A software patent is a patent on an invention implemented in software.

Software patents are very controversial. For many decades, patent offices around the world rejected most applications for software patents. In Europe, the European Patent Convention stated that there could be no patents on software "as such".

The exclusion of software from patentability did not suit the interests of many patent professionals and certain software firms, who continued to seek routes to exclusive rights over algorithms and more general software systems. Gradually, cases began to appear in various jurisdictions (such as the United States, Japan and Australia), holding that software could be patented in various ways. The European Patent Office (responsible for granting European patents, and separate from the European Union) decided that it could grant patents on software using a politically controversial interpretation of the European Patent Convention.

Had the story remained typical of the history of intellectual property laws, the alignment of intention between key corporations (especially IBM and Microsoft) and the patent offices of the US, Europe, and Japan, would soon have lead to mandatory software patents under international law.

At the present moment, however, armed with evidence suggesting that software patents are likely to be economically harmful, coalitions of interest groups including the free software and open source movements, many software firms without large patent portfolios, and many non-software businesses, are attempting to reverse the trend of patent expansionism. This conflict has been played out over the EU Directive on the Patentability of Computer-Implemented Inventions.

Table of contents
1 History and Rationale for Software Patents
2 Opposition to Software Patents
3 Dealing with Software Patents
4 Notorious Software Patents
5 References & Resources

History and Rationale for Software Patents

In the 1950s, 1960s, and 1970s, the United States Patent and Trademark Office (PTO) did not grant a patent if the invention used a calculation made by a computer. The PTO's rationale was that patents could only be granted to processes, machines, articles of manufacture, and compositions of matter; patents could not be granted to scientific truths or mathematical expressions of it. Since the PTO viewed computer programs and inventions containing or relating to computer programs as mathematical algorithms, and not processes or machines, they were therefore not patentable. This view was upheld by the U.S. Supreme Court in Gottschalk v. Benson (1968) and Parker v. Flook (1975).

In the 1981 case of Diamond v. Diehr, the U.S. Supreme Court ordered the PTO to grant a patent on an invention, even though computer software was used. The invention in this case involved a method for determining how rubber should be heated to be best cured, using a computer. The Supreme Court stated that in this case, the invention was not merely a mathematical algorithm, but was a process for molding rubber, and hence was patentable.

After this point, more patents on software began to be granted, albeit with conflicting and confusing results. The Federal Circuit attempted to clarify the rules; requiring that the computer program must have a practical application. However, since all software is written to perform some useful activity, many believe this to be the exception that swallows the rule.

Meanwhile, the Clinton administration pushed software patenting from the administrative agency side, by appointing Bruce Lehman as Commissioner of the Patent and Trademark Office in 1994. Unlike his predecessors, Lehman was not a patent lawyer but the chief lobbyist for the Software Publishing Industry. In 1995, the PTO established some broad guidelines for examining and issuing software patents. The PTO interpreted the courts as requiring the PTO to grant software patents for an extremely broad variety of circumstances, including those that are essentially algorithms only distantly connected to physical processes. Note, that although the US Congress has never legislated specifically that software is patentable, the broad description of patentable subject in the Patent Act of 1952 and the failure of Congress to change the law after the court decisions allowing software patents, has been interpreted as Congressional acquiescence.

Another impetus for software patenting was the growing recognition that using the copyright law to protect non-literal infringement of computer programs (rather than just piracy) was getting out-of-control. When comparing patent protection to the use of non-literal copyright infringement, many commentators argued that many protections for competitors are built into the patent system that are lacking in the copyright laws. Specifically, these commentators pointed out copyrights are not examined, but patents must first be examined to determine if the program is both novel and non-obvious; the scope of patent rights is defined by the patent claims, while the scope of non-literal copyright infringment is unclear; and the patent term of 17 or 20 years is much shorter than the copyright terms. When courts began to permit software invention to be patentable, other courts also began restricting the use of copyright law to obtain patent-like protection of software.

Those who favor software patents believe that software are inventions to the same extent as hardware and that the law should and, in practice is not able to, distinguish software inventions from hardware inventions. Proponents also argue that the patent system rewards inventors of innovative approaches in software, and thus promote innovation. This belief is important in the US, because this is the only permitted reason for a patent to be granted according to the US Constitution. More specifically, the Constitution only permits Congress "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."

Opponents charge that software patents are particularly favored by lawyers, who financially benefit from patent litigation, and by some (though not all) very large software companies, who hope to use patents to prevent competitors from using the patented technology.

Opposition to Software Patents

However, there remain many opponents of software patents, including an overwhelming majority of professional software developers. For example, Burton Systems Software conducted a survey of professional programmers, and found that by a margin of 79.6% to 8.2% (10:1), computer programmers said that granting patents on computer software impedes, rather than promotes, software development (the remaining 12.2% were undecided). By 59.2% to 26.5% (2:1), most went even further, saying that software patents should be abolished outright.

