Linux is the granddaddy of all open source projects, the blueprint for the decentralized development processes. And some of those who use the Linux code, free for the taking, don't give back in equal measure. Still, the time for cajoling those users -- even commercial projects like Ubuntu leader Canonical -- into participating is over, says Jim Zemlin, executive director of the nonprofit Linux Foundation.
FULL Q&A TRANSCRIPT: Zemlin talks about Torvalds' leadership, HTML5 and mobile's future
Zemlin, who spoke with Network World editors at the recent LinuxCon event, used to preach that contributing back was important on moral grounds, as the "right thing to do." But now he says, "It doesn't matter. I don't care if anyone contributes back." Sooner or later, he believes contributing will become an obvious business decision. It's "not the right thing to do because of some moral issue or because we say you should do it. It's because you are an idiot if you don't. You're an idiot because the whole reason you're using open source is to collectively share in development and collectively maintain the software. Let me tell you, maintaining your own version of Linux ain't cheap, and it ain't easy," he says.
He points out that Red Hat is one of the largest contributors to the kernel and also one of the most successful Linux distros. "So if some aren't giving back as much as others today, I just think it will naturally happen over time. It always is in their business interest to do so," Zemlin says.
Canonical is the popular example in the Linux distro community of relatively paltry contributions to Linux (and also to GNOME, although it recently downgraded GNOME in favor of its homegrown desktop, Unity). In the latest tabulation of the biggest kernel contributors, Microsoft landed in the Top 10 while Canonical wasn't even in the Top 30, according to LWN.net. This has been going on for years. In 2008 Linux kernel developer Greg Kroah-Hartman rather famously dressed down Canonical in his keynote speech at the Linux Plumbers Convention.
Canonical argues that its contribution is the popularization of Linux, which Zemlin does not dispute.
"Just to be clear, Canonical staff, engineers, management are not idiots. They get open source well and as they grow, I think it will be in their business interests to give back," he says.
While the profitability of privately held Canonical isn't known (the general understanding is that it's not profitable), there's little doubt that Ubuntu is popular and growing more so. It claimed more than 12 million users by the end of 2010 and is the fourth most popular Web server distro, according to W3Techs.
For enterprises wondering how much contribution is fair to the open source projects they use most, if Linux can be seen as an example, the answer is: as much or as little as you want.
On the other hand, Linux isn't hurting for contributors. In 1992, 100 developers were working on the kernel. By the end of 2010, 1,000 were.
Zemlin says that Linux is a success because the people building it don't plan.
"We have been a part of something that wasn't exactly a master plan. Linux has jumped from one form of computing to another somewhat seamlessly because self forming communities pop up," he says. Linux has moved from high performance computing and servers to embedded to mobile because someone has grabbed the code, started a project and attracted the committers it needed. Neither the Linux Foundation nor the kernel developers controlled that.
Indeed, Linus Torvalds told attendees at LinuxCon in August that he can't predict what's next and doesn't think of himself as a visionary kind of guy. He is focused on the changes to the next release and perhaps the one after that. As the kernel team issues a new release about every three months, that's a mere six months of insight at a time.
It is this lack of planning that could make Linux the first operating system not to be killed off by the Next Big Thing, Zemlin believes. Unix hit its rockslide by missing the desktop, he says. Microsoft hit the desktop but, so far, is struggling with mobile. "Linux, because it has these self-forming communities, hits it. Nobody is predicting and trying to shift course, it just naturally happens."
Trouble is, it can be hard for enterprises to create three- to five-year master plans and secure budgets for them when the technology leaders of their chosen platform are neither visionaries nor planners. The battle between open source virtualization platforms Xen and KVM is one example. Xen has more users, but KVM was the first to be accepted into the mainline kernel. Until Xen also made it into the kernel this summer, many in the Linux community questioned its survival and its users were left wondering if they should migrate or not.
NOD OF APPROVAL: Xen Hypervisor Goes Mainstream
Maybe the plans of Linux's leaders, even Torvalds, are immaterial because the people themselves are optional.
When asked if Linux without Linus would be better off than Apple without Steve Jobs, Zemlin's response is, "Unquestionably. Regardless of the hierarchy in Linux, there's a clear, collaborative spirit and structure. There's much more of a democratic nature even though there has to be someone at the end of the day that makes a call." Does "command-and-control" work better than "decentralized development"? "History will be the judge and I don't know," Zemlin says.
It will likely be many years before history can make that call. Although Jobs just resigned from his day-to-day CEO job at Apple, Torvalds has publicly promised that his retirement is many years away.
Even so, if Torvalds were to stop today, on whose head would the ball drop? Zemlin surprisingly says it's not Kroah-Hartman but Andrew Morton. "Andrew has one of the toughest jobs in Linux. He's the guy responsible for the development tree. Linus trusts Andrew Morton," he says.
Despite Zemlin's disclaimers that he can't see the future, that hasn't stopped him from trying to predict it. During his keynote speech at the recent LinuxCon event he poked fun at himself and his annual proclamation that this would be the year of Linux on the desktop. The desktop is one area where Linux can't claim success.
Zemlin says desktop dominance grows less important and we all have Apple to thank for it. "Apple did something for Linux that was really good. They have clearly shown that the desktop doesn't matter. Microsoft's stalwart position doesn't matter, and that has really enabled a lot. They showed that there are alternative ways of accessing information and they have definitely been breaking ground on this move toward streaming videos and music. Android is a result of that and it's all Linux."
Nevertheless, he quips he would be "completely remiss if I didn't predict the year of the desktop every single year."
Julie Bort is the editor of Network World's Open Source community. She also writes the Source Seeker blog, the Odds and Ends blog for Cisco Subnet and the Microsoft Update blog for Microsoft Subnet. Follow Bort on Twitter @Julie188.
Read more about software in Network World's Software section.