On October 30, 2008, the U.S. Court of Appeals for the Federal Circuit handed down a decision that purports to clarify the standard for determining whether a "process" is patentable. Section 101 of the Patent Act provides that whoever invents any new and useful "process, machine, manufacture or composition of matter" may obtain a patent therefor, assuming all other requirements for patentability (such as the requirement that the patent application fully teach the invention) are met. The words "machine, manufacture and composition of matter" have been included in the Patent Act since it was first enacted in 1793. The word "process," however, was substituted for the word "art" in 1952. Thus, the question before the court in In re Bilski, 545 F.3d 943 (Fed. Cir. 2008)--and the question with respect to the patentability of any so-called "business method patent"--is: What is meant by the word "process"?
The court in Bilski held that the word "process" means something that is (1) tied to a particular machine or apparatus or (ii) transforms a particular article into a different state or thing. The court referred to this test as the "machine-or-transformation test" and argued that it is consistent with U.S. Supreme Court precedent. The Federal Circuit acknowledged, however, that future developments in science and technology may "present difficult challenges" and that the U.S. Supreme Court may ultimately decide to come up with a new test. This author had the sense in reading the Bilski opinion that the Federal Circuit was doing its best to find consistency in past U.S. Supreme Court opinions where such consistency may not exist and that the court itself did not necessarily believe that the test it was espousing was one that would stand the test of time.
In this author's view, the "machine-or-transformation test" will prove extremely difficult to implement. With one exception (noted below), the court did not answer the question as to whether, and under what circumstances, data or an electronic signal is transformed into a "different state or thing." Similarly, the court expressly declined to answer the question whether or when including reference to a "computer" in a claim will satisfy the requirement of tying the invention to a machine or apparatus. Thus, in this author’s view, inventors seeking to protect their software or business method inventions have no further guidance today than they did prior to the Bilski decision.
As stated by Judge Newman in his dissenting opinion, "These new uncertainties not only diminish the incentives available to new enterprise, but disrupt the settled expectations of those who relied on the law as it existed." Judge Newman argued that the majority overstepped its bounds in imposing a limitation on patentability that is not reflected in the statute. According to Judge Newman, "The breadth of Section 101...reflects the legislative intention to accommodate not only known fields of creativity, but also the unknown future." In Judge Newman's view, the machine-or-transformation test will exclude forms of information-based and software-implemented inventions arising from new technological capabilities that may not involve a machine or even a transformation of matter.
This author acknowledges and agrees that some form of "check" on the patentability of business methods is appropriate; however, the test articulated by the court in Bilski goes far beyond dealing with problematic business methods and threatens to affect the patentability of all computer-implemented inventions. For now, patent practitioners would be well advised to include both system and method claims in software patent applications and, if the software involves the use of one or more algorithms (as most software does), ensure--to the extent they can--that the claims do not preempt all uses of the algorithm(s). According to the court, a claim that attempts to preempt all uses of an algorithm is, in essence, an attempt to patent the algorithm itself.
The court's attempt at providing some level of guidance to patent practitioners and would-be software patent holders fell far short of the mark. According to the Bilski court, an algorithm is a "fundamental principle," and fundamental principles are not patentable. A specific application of a fundamental principle (or algorithm) is patentable, however. That sounds straightforward enough, but then the court went on to say that if a particular algorithm has no utility other than operating on a computer, then claiming the algorithm in connection with a computer will not be enough. With respect to the "transformation" requirement, the court noted "for clarity" that electronic transformation of data into a visual depiction is sufficient to meet the transformation requirement. It remains to be seen how the patent office and courts will interpret and apply the machine-or-transformation test.
