Though most of the attention to yesterday’s decision in Mayo Collaborative Services v. Prometheus Laboratories Inc. has focused on its immediate implications for medical providers, the broader effect of the case probably will be on the software industry. Less than two years ago, in Bilski v. Kappos, the Court affirmed the Federal Circuit’s conclusion that a process for hedging risks of price changes was not patentable. Because that case seemed so easy under the Court’s precedents, the Court’s explanation of why the process was not patentable shed little light on whether the Court would intervene to cut back on the Federal Circuit rules that have permitted the regular (and rapidly burgeoning) issuance of software patents for the last fifteen years. It has been, thus, in the post-Bilski world, business as usual in the Federal Circuit.
But the stern tone of Mayo – with a tone of surprise that the Federal Circuit saw nothing in Bilski to undermine its willingness to affirm the patentability of the diagnostic tool at issue here – suggests that some serious rethinking of the Federal Circuit’s (and PTO’s) treatment of software patents is in order. The problem lies in the analytical frame the unanimous Court uses in Mayo to explain why the patent falls victim to the prohibition on patenting natural laws. The Court repeatedly underscores a high value-added threshold for any patent that includes as part of its operation a natural law: “If a law of nature is not patentable, then neither is a process reciting a law of nature, unless that process has additional features that provide practical assurance that the process is more than a drafting effort designed to monopolize the law of nature itself.” What is necessary, the Court explains, is that the “other steps” with which the law of nature is combined must “transfor[m] the process into an inventive application of the formula” (my emphasis).
Because the Court’s cases regard algorithms as natural laws akin to the metabolic process at issue in Mayo, many software patents have the same conceptual structure as the Mayo patent: a patent on a software program that applies an algorithm often claims the software necessary to make the algorithm practically useful. But if the invention is discovery of the algorithm – a simpler, faster, or more elegant arrangement of steps to manipulate data or make calculations – a patent claiming the software that implements the algorithm is highly dubious after Mayo. To put it another way, if the purpose of the patent is to control use of the newly discovered algorithm, the offhand tone of Mayo suggests the patent is an easy and obvious case for invalidity. But that easy and obvious case surely extends to a large share of the software patents that have been enforced in litigation over the last two decades of the Federal Circuit’s developing tolerance for patents in that area.
Software patents have been controversial for decades, as scholars and policymakers debate whether the incentive for creation they provide is enough to justify the limits they place on competitors. But for better or worse, they play an increasingly important role in allocating competitive control over important innovations. The multi-national litigation over Apple’s “slide to unlock” is only one example. A first reading of Mayo suggests that it casts a long shadow over both the jurisprudence of the Federal Circuit and the PTO’s regular issuance of those patents. History suggests that the PTO will readily take the opportunity to cut back on patents in this area. How quickly the Federal Circuit will get the message is of course a different question.