Back to article
LinuxCon: What Is the Future of Linux Development?
Too Hard for Noobs?
August 11, 2010
Here at the LinuxCon conference, a panel of Linux kernel developers from Red Hat, Google, Novell and Oracle discussed what's next for the ecosystem, and why not every kernel debuts with a big new feature.
Google developer Ted Ts'o said that the coming of very large multiple CPU cores on chips will be a challenge that Linux developers will have to address. He added that scalability, a concern that Linux kernel developers had largely considered a solved problem, will require work for the next generation of chip scalability. Ts'o also noted that the increasing growth of faster networking speeds will bring additional pressure to the kernel.
"I don't think that we have any lack of interesting technical challenges moving forward," Ts'o said. "There is the concern that we don't have any loyal opposition to challenge us, the way that we did when Sun was funding Solaris to the same degree it used too."
While the kernel itself iterates rapidly with new releases approximately every three months, not every kernel has big new features in it that users will notice right away.
"When we ship Fedora updates the story is more of the same, more drivers and fixes to existing drivers," said Dave Jones, Red Hat's Fedora kernel maintainer. "There are some great features in every release, but it's mostly more of the same from our perspective."
Overall the increasing level of complexity and quality is also making it more difficult for new people to contribute to the kernel. Jones believes that the barrier to entry for kernel developers has been raised in recent years.
"A driver that would have been merged in the early days would now get shot down, with request for re-writes," he said.
But Oracle kernel developer Chris Mason pointed out that there are still many ways for new people to get involved in kernel development. Mason is also the creator of the Btrfs next-generation filesystem.
Novell kernel developer James Bottomley suggested that for those who want to get involved in Linux kernel development they should find an area they are passionate about and then contribute.
Jones noted that there are certain areas where new kernel developers can make more of an impact than others.
"The kernel is big enough and there are whole areas are that are prime areas for people to come and clean up," Jones said.
That said, Jones suggested that new developers not choose to start with what he referred to as janitor patches that help to eliminate needless whitespace.
"Find something that is interesting and useful but at the same time easy enough to understand so you can get involved with it," Jones said.