When should software and business methods be patentable, if ever? Two courts are currently grappling with this very question. In the United States, the Supreme Court is expected to deliver its decision in Bilski v. Kappos within weeks. In Canada, the Federal Court recently heard oral arguments in Amazon.com, Inc. v. The Attorney General of Canada et al, the Amazon 1-Click appeal. Both cases are likely to shape the patent landscape for years to come.
Previous decisions in Bilski and Amazon both conspicuously broke with established patentability requirements and led to the current appeals. Each discarded earlier subject matter tests that looked primarily to the result of a method to determine whether the method was patentable. The mere presence of economically useful activity is no longer sufficient. Rather, the means used to accomplish the method are now a relevant consideration.
In the U.S., the means must be a "machine" or "apparatus" under the newly-developed machine-or-transformation test. Jurisprudence there has established that a general purpose computer, when programmed to perform a particular task, is indeed considered a "particular" machine. This semantic trick belies a deep insight into the nature of modern industry.
Fewer and fewer modern machines can be found today without at least some software in them. This is no surprise, as building in flexibility through software is incredibly useful. For example, modifications to a machine might have previously required custom machining of parts, or redesign of a circuit board. Today, many of the same changes can be made quickly and easily with modification to software code. Rather than recall thousands of defective devices, manufacturers can effect repairs with a simple patch.
Thus, the trick that allows a computer to be considered a particular machine recognizes that software is, in many cases, a direct and superior replacement for the physical machines that used to power industry. As such, it should be entitled to the same protections.
However, there are complaints that a strict machine-or-transformation test is too rigid. Others argue that no software should be patentable at all. It is difficult to craft a definition of patentable subject matter that accommodates the many useful applications of software while prohibiting mere abstract ideas or concepts.
The Amazon decision in Canada attempts to address this difficulty by adopting a more ambiguous "technological solution to a problem" definition, while also renewing the "change of character or condition of a physical object" test from the well-known Lawson case.
Reading the tea leaves
Based on transcripts of oral arguments, the U.S. Supreme Court appears genuinely determined to identify a middle ground between rigid application of a machine-or-transformation test and the previous "anything under the sun" approach. For instance, Justice Sotomayor asked how the Court could limit patentable subject matter to "something that is reasonable" if not to "technology" specifically. Several of the justices were critical of the Federal Circuit's machine-or-transformation test and it is likely that it will be reformulated, if not replaced outright.
If the Court does adopt the middle ground, software patents should emerge safely and perhaps with a better-defined test for patentability.
In Amazon, argument appears focused on a definition for "art" (method). Amazon prefers the much more expansive definition cited in the Supreme Court's Shell Oil decision: "new and innovative methods of applying skill or knowledge provided they produced effects or results commercially useful to the public."
For its part, the Crown argues that the Shell Oil definition was limited only to a sub-class of patent methods. Therefore, the Lawson "change of character or condition" formulation ought to apply in other cases.
The next few months will tell if the Federal Court adopts an approach similar to that predicted in the U.S.