Opponents of software patents argue against them for a diverse range of reasons. Here are some of the reasons opponents give for opposing software patents:

  • There is no evidence that software patents actually encourage innovation. The 1950s, 1960s, and 1970s included a large number of software innovations, when software patents were not permitted. These innovations can be measured both as published papers and as new kinds of products.
  • Many in the computing field believe software patents actively impede innovation. In 1991, Microsoft's Bill Gates wrote a memo saying, "If people had understood how patents would be granted when most of today's ideas were invented and had taken out patents, the industry would be at a complete standstill today." (Mr. Gates' company now acquires a vast patent portfolio, since to do otherwise would be suicidal, and that portfolio may be helpful in preventing competition). Donald Knuth, a highly-respected computer scientist, stated that "If software patents had been commonplace in 1980, I would not have been able to create [the TeX system used by 90% of all books and journals in mathematics and physics], nor would I probably have ever thought of doing it, nor can I imagine anyone else doing so."
  • Professors Bessen and Maskin, two economists at the Massachusetts Institute of Technology (MIT), have demonstrated that introducing patenting into the software economy only has economic usefulness if a monopoly is the most useful form of software production. This is concerning, because few believe that a monopoly is truly the most useful (or desirable) form of software production. Bessen and Maskin also demonstrated a statistical correlation between the spread of patentability in the United States and a decline in innovation in software. In particular, between 1987 and 1994, software patents issuance rose 195%, yet real company funded R&Ds fell by 21% in these industries while rising by 25% in industries in general.
  • Some believe that the problem besetting the software field is not a lack of innovation, but difficulty in developing the large number of desired products. The patent process interferes with, not aids, the development of useful products.
  • They believe the standard for "obviousness" in other fields is inappropriate for software. Because software is malleable, small, incremental changes and generalization are normal and obvious to practitioners. However, the PTO normally grants patents to small, incremental changes, even if they would be obvious to practitioners.
  • Many techniques are considered too obvious to publish by practitioners. However, a patent may be granted later by the PTO, because no paper was found by the PTO discussing the topic.
  • Some believe that switching from a copyright-based system to one permitting patents puts established experts at a severe disadvantage. Experts cannot patent many concepts because they are obvious (and sometimes verbally shared among peers)—yet they can be patented by novices because they are not as obvious to novices. Dan Bricklin, inventor of the spreadsheet, is a well-known proponent of this position.
  • Many believe that the cost structures for software development are fundamentally different. Extremely complex software systems with hundreds of thousands of parts are often built for small amounts of money compared to physical products. However, the costs of dealing with the patent system presumes that complex systems will result in large profits, on the order of those for physical products. For most software systems, this simply is not true.
  • Many believe the risk of a lawsuit greatly reduces the incentive to innovate new products. This risk is exacerbated because software patent searches are prohibitively expensive and unreliable. Besides, patents may be granted to another after the software has already been written, so even a perfect search would not prevent risks to software developers.
  • Patent licenses are especially harmful to open source software / Free software, which are becoming an increasingly important type of software and in many markets are the only alternative to no software or establishing a permanent monopoly in a functional area.
  • Software patents reward those who employ a deceptive practice known as submarine patents. In this approach, the patenter files for a patent and ensures that it is not made public by the PTO for some time through various paperwork processes, or simply words it so that it is not noticed by the community the patent would apply to. Patenters then attempt to ensure widespread use of the patented approach, e.g., by working with standards bodies and implementers to use the approach. Then, once the approach is widely used, they then announce the patent and sue all users, who will find it difficult to switch to other approaches once they are widely embedded.
  • Patent licensing strongly discourages, and in somes cases prohibits entry of newcomers into the software field. Large companies collect patents and attempt to force cross-licensing with others to protect themselves from software patents. But this means that small companies, without a large body of patents to cross-license, may be forced to license from a large number of companies to develop software at all. The total of these royalties could exceed all possible benefits, permanently blocking newcomers from the software field.
  • Small litigation companies (whose only contribution is to buy patents and sue other companies) can threaten large companies, even if those companies cross-license patents. Thus, even large companies can be at risk of a patent suit. However, these companies may exist solely to create patents of previously existing or obvious ideas, and litigating these patents can be more expensive than the product is worth.
  • Patent officers tend to be paid less than they could make doing other activities in software, so they tend to be less skilled. In addition, they must be generalists, so they are unlikely to be aware of well-known approaches in any particular area.
  • Databases of prior work are inadequate for the task of determining if something has already been done before.
  • The patent process has little incentive to identify pre-existing work. The process rewards patent requesters who do poor research, since by doing poor research, they will not find preceding work that would invalidate the claim. However, since patent officers tend to be less skilled, have inadequate databases, work under significant time pressure, and must of necessity be generalists, it is difficult for them to find preceding work. These resultant patents can still be useful to patent-holders as threats, since court cases are expensive and uncertain.
  • Patent offices are notorious for granting absurd patents, yet once they are granted they can be enforced by simply the threat of an expensive lawsuit. For a non-software example, Patent 6,368,227 is a patent on a particular method for swinging on a child's swing, one that has no doubt been used by children for decades. In Australia, one man patented the wheel.
  • The patent system diverts many able-bodied experts into processing patents instead of innovating.
  • Patent litigation is extremely expensive, and owners of patents that should never have been issued can nevertheless impede innovation or cause others to pay unnecessary fees to avoid the cost of litigation.
  • The term of patents (20 years in the US) is considered by many to be inappropriately long for software; software is quickly superseded and unable to be used long before hardware wears out.

