The question of whether software should be patentable stirs up strong feelings in the tech community, on both the pro and con sides, a divide that also extends to the law, which has struggled for almost 50 years to provide concrete guidance on how software should be treated by patent law, if at all.
An opinion this past summer by the nation’s highest patent court has brought that debate to the forefront again. The U.S. Court of Appeals for the Federal Circuit split on whether four patents covering a financial services platform covered subject matter that was eligible for a patent. The lack of consensus among the judges reminded us that even among patent scholars, the boundaries of what is and is not patentable can be difficult to draw.
Abstract ideas, mental processes and mathematical formulas are without question outside the boundaries of patentable subject matter. Software is essentially a collection of abstract ideas, algorithms, and mathematical formulas, so what else is required to cross the line into patentability?
Whatever your thoughts on the patentability of software, understanding the rules of the game are critical to protecting your company’s intellectual property, and keeping you safe from infringement liability.
Patentable Subject Matter Through the Ages
The modern contours of software patentability date to the 1970s and the dawn of the personal computer era. When the first software patent issued to Marvin Goetz in 1968, Phil Esposito was in the middle of hockey’s first 100-point season, Seymour Cray had just finished his second supercomputer, and Steve Jobs was in high school.
A few years later, in 1972, the U.S. Supreme Court found that a method of converting numerals from base ten to base two was not patent-eligible because it was directed to an algorithm alone. This appeared to be the end of the “pure” software patent.
After a few twists and turns, the Court decided in 1981 that a computerized patent covering what would ordinarily be patentable subject matter does not does not become unpatentable simply because it uses a computer program or a digital computer. The pendulum swung in the other direction just a bit.
Where a computer is involved, Court said that the invention as a whole was to be examined to determine whether it performs a function that the patent laws were designed to protect (e.g., transforming or reducing an article to a different state or thing). If so, patentable. If not, unpatentable. This became known as the machine-or-transformation test. Having been an IP attorney for over ten years I can confirm that patent attorneys like to talk about the machine-or–transformation test. A lot.
You can’t fudge this by tacking on an insignificant machine or transformation in an attempt to clear the patentability hurdle. Later Supreme Court cases clarified that the machine must impose a meaningful limit on the process and the use of the process has to involve more than “insignificant extra-solution activity.” Printing your results at the end of the calculation isn’t going to cut it.
Fast-forward to 2010 and Bernard Bilski, who invented a computerized method of hedging investment losses in the energy industry. Sounds algorithm-y, right? The USPTO rejected the claim because it was not implemented on a specific apparatus, merely manipulated an abstract idea, and solved a purely mathematical problem. The Supreme Court agreed. Unpatentable.
The Supreme Court went a little further, commenting that while the time-honored machine-or-transformation test was a “useful and important clue or investigative tool,” it is not the only test for determining patentability. This caused an audible gasp in the patent community in 2010 because it opened the patentability door just a bit, but without telling you how to get through that door. Bilski is another thing that patent attorneys like to discuss.
Down The Rabbit Hole
Beginning in 2002, an Australian company called Alice Corporation accused CLS Bank of infringing a handful of patents covering minimizing risk in computerized financial transactions. After several more letters, a district court proceeding, and several years, the contretemps (patent litigation can be a genteel, mannerly pursuit) ended up before a 10-judge panel of the Federal Circuit. (A hearing by more than the typical three-judge panel indicates that the Court recognizes the case to be of exceptional importance.)
The Court agreed that the patented methods were unpatentable, but couldn’t come to consensus on the patented systems.
What was remarkable about the case was that it generated five separate opinions from the panel and showed a division within the Court on how the patentability determination should function as a gatekeeper (is it a coarse filter or a fine mesh) and how it relates to the other hurdles of novelty and non-obviousness that must be overcome.
Seven judges agreed that the method claims were patent-ineligible and invalid. The vote on the system claims was tied 5-5 meaning that the lower court judgment was affirmed and those claims remain valid.
In finding the method claims to be unpatentable, one opinion suggested that many of the claims described an algorithm and tacked on generic computer functionality to lend speed and efficiency. “Abstract methods do not become patent-eligible machines by being clothed in computer language.” (Id. at 37-38)
Given the differing opinions within the Federal Circuit, and the current Supreme Court’s interest in patent law, further review seems likely. But what should you do in the interim?
Now What?
After the Bilski decision in 2010, the Patent Office issued guidelines listing the factors that it would consider when evaluating patent eligibility. These guidelines are still helpful in thinking about whether your invention is patent-eligible, or whether the company behind that infringement notice letter is all hat and no cattle. (A lot of patent cases are filed in Texas so you need to know these aphorisms.)
The law for an application filed today may be different than the law three years from now when that application issues as a patent. For inventors, a patent application filed now should hedge against changes in the law, and be broad enough to support any of the possible approaches that the Supreme Court could take. At the narrower end of the analytic spectrum, the process should be described in the context of the computer to which it is tied. Describe a meaningful connection between the process and the computer. Is the computer a general purpose computer or a dedicated device? How does the computer perform the claimed method steps?
The Supreme Court could decide whether to grant review in CLS Bank as early as next month. The ballet of 35 U.S.C. § 101 continues.