OSDL Success

Matt Asay today on his blog made the statement (links added) that “Open Source Development Labs (OSDL) and other industry organizations have miserably failed”.  I take exception to that. OSDL, today the Linux Foundation (LF), has had an important and positive impact on Linux. I think people forget the original premise of the organization when it was founded, what it has done to advance Linux, and how it has evolved to serve the changing needs of the Linux vendors and the Linux community over its lifetime.

OSDL was originally founded to make advanced hardware (hence the “Lab” in OSDL) available to the broader Linux kernel development community. The original idea being that access to a variety of hardware, in particular large-scale hardware with (at the time) advanced features such as 4-way and greater SMP systems, was necessary to advance Linux and not generally available to most kernel developers. This enabled, for example, the first TPC workload testing tools for Linux available to the community. The array of systems available also allowed OSDL to develop and make available to the Linux kernel development community regression testing to help converge stable kernel releases.  I could go on, but many of the successes of OSDL during its “Lab” phase focused on the infrastructure of Linux, and are not as likely to be visible at the open source application level (like Alfresco), but were still important in the broader success of Linux, particularly in the Enterprise data center. This doesn’t even consider services such as providing a safe and neutral environment place for people like Linus Torvalds.

Over time it became apparent that the “Lab” function of OSDL was becoming less necessary to the success of Linux.  I believe, in part, due to the fact that more and more Linux development was moving to paid employees at larger companies who could provide access to the types of systems OSDL was hosting. OSDL, preserving the original vision of advancing Linux, began to look at other barriers to the adoption of Linux.  IP is one example where OSDL was able to respond to industry needs, creating the Linux Legal Defense Fund in response to the SCO lawsuits. In addition, OSDL addressed patent issues with the Open Source as Prior Art (OSAPA) and Patents Commons Project initiatives.

OSDL Workgroups were also formed to bring together people interested in enhancing Linux to make it suitable for new and broader markets.  These included everything from development efforts to support telcos and desktops.

Late in 2006 it became apparent that OSDL needed to undergo yet another transition to continue its broader mission of advancing Linux.  In early 2007 OSDL merged with the Free Standards Group (FSG) to form the Linux Foundation, finally bringing the important role of standards together with the other efforts to foster the growth of Linux.  Today the Linux Foundation incorporates three broad classes of activities:

1. Protecting Linux - LF continues the efforts around IP, patents and legal protection launched by OSDL, including as I already mentioned the Open Source as Prior Art project, the Patent Commons Project, and sponsorship of the Linux Legal Defense Fund.  In addition, the Linux Foundation also manages the Linux trademark.

2. Standardizing Linux - LF produces the Linux Standard Base (LSB) and the Linux Developer Network.  All major Linux distributions comply with the LSB.

3. Promoting Linux - LF serves as a neutral spokesperson to advance the interests of Linux and respond with authority to competitors’ attacks. It also fosters innovation by hosting collaboration events among the Linux technical community, application developers, industry and end users.

So while one might argue that OSDL/FSG/LF might have done more or been more effective, saying it “failed miserably” is I think totally unwarranted. I’m a firm believer in Jim Zemlin and his vision for the organization today, and trust that with the support of board members like Mark Shuttleworth, James Bottomley, Tim Golden and many others he will continue to help Linux grow.

Comments

[Matt Asay]

Hmm...kind of sneaky not to link back to my argument. :-)

I still think OSDL was a complete failure. I was involved through my then-boss, Jeff Hawkins, and watched the futile attempts to be more relevant than Red Hat. They were...futile.

I wish Jim and the LF luck, but I don't personally see much of a role for the organization, and think Jim's many talents would be better served elsewhere (like in running another company). Or, rather, I don't see much of a role that it will be able to successfully fulfill. It tries to keep ahead of two trains: Linux kernel development and Red Hat, both of which largely do what they want to.

The LSB does not guide Red Hat (or SUSE, really, if memory serves). It's a nice to have, but it's not making any purchasing decisions for enterprise IT.

As for protecting against attacks, customers do that best. I'm not sure anyone is looking to the LF to be the shield.

In short, I politely, but firmly, disagree. We don't need more committees in open source (or Linux). We need more winning companies.

[lma]

Not trying to be sneaky, it was just late at night for me. ;-) For those of you that missed it, here's Matt's post where he incorrectly calls OSDL a complete failure:

http://blogs.cnet.com/8301-13505_1-9760440-16.html

I'll add the link into my original post.

Back to the merits of the argument, I think Linux is better off with the LSB, LDF, OSAPA, CGL, DCL, Portland Project and the many other LF funded/supported initiatives than without. LF does not try to keep ahead of Red Hat or the Linux kernel community. Rather, it tries to help both be more successful by solving issues neither are well equipped to handle on their own.

[Don Marti]

Pre-2006 OSDL and today's LF are very different organizations. The group has made some big changes since the kernel developers got involved -- see http://www.kroah.com/log/linux/osdl_board_proposal.html . Before 2006, OSDL wasted a lot of time and effort organizing "workgroups" to put together enormous documents of things that developers are supposed to do, which developers didn't read. Now, it seems that instead of trying to "keep ahead" of the process, LF is providing infrastructure where needed in areas such as the developer travel fund, paying Linus and docs editor Rob Landley of course, and NDAs for drivers. Less ambitious, more useful.

[lma]

Don, I agree with you that today's LF is a more efficient organization better addressing some of the needs of kernel developers. I also agree that some of the OSDL Workgroup efforts in the past were not handled correctly. Some of them tried to treat kernel development like a proprietary software process and failed as a result. But give OSDL credit for the positive things it accomplished, as well as credit for listening to the input from the kernel developers and restructuring itself (in part through the merger with LF) to better serve the needs of kernel developers, Linux vendors, and Linux users. The new LF kept was working at OSDL and changed what wasn't. Overall, I think Linux benefitted from OSDL and will hopefully benefit even more so from the new LF.