United States: Processes Effected With Software Are Patentable In The US - What Are The Limits?

From country to country around the world there are wide divergences in legislation and judicial rulings governing the patentability of computer software. In the United States, where the pertinent legislation has remained essentially the same since the late sixties when software began to assume commercial importance, a progression of court decisions handed down between 1972 and the present show an increasing tendency to reward software inventions with patents.

The most recent decision in this progression, State Street Bank and Trust Co. v. Signature Financial Group (see Endnote 1) has been widely discussed, analyzed, criticized, lauded and otherwise scrutinized. Because its claim to fame -- apart from having emanated from the U.S. appellate court charged with hearing all appeals in patent cases, a court answerable only to the Supreme Court of the United States -- is that it abolishes an always somewhat shaky and less than consistently invoked maxim of U.S. patent law under which methods of doing business were deemed per se unpatentable. Since the Supreme Court, which had never passed upon the "business method" rule (a creation of lower level U.S. courts) in the first place, declined to review the ruling, State Street is an authoritative precedent for what it actually holds, at least until such time as its guiding principle is reviewed by the Supreme Court and is reversed. Accordingly, software developers interested in entering the U.S. market for software and entities that provide their funding -- both institutional and individual -- are well advised to cultivate a basic understanding of State Street and certain of its predecessors that underpin the current state of software patentability, in order to reach sound decisions about patenting software in the U.S., including its risks and benefits.

To begin with, State Street does not hold that all business methods embodied in software or all software methods for manipulating numbers are patentable, despite the existence of various reports to the contrary.

State Street, in essence, makes eligible for U.S. patent protection those business methods conducted with software that meet the established U.S. statutory criteria for patentability. Those criteria may be summed up as novelty, unobviousness and practical utility. The novelty criterion is fairly simple (albeit it has complicated ramifications in some contexts) and it requires that an identical process to the one sought to be patented not have been previously known, invented or used by others. The unobviousness criterion is often complex to apply; basically, it requires that the process not be so much like one or more of those previously known that a person of ordinary skill in the business to which it applies would find the programmed method self-evident (or "obvious"). The practical utility requirement is perhaps the most difficult to apply with confidence because of the U.S. history relating to software patentability and the lessons that legal experts have often drawn from the court decisions embodying that history.

In the specific contexts of diverse technologies unrelated to software, the U.S. Supreme Court has consistently ruled for over a century that mere concepts lacking an explication of how to achieve them, abstract ideas, principles of nature, mental steps and scientific or mathematical principles that are building blocks of human knowledge cannot be patented. In its first computer software decision, (see Endnote 2) handed down in 1972, that Court held that a very simple mathematical algorithm embodied in software was not patentable, for the same reasons that the multiplication tables, newly embodied in software, or the principles of gravity or magnetism could not be patented. Many lawyers for a time viewed this decision as mandating that mathematical transformations of any sort could not be patented if embodied in a computer program. Another group of lawyers pointed to a logical anomaly much argued and discussed during the early 1970’s when designing and building special purpose computers was still somewhat in vogue. The point was often advanced that one could build a special purpose computer hard-wired to accomplish the conversion of binary-coded decimal numbers to pure binary numbers (which was the object of the algorithm embodied in the Gottschalk v. Benson software) and could readily obtain a patent on the resulting device, just as engineers of earlier generations had been able to build and patent adding machines (which also did subtraction) and more sophisticated calculators additionally capable of multiplication and division. A question often posed was why the Gottschalk v. Benson software for a general purpose computer that conferred upon it the same functional capabilities that were achievable by hard-wiring to create a special purpose computer was not equally patentable with the special purpose computer. The Supreme Court in Gottschalk v. Benson was confronted with this anomaly, but did not specifically discuss its reasons for holding the software unpatentable even while tacitly conceding the patentability of the hardware.

In 1981, in Diamond v. Diehr, (see Endnote 3) the Court held that when an algorithm implemented in software and its computer execution were made one step of a multi-step physical process for curing synthetic rubber, the overall process for curing rubber could be patented even though other steps in the process were old or of a routine character. In essence, the Court’s reasoning was that Diehr was not seeking to preempt the algorithm for all possible purposes, known and as yet undiscovered, but was merely asking to patent a rubber curing process, an essential element of which was to use an algorithm embodied in software implemented by a general-purpose computer. The Court particularly noted that achieving cured synthetic rubber is a tangible, useful result. The legal commentators assumed, accordingly, that processes implemented by software must produce a physical change in a tangible medium in order to be patentable.

Following Diehr, the United States Supreme Court has not again considered any facet of the patentability of computer programs. The Court of Appeals for the Federal Circuit, which Congress created in 1982 legislation, and which was given exclusive appellate jurisdiction over patent issues arising in federal district courts and federal administrative agencies, has repeatedly considered issues of software patentability during the interim period.

