November 14, 2003
©2003 Rich Stillman, Waystation Partners
Do you like this article? You can read more. Better yet, subscribe to my mailing list and you'll receive reviews of new devices and technology that can make your electronic life better today!
The rising level of business interest in Linux, along with the press coverage that has attended the acquisition of SuSE by Novell, has reminded me of some concerns I have been meaning to raise about the long-term direction of this very important piece of software. As much as the acquisition is a positive step for the future of Linux, I believe there are risks in the operating system's journey from an Open Source proving ground to a source of revenue for commercial enterprises.
I'm a strong believer in Linux, largely because of the inevitable effect of competition in the OS and primary applications space. There is a tremendous amount of energy in the Open Source community. The speed and agility of Open Source teams has generated products that have approached or reached parity with Microsoft in product features. As Open Source products close in on Microsoft's feature set - easier to do now that Microsoft has generously offered them a relatively stationary target until 2006 or so - innovation will turn to new products and product categories. The rapid pace of innovation that comes with Open Source development will be difficult for Microsoft or any other company to match. The desktop world of 2005 may actually force buyers to choose between Linux and Windows not on the basis of feature-by-feature comparisons, but by which platform offers a larger subset of the tools they need to get their jobs done.
And this doesn't even bring up the area of servers, where Open Source software competes with Microsoft on an entirely different, and far more mature, level.
Still, as Linux distributions and other Linux-related products become increasingly commercialized, I can't help pointing out that there are potential clouds on the horizon. One is the impact of competition on the Open Source model and has been fairly widely discussed. The other factor, one that has not been considered in the press at least, concerns the risks of succession - what happens when Linus Torvalds is no longer the spiritual leader of the Open Source movement.
It's not hard to think back to the old days of proprietary Unix - not hard at all, since they haven't yet ended - and remember the deliberate incompatibilities layered on top of (and in) the kernel in the name of "competitive advantage". When Digital, Data General, Prime and other minicomputers roamed the earth, Unix offered a real alternative to VMS and other proprietary operating systems. It never reached its potential, not because it wasn't good enough, but because each variant competed in its own bubble, winning by its own strengths and losing by its own weaknesses. Software and skills were largely not portable from one variant to another. Repeated attempts by industry groups to set standards in important parts of Unix failed because they ran up against the competitive interests of individual Unix vendors. As a result, the pace of evolution in the Unix world was glacial, and Unix was never a threat to displace proprietary operating systems even on the minicomputers that were its natural market at the time.
The success of Linux shows the impact that a unified Unix could have had on the market. Linux, of course, had advantages that allowed it to succeed where proprietary Unix failed. Beyond the clear advantage of running on cheap, generic hardware, three factors allowed Linux to evolve to the point of practicality:
These factors have a strong parallel in the development of the Internet, which started life as a research network funded by the government and developed by academics with little interest in the commercial potential of their work. The committees and standards boards that made up the Internet Engineering Task Force (IETF) and its sister organizations were the Internet's strong central figure, guiding the development of the Internet and reminding people that, in a undertaking as dependent as the Internet on the contributions of large numbers of independent developers, standards are A Good Thing. Unchallenged standards bodies have allowed the Internet as a technical design to remain more or less whole in spite of the retirement of almost all of the Internet pioneers and the massive commercialization that has all but replaced the academic orientation of the original developers.
The first two factors, critical as they are to the current structure of Open Source development, are as obviously undermined by the commercialization of Linux as they were by the commercialization of the Internet. However, as the progress of the Internet has shown, they are not necessarily critical to success as a product like the Internet, or Linux, matures.
The third factor, however, is critical. There is no question that the central figure in the world of Linux - Open Source's analogue to the IETF - is Linus Torvalds. Linus resolves disputes that threaten to fork the development of Linux components, sets the overall course of development, and makes the ultimate decisions about the overall project's direction in response to unforeseen external and internal developments like DRM and the issue of direct competition with the Windows desktop. The recent profile of Linus in Wired magazine shows how important Linus is to Linux - how, in spite of his naming of Andrew Morton as a potential successor, Linus is still clearly the guy in charge:
"Earlier this year, Torvalds asked Morton to take over informally as number two. Morton, who for several years ran software development teams inside Nortel Networks, is now overseeing the release of Linux version 2.6, expected by the end of this year. But that arrangement is represented more clearly on an organizational chart than in reality. Some people, it seems, still send potential 2.6 fixes directly to Torvalds - and he'll respond rather than defer to his lieutenant."
As Linux becomes a prime-time player through ever-better releases and corporate buyouts, the central role of Linus Torvalds becomes more critical - and more dangerous. In a traditional business, a founder's departure from the head of the company he or she created can cause upheaval, even if a succession plan is in place and in spite of the obvious fiscal consequences of a poorly planned transfer of power. The Linux community clearly has no succession plan, and no universally accepted authority beyond Torvalds. Further, the fundamental structure of the open source community assures that there is no business oversight of organization or products.
Torvalds' departure from the leadership of the Open Source community, an inevitable development at some point in the near or distant future, is a potentially cataclysmic event. The Open Source community has not, as far as I know, planned for it. The social dynamics of this community, combined with the passion and ego of the next-tier leadership and the financial pressures brought on by commercial success and mainstream acceptance of Linux, make such planning imperative.
So I'll continue to watch and hope for the commercial success of Linux and related Open Source products. But be concerned until a plan has evolved to ensure the long-term future of those products, a plan that does not depend on the continued involvement of one person. The people at Novell and other companies who are investing in Linux and charting its course as a commercial product should share that concern.