TCS Daily

Open Agnosticism

By James V. DeLong - December 10, 2002 12:00 AM

Julian Sanchez characterizes me as an "enemy" of Open Source (OS). "Open source agnostic" would be a better term than "enemy." Opposition to proposals for government preferences are not necessarily based on enmity for the Open Source movement, which has some impressive strengths and has produced some significant accomplishments. But I do not share the messianic dedication that infuses many of its proponents.

The movement must be given due credit. Much fine software has been developed by programmers at universities, governments, and private companies, working cooperatively, either on their own time or with the blessings of their employers. A large group of smart people each devoting a small percentage of his time, linked by the Internet, can be highly creative. The results offer many possibilities for the future.

But the elephant in the living room, seldom mentioned by OS advocates, is that the programmers must be supported somehow. Programming is hard work. It requires substantial investment in human capital to get up to speed, and it requires continuing support while the work is performed. Programmers are not sleeping on steam grates and flipping hamburgers for a living while they toil. They are paid, either directly for writing code or for some other product that is sold on the market.

Because of this need for money in the system somewhere, the OS movement has several different parts:

  • The Free Software Foundation, which is the most ideological part. The purists believe that software should not have owners, should be produced entirely by cooperatives, not for the market, and should be free.

  • Academicians have complex motivations. Some are FSF believers. Many are simply concerned with producing good tools for their own use and that of their colleagues, and see no reason they should be limited by the incentives, schedules, and limitations of the commercial software producers. They "do software," the way scientists "do science" and other academics do their thing. Academics are not Franciscans, btw. A tenured chair at a major university is one of the most valuable pieces of property one can own in modern America.

  • Corporate IT professionals share many of the characteristics of academicians. They like software, enjoy exercising their skills, and belong to an exciting community of like-minded aficionados. Open Source is useful and fun. It also makes their jobs easier, because they can adapt it to their particular needs without going through a vendor's help line or waiting interminably for a bug fix. One respondent to an earlier piece of mine described part of the reason for going with OS: "Then there is Support. MS has been so bad we do not even try anymore: we know more than the people on the phone do. Wait on hold go through procedures pay $$$ and then be told things you already know. . . . Linux support, run by volunteers, gets me replies from subsystems' authors within hours. Try that with "big corporate" software."

  • Software developers who want alternatives to Microsoft. Many programmers want to write and sell software without being limited to Windows. They support Linux, the Open Source operating system, but they do not want to extend requirements that software be open to the applications that they write.

  • Computer hardware makers, who see Open Source as a way to avoid sharing the computer business with proprietary software companies. IBM, in particular, is putting billions into Linux. IBM then makes its money through sales of hardware, services, and proprietary aps that run off Linux.

  • Sellers of Open Source-based services. A slew of companies, most famously Red Hat, sell Linux and support for it. They cannot sell the program itself, but they can charge for its distribution, for providing warranties, and for follow-up consulting services.

  • Corporate, government, and individual free riders. Many denizens of these worlds are enthusiastic about OS software. Some regard it as of higher quality. As noted above, many IT people have good reason to like it. And, above all, CFOs like it because they do not pay the costs of production, which makes it cheaper.

Yet another distinction must be made. Open Source software comes in two basic flavors: non-viral and viral:

  • Non-viral means that the software is in the public domain. Anyone can use it for anything, including incorporation into proprietary applications. Of course, it cannot be appropriated by the user, but it can be freely mixed with software that remains the intellectual property of its creator. The Berkeley Software Development License may be the best-known example.

  • "Viral" is the term coined by Richard Stallman of the FSF to describe Linux and other software covered by the General Public License. Under the GPL, the licensee agrees not to sell or otherwise limit the reproduction of the software (though one is allowed to charge for costs of distribution, warranties, and services, as noted above in the discussion of Red Hat). It also provides that any software that incorporates GPLed software also becomes subject to the GPL. This creates problems for writing applications, and led to the creation of a Lesser GPL that eases this provision. (These licenses are available at the FSF website.)

Logically, given the motivations of the players, it would have made sense for the GPL to be limited to the software connected with the FSF movement and for a non-viral license, such as the BSD, to be used by other parts of the movement, which are more receptive to the need to make money somehow. But that is not what happened; Linux, the linchpin of the Open Source movement, is under the GPL. This does not make it impossible to combine Linux with moneymaking operations, but it vastly complicates the effort.

In this chaos, the fundamental argument for the primacy of Open Source would run roughly as follows: The academicians and IT professional will keep producing it for their own purposes. Once software exists, the cost of reproducing it is close to zero. Therefore, why not distribute it to corporations, individuals, and governments at this marginal cost?

It is an interesting argument, but I am not convinced.

First, I am not convinced of the initial premise, that a cooperative effort to produce software is superior to commercial efforts even for the professionals and researchers. As computers mature, everyone might be better off with more division of labor, and even the academics might prefer to deal with professional developers rather than with a coop. They will insist on access to code, however, and on responsiveness. Microsoft itself is recognizing this and is moving toward "shared source" with some of its customers.

In addition, it is hard to see the cooperative effort working over a period of years in an environment in which hardware changes continually and software must be modified in response. Not all code writing is fun. Much of it is good old-fashioned scut work, of the type that does not lend itself to volunteer efforts. The OS advocates need to study a little more history - perhaps Max Weber on the problems of the routinization of charisma.

Beyond this, a model that depends on free riding by corporations, individuals, and governments is not likely to be sustainable. After a time, the programmers are going to notice that they are producing something of great value to all these groups, and that there is no reciprocity. For example, why should programmers provide free software to General Motors? Is GM going to give them cars for nothing? Will the IRS remit their taxes in consideration of the value received? The system will break down over the lack of balance, especially if extra scut work must be performed continually to keep the software useful to the free riders.

Cooperative efforts are fine things, but the assumption that this means that we dispense with money and markets is fallacious. Money and markets are precisely the way in which our society administers cooperative efforts. Programmers write software for GM, which makes cars for the programmers. Reciprocity, with money as the administrative mechanism. Markets are not an alternative to cooperative efforts; they are the best mechanism for promoting them.

Of course, it may turn out that the corporate supporters of OS put in enough money to keep the programs current, as a part of their efforts to sell hardware and services. This is certainly their privilege, and it may turn out that this system - which is simply a form of bundling of open software with proprietary hardware - is a viable model. But it is only one possible model, and it also may be that proprietary software produces a superior product.

There is no reason for governments to pre-judge the issue because of a view that OS is somehow morally superior. It is not. So let the playing field remain level, and let's see what happens.

James V. DeLong is Senior Fellow and Director of the Center for the Study of Digital Property at the Progress & Freedom Foundation in Washington, DC.

TCS Daily Archives