The Free Software Foundation was founded when software was becoming increasingly proprietary and restricted, with software vendors asserting more and more control over what customers could do with their software.

Resisting that trend, FSF founder Richard Stallman wrote that software should be "Free as in Freedom"; users should have the freedom to use software for any purpose, to have source code available, to modify the source code, and to share their improvements.

Although today software is not as "free" as Richard Stallman would wish, in many ways his broader vision has come to pass. Thousands of useful programs are widely available and, if not "free," at least open source. Many such programs enjoy widespread commercial use. Software that is developed under review by the community is believed to be of higher quality because of the number of experienced eyes finding bugs. Cost-free availability promotes reuse and proliferation, and so open source software can set de facto standards due to network effects.

The benefits of open source are so appealing that many commercial vendors, especially those who are not market leaders, have released proprietary code to the community. Examples range from the Netscape browser to the Real Networks' player, and most recently Sun's announcement that Solaris will go open source. These companies typically do not sell their open source code. Instead they derive revenue from the maintenance and support business or from compatible products, trading revenue for the benefit of community involvement in developing, debugging, distributing, and marketing their products. They also put tremendous pressure on a proprietary market leader, who must provide an overall better offering without the free help of the community.

A market leader facing serious open source competition must adapt to match the advantages of open source. We've seen Microsoft searching for a compromise with its "shared source" program, and now it is testing the waters with open source releases to the community. Within five years - you heard it here first! - Microsoft will, I predict, release most of what we now know as the Windows operating system as open source. Not out of altruism or a belief in software freedom, but to prevent Linux or open source Solaris from becoming a standard. Applications and special functionality that sit on top of the operating system will likely remain proprietary.

You might think that this would represent the ultimate victory for free software. Indeed, picturing the über-capitalist turning over its flagship product to the community seems revolutionary today. But, by then, it won't be a big deal to Microsoft, because its proprietary platform will have moved "up the stack." Let me explain.

Professor Clayton M. Christensen has hypothesized that the commoditization of a technology product pushes proprietary development (and the value) elsewhere in a product chain: either down to components of the commodity product or up to aggregations of the commodity product with other products. He calls this the "Law of Conservation of Attractive Profits." Open source represents the ultimate commoditization of software, Tim O'Reilly has observed, and value in the computer industry is now moving into networked applications like Google, Amazon, and eBay, which make use of open source.

The impressive-sounding features Microsoft touts as long-term goals in discussions of .NET, Longhorn, and Blackcomb - such as identity management, and information aggregation, search, and retrieval - will be implemented as server-side and/or networked applications. The applications will create a new platform for Microsoft computing, one that doesn't require a proprietary base operating system. The code and the interface that provide the new, innovative networked applications, and the data underneath them, could remain proprietary on top of an open source base.

At that time, even more open source software (including parts of Windows) will be available to all, but the new, noncommodity computing functionality will still be proprietary. The community will need to figure out how to provide alternatives to those proprietary networked application service offerings - not impossible by any means, but a new challenge. In the fight for free software, these developments will mark great victories, but the war will not yet be over.