Free Software Matters:
The Patent Problem

Eben Moglen

October 9, 2000

Free software is about open standards. The high quality and reliability of free software emerges naturally and directly from the complete sharing of all information and the opportunity to improve the product. Closed protools and closed code radically reduce the number of people who are available to develop, test and repair software, which in turn means lesser quality and reliability. And because code is not freely available for modification, someone who has a better idea has to start over from scratch, which means that fundamental improvements happen much more slowly.

These are pretty basic observations, for anyone who has been familiar with free software for any length of time. But simple as they are they cut completely against the basic doctrine that underlies the law of intellectual property. Rules that give someone a monopoly over an idea or an expression are most often justified on the basis of the ``incentive'' for creation they allegedly produce. In the United States, this principle is made explicit in the Constitution, which provides that Congress has the power to make patent and copyright law ``to promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.'' But free software challenges the identification of exclusive rights and promotion of progess.

There is a crucial distinction between the way patent and copyright concepts respond to the challenge free software poses. Copyright law, as the US Supreme Court has said, is primarily intended to cover expressions, not ideas. So if in a particular instance software copyright inhibits progress in making better, more reliable, or more effective software, the inhibition can be overcome: It is always possible for programmers to sit down and rewrite from scratch whatever program needs to be available in a freely-modifiable version. This may be time-consuming, but it cannot be forbidden. But patent law prohibits anyone from ``practicing the teaching'' of the patent--or using the technological ideas the patent describes--without license. It does not matter how you came by the idea the patent discloses, even if you invented it for yourself in complete ignorance of the patent and the prior art it references: Without a license you cannot embody your idea in any way covered by the patent's often very general claims.

Software patents, then, can achieve the complete reverse of what the patent system is supposed to be for. Independent reinvention can be prohibited, and the owners of inferior programs of bad quality can prevent anyone from making better and more reliable versions, and can make fundamental improvement in design all but impossible.

In the United States, software patents have already become a terrible problem. The US Patent and Trademark Office has been granting patents on programming ideas with wild and uncritical abandon. Basic ideas that have been in use for decades have suddenly been awarded to one implausible claimant or another, who thus gains the leverage with which to threaten everyone else doing what has long been known and familiar in the trade. Commercial manufacturers respond by buying licenses, which are far less expensive than lawsuits, or by ``cross-licensing'' their own ridiculous patents in return for peace and quiet from the demands of other lucky lottery winners. But free software developers cannot afford to buy licenses.

Why this outpouring of patents? In the US, the standard for granting a patent is that the technological idea it contains must be novel, that is not already anywhere in use, and non-obvious, which means not an evident incremental extension or improvement to the body of knowledge readily available to those ordinarily skilled in the art. It is often said that the Patent Office has been unable to hire and retain examiners deeply knowledgeable in the art of programming, because competition for such people elsewhere in the economy has been so intense, and that the Office has therefore been easily led by outside patent lawyers to the conclusion that developments were novel and unobvious that were in fact nothing of the kind. Others have speculated, more darkly, that the bias towards issuing patents has been a form of covert government subsidy to politically-influential industries and firms.

However one views the record in the US, however, the situation elsewhere has been rather different. The countries of the European Union have shown less enthusiasm for patents exclusively involving software. But the US government is stepping up pressure on other nations to conform more closely to its patent policies, and even to give effect to its patent judgments in their own courts. The intention is not specifically to advantage US software companies: the pharmaceutical industry is even more determined to eliminate diversity in international patent law. But the effect of a further expansion of American-style software patents could be a lessening of the quality and inventiveness of software around the world, and a strengthening of the hand of companies like Microsoft at the expense of the users and programmers everywhere who collaborate to make and improve free software.

What can be done about the patent problem? Certainly those of us who understand the free software movement's enormous importance to the spread of the Internet revolution's benefits can participate in the political system in our home countries, raising questions about whether unrestricted patenting of computer programs promotes or retards technological progress. But there are other measures that the free software movement can take to protect users and programmers against inappropriate interference from the patent system, which I will discuss next month in Free Software Matters.

Eben Moglen