Software patents tend to be opposed by individual software developers, who view software patents as a risk to their livelihood: if enough patents are granted, they will not be able to practically develop software. Some large software companies also oppose patents, fearing that they will be sued for implementing obvious techniques, resulting in continuous payments to avoid court costs or steep fees for court battles. Well-known opponents of software patents include Richard Stallman (author of the gcc compiler), Dan Bricklin (inventor of the spreadsheet), Donald Knuth (an expert on computer algorithms and the author of the TeX typesetting software), Hartmut Pilch, Eurolinux Alliance, Mitch Kapor, Michel Rocard (former Prime Minister of France), Adobe and Oracle.

Dealing with Software Patents

Most software development companies have decided to acquire software patents, even if they oppose the granting of them. Their motives include acquiring a patent before someone else does, or forcing competitors who acquire patents on obvious approaches to cross-license with them. Often these patents are only used defensively, e.g., they are only used against someone who first sues the company. Some organizations and licenses have formalized a nonaggression policy (a policy of never pursuing or profiting from aggressive software patent suits) and/or of mutual defense (in which a pool agree to this). Such systems, however, provide little defense to individual developers or small businesses, and it is unclear if they will prevail once companies come into financial hardship, needing patent revenues to persist. Often a patent can be worked around once the patent is known, but this can be a significant hardship if there is a significant amount of data in a format requiring the use of the patented algorithm.

A recent concern is the role of patents in the standards process. Some standards bodies have no patent policy; thus, it is possible for a member to convince a standards body to make certain technologies required by a standard while at the same trying to get a patent on that techonology. As a result, many standards bodies (such as W3C)are now requiring their members to promise to grant either reasonable and non-discriminatory (RAND) or even royalty-free licenses on their patented technology that is incorporated into the standard.

Notorious Software Patents

Notorious software patents include:

  • Unisys acquired a patent on LZW compression, a fundamental part of the widely-used GIF graphics format. Unisys published the compression algorithm, but did not make it clear that it was applying for a patent. CompuServe then developed a graphics format, and, having no reason to choose otherwise, chose the LZW compression algorithm. Unisys waited until the GIF algorithm was widely used, and then demanded royalties for users and developers who until then had had no idea of the issue. It is uncertain if Unisys delayed intentionally or not, but it is clear that many developers were very angry at the Unisys response. This resulted in re-development of a new graphics format, PNG.
  • The popular music format MP3 is encumbered by a number of patents, which were only enforced once the format became popular. This resulted in several GNU/Linux systems dropping support for MP3, and in re-development of new audio formats—notably Ogg Vorbis.
  • British Telecom sued Prodigy under U.S. Patent No. 4,873,662 claiming that Prodigy infringed its patent on web hyperlinks. However, after costly litigation, a court found for Prodigy, ruling that British Telecom's patent did not actually cover web hyperlinks. [1] Hyperlinks were first described in 1945 in the landmark paper As We May Think, as well in the widely-known project Xanadu starting in the 1960s.
  • Cadtrack's US Patent 4,197,590 covers drawing a cursor on a screen using XOR, which allows removal by XORing it again, thus eliminating the need for backing store.
  • Amazon vs. Barnes and Noble: "One click buy"
  • EOLAS vs. Microsoft: "browser plugin patent"
  • Compton's NewMedia was awarded a submarine patent in 1993, which had been filed five years before, for "A search system [that] uses a multimedia database consisting of text, picture, audio, and animated data." Compton's announced their patent at the height of the excitement over CD-ROM software and claimed this patent covered all multimedia software, and announced a royalty payment schedule. An outcry ensued that this was an attempt to patent something that had been in active use for many years, and the furor was so great that the Patent Office commissioner quickly started an "office action" to re-investigate the claim. It seemed that the PTO had only reviewed existing patents for prior art, and not the wide body of prior art in the field that had not been patented. The patent was voided in 1994.
  • Stac Electronics sued Microsoft for patent infringement when Microsoft introduced a data compression scheme into MS-DOS which resembled Stac's Stacker software. Stac was awarded $120 million by a jury in 1994 and Microsoft was ordered to recall versions of MS-DOS with the infringing technology. Subsequently Microsoft and Stac settled the case; Microsoft promised not to appeal, paid Stac $43 million, and purchased $40 million of preferred Stac stock.

References & Resources

Legal Resources

Groups against software patents

Groups in favor of software patents

  • large software and IT companies having built up a stock of software and other patents
  • Patent lawyer associations

Neutral resources