In the real world, two of its decisions have had noteworthy impact -- i.e., In re Alappat, (see Endnote 4) decided in 1994 and that in last year’s State Street case. Alappat is the product of an en banc court, wherein all active judges of the Federal Circuit participated in considering the issues, rather than just a panel of three of them, as occurs in most cases.

In Alappat, a majority of the Court found patentable a program that performed mathematical calculations that converted certain waveforms obtained as data samples into smooth and continuous lines for display on an oscilloscope. The majority of the Court also joined in opining that a general purpose computer becomes a new machine whenever programmed with new software having a different purpose -- a line of thought that has conceptual significance but does not conform to fact.

In State Street, the patent claims are for a general purpose computer programmed with software that enables calculation of individually allocated profits, losses and expenses, on a daily basis, of pooled mutual funds having a large number of owners. Because of the magnitude of financial instruments involved and the magnitude of the individual ownership, the ability to thus calculate and allocate to individual owners the daily profits, losses and expenses is not practically attainable if the operation is not performed by a computer. Nevertheless, the district court whose decision the Federal Circuit reversed, noted that the same functions could be performed by an accountant having pencil, paper, calculator and filing system -- but much less efficiently. This conclusion is theoretically true, but cannot be squared with the real world necessity of daily monitoring, calculation and allocation that United States tax law mandates, in order to permit the owners of the pooled assets to minimize their federal income tax burden. In effect, the Federal Circuit may be regarded as having abolished the somewhat tenuous per se prohibition against the patenting of business methods because the patentee’s software invention availed of technology to achieve a result that could not feasibly be attained otherwise. Traditionally, technical solutions to problems of infeasibility and impracticality have been viewed as necessarily imbued with substantial utility for practical purposes and potentially patentable under U.S. law (i.e., patentable provided novelty and unobviousness were also present), but these solutions heretofore normally have been embodied in the form of a tangible new or greatly improved machine, composition of matter, or article of manufacture. The Federal Circuit’s remand of the patent infringement and remaining validity issues in State Street is consistent with the fact that all it has so far decided is that a clearly very useful method of doing business is potentially patentable.

A great surge of patent application filing on business method inventions has inundated the U.S. Patent and Trademark Office in the wake of the State Street decision, as the Acting Commissioner has publicly acknowledged. Whether that Office is equipped to handle those applications and make sound individual patentability decisions remains to be seen. It has long been recognized that the Office’s access to software technology defining the state of the art is sketchy and far less than complete. Whether it has sufficient personnel who are trained in the technology is uncertain. How efficiently and fairly a business method application can be examined and processed remains to be seen.

Notwithstanding these uncertainties, software entrepreneurs intending to enter the U.S. market and those who invest capital in their endeavors are well advised to weigh the risks and benefits of filing business method applications in the U.S., at least in instances where a previously impracticable result has been rendered feasible of attainment. One of the current advantages such an applicant enjoys only in the United States is that patent applications are held secret while being examined rather than subjected to automatic early publication as occurs throughout most European countries and in Japan. Traditionally, this lack of publication has offered the U.S. applicant who does not file in an early publication country the option of choosing to hold the invention as a trade secret (with the possibility that it will be a long period before others discover or develop the same invention) in lieu of patenting it, after examination for novelty and unobviousness. Of course, if someone else does discover and patent the same invention, the elector of the trade secret option becomes a patent infringer under current law, notwithstanding his prior invention. Nevertheless, the option is foreclosed elsewhere and its availability deserves consideration in the risk/benefit analysis of possible U.S. patent filing.

An already fairly widely voiced criticism of allowing business methods to be patented is that the successful patentee in most instances will enjoy a 20-year exclusive right to exclude others from a whole segment of business, since the patent claims can readily be drafted to prevent others from acquiring competing software patents. Some business gurus are essentially simultaneously suggesting, based on such factors as the availability of LINUX, a freely available computer language that is not copyrighted or otherwise monopolized, that the world is about to enter upon a new era of business cooperation in which, for example, patents will be widely licensed to competitors rather than relied upon to assure exclusivity of their enjoyment by the patentee. Whether or not this occurs, it, too, deserves consideration.

Finally, no one can assure that any software-based patent relating to a business method will survive a post-issuance validity attack, launched either in the form of a patent reexamination request or in response to a charge of infringement. While this is also true of many patents in other fields, the uncertainties in this area of endeavor are greater than in patents on machines, compositions of matter and articles of manufacture. But in deciding whether to seek a software patent of significant potential impact, one must also weigh the high stakes sometimes available to reward the successful patentee.

1. 49 F.3d 1368 (Fed. Cir. 1998), cert. denied, 142 L.Ed.2d 704 (1999).
2. Gottschalk v. Benson, 409 U.S. 63 (1972).
3. 450 U.S. 175 (1981).
4. 33 F.3d 1526 (Fed. Cir.1994).

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.