The Linux Foundation–the new organization formed this year from the merger of Open Source Development Labs and the Free Standards Group–is holding a summit this week at the main Google campus.
I think we can already call the summit a success, and an indication of the Linux Foundation’s acceptance by its community, just on the basis of the many busy, well-known people who chose to show up. It was a great place to make connections among both speakers and attendees.
The first day was a pretty public affair, with an audience of 200 that included journalists, so the discussions were basically polite and stuck to acceptable debates such as how to recruit more developers. I could not, unfortunately, attend the sessions later in the week where (to use Executive Director Jim Zemlin’s metaphor) the sausage would be made. But some interesting points came up anyway.
Role of the Linux Foundation
It was impressive to see the thanks tendered by Linux kernel developers to the Linux Foundation, especially for its work to recruit developers from the Pacific Rim. Greg Kroah-Hartman and others spoke about how valuable it was to travel on the Foundation’s dime to Japan and encourage development groups there. Bringing East Asian developers more into the kernel community will be a major focus.
Other tasks Zemlin defined for the Foundation include promoting and defending Linux and free software. It may be hard to remember that free software is still subject to legal threats. But the SCO versus IBM cluster of cases drag on and patent wars escalate. Under such circumstances, one can’t expect FUD to go away either–why are U.S. governments so slow to adopt free software and open standards?
GPL debate, version ad infinitum
A stark contrast emerged between the attitude of Linux developers toward the soon-to-be-released GPL version 3 and the attitude of prominent lawyers in the free software community who worked on the GPL over the past couple years.
The Linux kernel developers kept a solid front, backing what Linus Torvalds has said about version 3 from the beginning. The most positive statements they could muster is that the Free Software Foundation listened to its critics, that the final GPLv3 is much better than its early drafts, and that (in the words of Ted Ts’o) it would be “basically OK for userspace programs” (not the kernel, in other words).
James Bottomley presented the kernel developers’ digging in as a pragmatic decision. Given that individual developers keep the copyright on their contributions, it would be too much trouble to try to get them all to relicense their code. But everyone made it clear they didn’t want to change from v2 to v3 anyway.
Later that day, attorneys Karen Copenhaver and Mark Radcliffe had the chance to discuss their work with the Free Software Foundation on v3. They both praised the new license and suggested that it would spread gradually to projects over time. Everybody at the conference suggested that the community “chill” (relax and let things take their natural course). I talked to O’Reilly colleague Allison Randal, who has worked a lot on v3 from her perspective as a Perl language contributor. She confirmed that from a legal standpoint, the GPLv3 stands on much firmer ground than v2, but pointed out that even the lawyers who supported the license didn’t predict quick, widespread adoption.
Shuttleworth gave a quietly very inspiring talk that straddled the important themes of community building and formal tool support. He pointed out that traditional, centralized version control tools such as CVS and Subversion divide the user community into two parts: those with the rights to commit changes and those who can merely download. For free software, he promoted the use of the newer distributed tools such as git and Mercurial, where anybody can make a clone of the software and people can choose whom to accept changes from. “Everybody’s important,” he said.
Free software developers have created superb tools for sharing bug reports and other information with collaborators around the world on a single project, but we need to do more to allow such sharing between projects. Shuttleworth contrasted our development silos with Microsoft, where people on widely distinct projects can get in touch with each other over IM within minutes.
To reach this stage, we need to create standards for data exchange between different bug trackers and other tools (as well as between data formats), and end the current situation where someone who wants to submit bug reports to different projects has to create an account on each project. Sharing should be accompanies by digital signatures to prevent the projects from becoming vulnerable as they become more open.
He also praised formal planning, saying that when starting the Ubuntu project he was inspired by the GNOME developers’ ability to explain to users what they were developing and to deliver the projects relatively on time; that was also one reason Ubuntu was released with GNOME support first.
Finally, while expressing satisfaction with the 2-3 month release cycle of the Linux kernel, he asked that larger cycles of experimentation and bug-fixing be layered on top of it.
There were certainly warnings about things that need to be fixed. Brian Aker of MySQL AB said that the new Linux threads were buggy and that most MySQL users had reverted to the old threads implementation–or in some cases switched to OpenSolaris. Aker also said support was difficult because the kernels in different distributions had different bugs, even if they were compiled from the same kernel version.
Although the Linux Standard Base makes development and support much easier, as reported by some major vendors such as Real Networks, adoption is still slow.
I’ll end on some pleasant statistics voiced by industry leaders. Dan Frye of IBM said Linux was ahead of its competition in real-time. (A patch now brings real-time into the kernel core.) Christy Wyatt of Motorola said her company would base 60% of its upcoming phones on Linux, and reported that “even the most conservative industry predictions” says Linux will soon tie for first place in the mobile market.
My impression is that Linux and free software are expanding their hold in areas where they have traditionally done well, such as in the back office and on specialized devices created by sophisticated developers who have time to tinker. But there are still big challenges to overcome in reaching the desktop, user tools, and other areas less well understood by the free software community. That doesn’t mean the proprietary vendors have it all figured out and give users everything they want. It just means that free software doesn’t have a strong enough basis for overturning the lead of the proprietary software.