical. Since then, Canonical has faced claims from rivals that it does not contribute to Linux as much as it should given its popularity.

Recently Canonical founder Mark Shuttleworth told The INQUIRER that his company has no interest in contributing to the Linux kernel.

Previously Scott Crenshaw, VP of Red Hat's cloud business had told The INQUIRER that Canonical does not write code. When asked, Shuttleworth responded to that by saying, "I think Red Hat feels [that] if they keep saying that enough maybe it'll be true."

Shuttleworth continued, "It's absolutely true we have no interest in the core fundamentals of the Linux kernel, none whatsoever. The Linux kernel was flying before Ubuntu was founded, what was missing at the time was the commitment to the end user experience, the quality of the whole integration that Ubuntu essentially brought. I don't think anybody who thinks about that seriously would say that the enormous amount of work that we do on that is not a contribution."

Shuttleworth said that the kernel community doesn't buy Red Hat's claims of Canonical not submitting Linux code. "When you go to a kernel conference, 70 per cent of the people are running Ubuntu. That means the work they are doing on the kernel is enhanced by the fact that their laptops all just work, security updates all just work. So even in the kernel community there's a rolling of eyes every time Red Hat trots that statistic out," said Shuttleworth.

Then Shuttleworth went on the offensive, claiming that Red Hat uses Canonical code in the core of its products. He said, "Red Hat uses Canonical's core cloud initialisation technology, Red Hat uses Canonical's system initialisation technology. In other words every time you start [Red Hat Enterprise Linux] it either starts or not depending on whether Canonical's technology works. It is mission critical, it is process [ID] one, it is the thing that starts everything else on your computer, it is as critical as the kernel and Red Hat uses Canonical technology for that."

Shuttleworth claimed Canonical has strong views on where it should contribute, saying that should not be held against it. Shuttleworth said, "So on the cloud, in the enterprise version, in the desktop Red Hat uses an enormous amount of Canonical technology because we're investing very substantially in things that we think are important, and the fact that we are very opinionated about the fact that we think the kernel community is vibrant and healthy and mostly led by the hardware people who care about it shouldn't be held against us."

As for Crenshaw, when asked about Shuttleworth's 'process one' claim, he told The INQUIRER, "I'm glad he can find one specific example, I'm sure there a few others, but the fact remains when the world looks to innovators, companies that invest in engineering, Red Hat stands out as the leader and it is very difficult to find Canonical as a contributor of substance."

Red Hat's Crenshaw did say Canonical would be welcomed should it decide to contribute to the Linux kernel and it would be mutually beneficial. "We welcome more contributions from them. We think if they contribute as much code as they have contributed to marketing hype we think the community would be richer and all of our customers would benefit, all the community would benefit, but the numbers bear it out, it's very clear that even [Microsoft] that branded Linux as a cancer is now a bigger contributor to Linux than Canonical is," said Crenshaw.

With Shuttleworth saying that Canonical has no interest in developing the Linux kernel, the firm is clearly positioning itself as an integrator, which is not necessarily a bad thing. The problem Canonical has is that its latest efforts at polishing Ubuntu have mostly been seen through its Unity interface, which has had a troubled introduction. Some may question how much its focus on end-user experience delivers more than rivals such as Red Hat's Fedora distribution.

Shuttleworth clearly has a point about the Linux kernel doing fine on its own, and conversely Canonical could argue that Red Hat does little for consumer Linux by working on things for which mainly enterprises will pay handsomely. From Red Hat's and Canonical's comments it seems that both firms are working on their own models to promote Linux, which is perhaps the best way to gain widespread acceptance. µ