In 1903, a patent attorney named George Selden sued Henry Ford and four other car manufacturers, demanding a royalty on every car sold. Although he had done nothing to advance automotive technology, Selden held a patent claiming to cover the automobile.
Patent trolls are nothing new. In the nineteenth century, they were called “patent sharks” and they went after farmers in large numbers, armed with patents on barbed wire and sliding gates. But lately, patent trolls have become a hot-button issue in a way they haven't been for decades.
That's because trolls are filing an unprecedented number of expensive lawsuits. Over 5,000 firms were named as defendants in patent troll lawsuits in 2011, costing them over $29 billion out-of-pocket. Today’s patent trolls are wreaking damage on a scale not seen in the past. And there's a specific reason for this: The last two decades saw a dramatic increase in the number of patents on software, and these patents are particularly prone to abuse, both by trolls and by other types of patent holders.
Policymakers are increasingly focusing on the problem of frivolous patent litigation. But so far, policymakers haven't given enough attention to the fact that the patent crisis is mostly about patents on software. Understanding that fact is essential to fixing what's wrong with the patent system.
Why software patents lead to litigation
Last week, Tim Lee wrote about a recent study from the Government Accountability Office found that a disproportionate share of patent litigation concerns software patents. The non-partisan government agency found that the number of defendants in patent lawsuits more than doubling from 2007 through 2011. Notably, this increase is specifically related to software patents — software patents account for 89% of the increase, according to the GAO's calculations.
So why are there so many lawsuits over software patents? The report states that “many recent patent infringement lawsuits are related to the prevalence of low quality patents; that is, patents with unclear property rights, overly broad claims, or both. Although there is some inherent uncertainty associated with all patent claims, several of the stakeholders with this opinion noted that claims in software-related patents are often overly broad, unclear or both.”
It is not hard to find examples of software patents in lawsuits that are unclear and overly broad. A company called Lodsys has been threatening to sue hundreds of smartphone app developers for patent infringement. Its patent cover “Methods and systems for gathering information from units of a commodity across a network.” App developers have little idea what that means and the specific claims within the patent are written in similarly abstract language. Needless to say, the patent’s inventor did not actually develop a practical technology based on this patent and the original patent filing was made in 1992, long before there even were smartphone apps.
In other cases, the patents relate to actual inventions, but are specified in such broad language that they can be interpreted to cover other technologies that are developed later. For example, one inventor developed a kiosk to be placed in retail stores for producing music tapes from digital downloads. During the early 1980s, he filed a patent claiming to cover an “information manufacturing machine” at “a point of sale location.” E-data, a patent troll, later acquired the patent and aggressively interpreted this language to cover digital e-commerce, collecting millions of dollars from all sorts of companies.
Both overly broad patents and unclear patents foster unnecessary litigation. An overly broad patent allows a patentee who did not really develop a new technology to sue the actual innovators. Unclear patent boundaries mean that innovators cannot avoid inadvertently infringing a patent. Also, unclear boundaries mean that patent examiners cannot really tell whether the patent is novel — if you don’t know exactly what the patent covers, you can’t tell if it is different from previous inventions.
For example, there are dozens of patents on ways to unlock smartphones. Some use the words “slide to unlock;” others use words such as “activate the function.” Neither patent examiners nor technology developers can tell exactly what these patents cover. Software patents with unclear boundaries are often invalid because they cover something already invented or are an obvious improvement over what was previously invented. Indeed, one economist estimates that at least 38 percent of software patents are at least partially invalid for these reasons; 59 percent of patents used by trolls are likely invalid. Moreover, there are now hundreds of thousands of software patents, making it impossibly expensive for developers to do a pre-development clearance search.
Why software is different
Fuzzy boundaries are not unique to software. Patents on mechanical devices and other technologies can also be written in vague, expansive language. These often cause lawsuits, too. But software patents are particularly prone to such abuses because software is inherently conceptual. Software is a technology that represents broad classes of interactions abstractly. That makes it inherently difficult to tie down a software patent to a specific inventive concept.
In contrast, mechanical devices are tied to tangible objects. If it is not clear what a word might mean in a patent for a mechanical device, the device itself serves as a reference. Indeed, in the past, the Patent Office required that patents on mechanical devices were accompanied by working models. If the model worked differently from some other device, then that other device did not infringe. It's similar for chemical patents: A chemical patent’s boundaries are determined precisely by the structure of the molecules it claims. If another chemical has a different structure, then it does not infringe the patent.
But for software patents, the conceptual description alone is sufficient to obtain a patent. Computers programs designed to solve dramatically different problems can be covered by the same patent. Patent lawyers make the problem worse when they use vague words, but there is an underlying difference to begin with.
Consequently, software patents are much more likely to have fuzzy boundaries and they are much more likely to be involved in a lawsuit. Research shows that a software patent is four times as likely as a chemical patent to be involved in litigation; a software patent on a method of doing business is thirteen times more likely to be litigated. Nor are these high rates of litigation a result of the newness of software patents. After hundreds of thousands have now been issued, the probability that a software patent will be used in a lawsuit has gone up, not down.
Abstract patents and the law
Since the nineteenth century, the courts have wrestled with doctrines aimed to exclude patents claiming abstract ideas. Samuel F. B. Morse, in his famous telegraph patent, included an abstract claim for printing characters using electromagnetism. The Supreme Court shot that claim down, while keeping his more specific claims. In 1972, citing the Morse case, the Supreme Court shot down a patent on a software algorithm. This decision was widely understood to exclude patents on most software innovations.
But during the mid-1990s, the appellate court for patents, the Court of Appeals for the Federal Circuit, reinterpreted the Supreme Court's precedents in ways that encouraged patents on software, methods of doing business and mental concepts. The number of software patents granted subsequently soared. The GAO estimates that in 2011 the majority of patents granted were software-related.
The Supreme Court has begun to reassert itself. Two recent decisions have reiterated the ban on patenting abstract ideas. However, judges on the lower courts, particularly the Federal Circuit, have not always been willing or able to implement this ban. Indeed, after the Federal Circuit invalidated a software patent in a recent case, one judge warned, “if all of these claims, including the system claims, are not patent-eligible, this case is the death of hundreds of thousands of patents, including all business method, financial system, and software patents as well as many computer implemented and telecommunications patents.”
Most of those patents probably should not have been granted to start with. But this creates a political problem for the courts. If they strictly enforce the Supreme Court’s doctrines, most software patents will be found invalid, disrupting powerful interest groups. If they don’t, then wasteful litigation will continue to grow, imposing large costs on society, costs that are already inhibiting innovation.
Congress and the White House, under pressure from large numbers of businesses hurt by patent trolls, are also attempting to craft responses. Seven legislative proposals dealing with trolls have been introduced in this session of Congress. Many of these proposals attempt to make it more costly for patent trolls to sue. While possibly beneficial, these proposals do not address the underlying problem. Other proposals do target software patents specifically, calling for a review process to allow these patents to be more easily challenged. These help, but they may not go far enough.
In the past, the scope of property rights has not always expanded when new technologies made boundaries difficult to ascertain. Since Roman times, whoever owned the soil also owned rights “up to the Heavens.” But when landowners attempted to assert their rights against airplane traffic, the courts limited those rights. Today, the explosion of patent troll litigation provides stark evidence that old patent doctrines are not providing clear boundaries for information technology. The resolution will require substantial changes to the way the law treats software patents.
James Bessen does research on technology and innovation at Boston University School of Law, including research on patent trolls. He is the co-author of Patent Failure: How Judges, Bureaucrats, and Lawyers Put Innovators at Risk. You can follow him on Twitter.