CNET también está disponible en español.

Ir a español

Don't show this again

Tech Industry

Open source or no: Let the market decide

The Progress & Freedom Foundation director James V. DeLong argues that government preferences in the awarding of open-source software contracts sets a dangerous precedent that will ultimately hurt consumers.

The latest anti-Americanism to sweep Europe is a broad hostility toward computer software that's produced by proprietary companies, mostly American, and an enthusiasm for "open source" software--programs written by networks of volunteers for which code is free, open to inspection and modifiable by any savvy user. International bodies, national governments and lesser political subdivisions are moving to legislate preferences for open-source over proprietary software. The proposals range in strength from mild boosts to complete mandates. At the moment, some 80 provisions are on the table in 40 different nations, including in the United States, where Oregon, Texas, and New York City are involved.

Offhand, one might wonder why any legally mandated preference would be necessary. "Would you like to pay for this software, or would you rather get it free?" is a question that seems to have only one answer.

However, the choices are more complex than this. For any computer program, license fees are a small part of the total cost of ownership, or TCO. They can be dwarfed by costs that are associated with employee training, bug-fixing and other support, specialized modifications, user needs, long-term utility guarantees and other factors.

For proprietary software, the provision of these services is dictated by the mechanisms of the market. A software developer that wants to stay in business either addresses these issues or it sells no software.

Open source is different. The movement is a hybrid of commerce and ideology. Take the Linux operating system, which is the flagship of open source. Companies such as IBM, which is a huge vendor of proprietary software, can also use Linux as a platform on which to hang proprietary applications. IBM, which sells more proprietary software than anyone except Microsoft, also uses Linux as a platform on which to hang its proprietary applications.

These commercial interests provide crucial financial sinew, but the origin of the open-source movement, and still its dominant ethos, was in an ideological commitment to sharing and to voluntarism rather than crass commerce.

A weakness of any volunteer movement is that people want to do what is fun, so open-source software has been long on neat code and short on documentation and routine stuff. A major contribution of the commercial players has been to ameliorate this issue.

The world benefits from competition between open-source and proprietary software and from the pressure that the strengths of each form puts on the other.
A more fundamental weakness of voluntarism is in the incentives for the volunteers. A close-knit group can cooperate in exchange for mutuality and approval, but when the benefits spread to people who are not contributing, the structure gets tottery. Why should the programmers sweat so that corporations or governments can get free software? Do auto companies give cars to the programmers in exchange? Do governments remit their taxes? After a period of time, the lack of reciprocity begins to grate.

These structural issues create doubts as to the long-term viability of open source as a mode of production. The cooperative turmoil of the computer revolution may have been highly creative, but things change as revolutions mature. A century ago, Max Weber wrote about "the routinization of charisma"--and in time, the motivations of the programmers are likely to falter.

Some established programs may continue to do fine, but it is not clear that the movement is capable of generating major new programs or major rewrites in response to hardware advances, except, of course, to the extent that work on these is supported by commercial interests.

Given these question marks, companies or governments that commit to open-source software are risking their fate on a series of uncertainties, given the lack of incentives in a volunteer effort. An alternative is to calculate that the user can maintain software internally, but this raises the costs of ownership hugely. Or, of course, the user can assume that the commercial interests will continue to subsidize software production and distribution out of other revenues.

This may be a good bet, but it means that the users must pay the costs indirectly, which brings one back to calculations of TCO, and of whether a user would rather pay for software directly or indirectly. This is not at all the same as the dominant assumption touted by the preference advocates, which is that open-source software is "free."

In this context, the movement for legal preferences makes more sense. Their advocates fear that the open-source model will not be able to compete over time, and want to persuade governments to use their buying power to rig the game. And governments, with their inefficient non-TCO-oriented budget processes, are prone to be persuaded.

It would be unfortunate for the preference movement to succeed. The world benefits from competition between open-source and proprietary software, and from the pressure that the strengths of each form puts on the other. Tilting the playing field in either direction is not necessary.

Preferences would also be unfair to programmers, who are, after all, the linchpin of the system. They truncate the market in which the programmers sell their wares, and in effect tell them they must donate their work, not sell it through proprietary companies: "You will volunteer."