How many Open Source developers does it take…
2000-09-01
There is certainly no shortage of companies offering up their products to the open source movement at the moment. Last week, IBM announced its plans to release the source code to its AFS enterprise file system to the “developer community.” A few weeks ago, as we discussed \link{http://www.it-director.com/00-07-03-3.html,here} Sun Microsystems admitted to difficulties in opening up Solaris. This relentless march of companies is delivering a growing stack of source code to the seemingly infinite resource pool that makes up the open source movement. The cathedral builders are handing over their plans to the bazaars, and the world will be a better place for it. That’s the theory anyway – in practice, as demonstrated by the recent announcement of a company-sponsored open source laboratory, it is the major corporations that rule the roost.
As far as we know there are no statistics concerning the breakdown of the open source movement into its constituent developer types. It seems reasonable that there are three categories:
- commercial developers, who are salaried workers tasked with the modifications for business reasons
- academic staff and students, with lectureships, research grants and undergraduate projects to spare
- miscellaneous others who dabble or spend their waking hours working on code.
The utopian idea of any of these individuals developing for the greater good of mankind is unrealistic: each has a goal in mind that may be commercial or personal. It is absolutely not be the case that every release, by a vendor, of code to the community is seized upon with delight and absorbed automatically into the great open source repository in the sky. Two facts are clear about the release of code – that it does not imply that a vendor is going to stop working on it (rather, that any work will be more transparent), and that there is no real loss of ownership. Just as Linus Torvalds still has power of veto over the Linux kernel, so Sun and IBM will still keep control over their own offerings.
In fact, when a large company “opens” its code, it amounts to little more than opening up its APIs. Sure, you can see the code and modify it if you like, but only of you are willing to take the time and effort to understand it in detail, not to mention construct the development and test environments necessary to enhance it. Anyone that has worked on file system development, particularly something as complex as AFS knows exactly how unlikely it is that anybody will understand the code, let alone want to modify it. In other words, IBM is nodding in the direction of the open source movement as a whole rather than facilitating anything in particular. Some companies may choose to download the code, and may even decide to build on it but it is unlikely they would do so without working in partnership with IBM. So – if it isn’t that generous an arrangement, what is it about? Making money, of course, though not directly.
There is nothing new here. Companies have been using the open source model for perceived competitive advantage for many years. Novell, for example, is releasing the source of parts of (note that) NDS in an aim to dominate the directory space against its main competitor, Microsoft. As we reported \link{http://www.it-director.com/99-09-01-3.html,here} Novell's own bugbear is TRG, which offers a NetWare-compatible file system as a free download from its \link{http://www.timpanogas.com,web site}. This product is open source and comes with the thinly disguised aim to take away Novell's market share. Similarly, it is no secret that IBM is putting its back into Linux, despite having perfectly workable Unix-a-likes of its own (and probably scuppering the 64 bit AIX replacement, Monterey, in the process). Its motives are to reduce the costs of infrastructure software and make life very difficult for companies who depend on such things for the bulk of their revenue.
Open source is not the sub-culture that so many would love it to be. This sub-culture exists, but it is by far a minority in an open source world that is controlled, as ever, by the corporations. The ultimate success of open source, to be adopted by the mainstream, will also be its doom.
(First published 1 September 2000)