As this issue of Linux Magazine goes to press, it’s been only a few days since the defeat of the Computer-Implemented Inventions directive — the so-called “software patent” directive — by the European Parliament. But it’s not too soon to reflect both on what happened and what still lies in store for software developers in the future.
First, a little history on software patents in Europe.
The European Patent Office (EPO) does not operate under the auspices of the European Union (EU). Instead, it’s an organization formed by convention (the European Patent Convention) by a number of European countries with the intent to bring uniformity to patent practices among the patent offices of the signatory countries.
Article 52 of the European Patent Convention states that programs for computers are not patentable as such. While interpreted by many to mean that software isn’t patentable, both the EPO and a number of its member country patent offices have been issuing so-called software patents for a considerable period of time. However, there’s been no consistent practice among all EPO member countries, and there’s been a good deal of uncertainty whether software patents are even enforceable. So, the European Commission stepped into the picture to propose legislation intended to bring harmony among EU countries on the subject.
The efforts of the EC were soon perceived by many independent programmers and small and medium enterprises, including members of the open source community, as an attempt to strengthen the practice of granting patents on software without addressing whether such patents should even be granted in the first place. Without recounting each step of the political process over the last several years, suffice it to say that, in the end, the interests of large industry players holding significant numbers of so-called software patents and the interests of enterprise and average citizens were at odds. The former wanted certainty and enhanced value without concessions; the latter wanted to end the practice of granting software patents consistent with Article 52.
While the European Parliament’s decision to reject the directive may appear to be a draw, with neither side getting the upper hand, that’s not actually the case. Only a few weeks ago, it appeared certain the issue was going to be steamrolled through the European Parliament, shepherded by highly-compensated industry lobbyists. The fact that the directive failed is in no small part due to the grass roots efforts of organizations like the Foundation for a Free Information Infrastructure and the European satellite of the Free Software Foundation.
While important in demonstrating the political muscle of small- and medium-sized businesses, independents, and the open source community, the successful defeat of the CII directive should not be viewed as the last battle. There are still many critical questions to be answered…
*Are patents really necessary to promote innovation in software?
Despite this common misperception, there appears to be no empirical evidence to support this contention. However, there are a number of studies that contradict it, including The Software Patent Experiment
by James Bessen and Robert Hunt (available online at http://www.researchoninnovation.org/softpat.pdf
*Does software merit protection by both copyright and patent law, unlike any other form of art? No other art or invention enjoys this dual privilege. Books are copyrighted. Mechanical devices are patented. Music is copyrighted. Chemical compounds are patented. If software is so unique that neither copyright nor patent alone are adequate to protect it, perhaps it requires its own novel means of protection. The concepts of copyright and patent are more than two hundred years old and neither contemplated the digital age.
*Does the extended life of either patent or copyright protection encourage rapid development or delay it? Patents and copyrights are public grants of a limited monopoly. As stated in the Constitution, they are provided to advance science and the useful arts. When the cadence of invention followed the pace of the 19th Century, a patent that extended for 17 years was not unreasonable. But in an age when Moore’s Law still defines the information technology industry, 20 years of patent protection represents 5-7 product life cycles. A simple examination of Microsoft release cycles for Windows offers the proof. To really encourage innovation, protection for software should be limited to no more than two product life cycles or about 6 years.
*Are such limited monopolies, as provided by patents, necessary to adequately reward software development? Anyone who pays attention to the costs of developing and obtaining United States Food and Drug Administration (USFDA) approval of a new pharmaceutical knows full well that, without the extended protection of patent, few companies would be willing to gamble hundreds of millions of dollars over a decade. But those numbers are foreign to the software industry, where the cost of obtaining a patent (typically about $35,000 in the U.S.) far outweighs the direct cost of the invention. Such invention costs could easily be recouped with a reasonable profit within a much shorter time period.
Patent reform in either the Europe or the United States is all well and good — if it benefits all participants, including consumers. But patent reform without addressing fundamental questions such as those listed above is leading the software industry down a self-destructive path.
Mark H. Webbink is the Deputy General Counsel of Red Hat, Inc.