The Patentability of Software: Myths, Facts and a Proposed Test

For more than a decade, the legal, technical and academic communities have debated the boundaries, legality and wisdom surrounding the issue of software patentability. The debate, to say the least, has been spirited with many organized movements, websites, articles, blogs and law review articles, as well as a lot of lobbying dollars, devoted to both sides of the debate. Yet, as the U.S. Supreme Court currently reviews the U.S. Court of Appeals for the Federal Circuit’s en banc May 10, 2013, decision in CLS Bank v. Alice, we take the opportunity here to dispel some myths, state some of facts and offer a test with respect to software patentability.

This is not your Parent’s Software Industry

There are those who argue against the patentability of software as a whole – never mind finding a test as to what software (or computer-implemented) claims should be patentable. We disagree. Why? Well, the notion that software should not be patentable necessarily indicates that the software industry itself is not capable of innovation worthy of patent protection! Yet, in a country where patent rights are guaranteed by the Constitution, should not all fields of innovation be treated equally under the law? Should we not avoid becoming a country where one field of endeavor (e.g., pharmaceuticals or electronics) is deemed more “patent worthy” than other fields (i.e., computer science and information technology)!? To answer these questions in the negative seems silly to us.

According to the U.S. Dept. of Commerce, the U.S. software and IT services industry had revenue totaling $606B in 2011, with overall research and development spending of $126.3B, and a U.S. workforce of nearly two million people. Further, a PriceWaterhouseCoopers report pinned the cumulative value of technology-related M&A activity for 2013 at $99.8B, with software representing 25% of this total value and 35% of the total deal volume. This is a substantial amount of commerce to exclude from the patent system!

In 2012, according to a report by the Computing Research Organization (CRA), there we over 1,620 PhD graduates, along with over 6,973 Master’s degrees and 13,055 Bachelor of Science degrees awarded in Computer Science. These numbers represent a double-digit percentage increase for the third straight year in bachelor’s degree production and the highest number of doctorate degrees ever awarded in a single year (an 8.2 percent increase from 2011). Are these graduates’ innovations less “patent worthy” than the 8,440 PhD (2012), 15,352 master’s degrees (2012) and 110,015 bachelor’s degrees (2010) recipients in the biological/agricultural sciences? Of course not! Why else would a 2013 report by PayScale show that the Computer Science college major ranks number eight for salary potential, with a median $59,800 stating salary and a median $99,300 mid-career salary!? In fact, another 2012 study shows that “computing careers will make up 73% of the new jobs in STEM careers this decade compared to 16% in (non-software) engineering, 9% in the natural sciences, and 2% in the mathematical sciences.” These numbers clearly show the value of a highly-skilled workforce to an important segment of the U.S. economy.

Those who argue against the patentability of software also seem to think that anyone (e.g., a teenager) can quickly program a computer to do just about anything, and that the level of skill in the art is low. Indicative of this misconception is the following exchange during the U.S. Supreme Court Alice v. CLS Bank oral argument on March 31, 2014:

JUSTICE KENNEDY: Suppose I thought … that any computer group of people sitting around a coffee shop in Silicon Valley could do this over a weekend. Suppose I thought that.

MR. PHILLIPS: You mean wrote the code?


MR. PHILLIPS: Well, that’s absolutely ­­ I’m certain that’s true. … But that’s true of almost all software.

What may surprise you is that Mr. Phillips was counsel for the patentee and arguing for software patentability!

These misconceptions are furthered by media stories about teenagers writing successful apps. For example, in March of 2014, it was reported that Nick D’Alosio – a 17 year-old from the United Kingdom – sold his mobile app to Yahoo! for $30M. The truth of the matter is that statistics show over 3 million U.S. high school students drop out each year, and international test results of 15 year-olds from 65 countries in 2012 show that U.S. students ranked below average in math among the world’s most-developed countries. An overwhelming majority of teenagers simply cannot be like Nick. He is a prodigy, and 98% of the U.S. population does not qualify as such.

What is Patentable?

What we do know for sure is that U.S. Patent Law recognizes four broad categories of inventions eligible for patent protection: processes; machines; article of manufacture; and compositions of matter. 35 U.S.C. Section 101. We also know for sure, despite the oft-quoted recognition that the patent laws were made to cover “anything under the sun that is made by man,” Diamond v. Chakrabarty, 447 U.S. 303, 309 (1980) (quoting S. Rep. No. 1979, 82d Cong. 2d. Sess., 5 (1952)), the U.S. Supreme Court has long recognized that there are three exceptions to these four broad patent-eligibility categories: laws of nature; physical phenomena; and abstract ideas. Id. This is where the certainty ends.

The Software Patentability Legal Voyage

In State Street Bank & Trust Co. v. Signature Fin. Group, 149 F.3d 1368 (Fed. Cir. 1998), while affirming that pure mathematical algorithms remain per se unpatentable, a three-judge panel of the U.S. Court of Appeals for the Federal Circuit (CAFC) held that any method that can be described as a whole in a manner in which it produces “a useful, concrete and tangible [i.e., ‘real-world’] result,” is eligible for patent protection. Id. at 1373. The State Street court, considered a clam directed to “a data processing system for managing a financial services configuration of a portfolio established as a partnership.” Id. at 1372. The court noted that, for the purposes of a 35 U.S.C. Section 101 patent-eligibility analysis, “it is of little relevance whether claim 1 is directed to a ‘machine’ or a ‘process,’ as long as it falls within at least one of the four enumerated categories of patentable subject matter [specified in 35 U.S.C. Section 101], ‘machine’ and ‘process’ being such categories.” Id.

Twelve years after State Street, the Supreme Court tackled the issue in Bilski v. Kappos, 130 S.Ct. 3218 (2010). The U.S. Patent and Trademark Office (USPTO) and the CAFC both rejected the Bilski application as being directed to a non-patentable process under Section 101 using the “machine-or-transformation” test. Under this test, a claimed process is patent-eligible under Section 101 only if: “(1) it is tied to a particular machine or apparatus; or (2) it transforms a particular article into a different state or thing.” The Supreme Court, however, unanimously ruled that: “[T]he machine-or-transformation test is a useful and important clue, an investigative tool, for determining whether some claimed inventions are processes under [Section] 101. The machine-or-transformation test is not the sole test for deciding whether an invention is a patent-eligible ‘process.’” 130 S.Ct. at 3227. The Supreme Court reasoned that Section 101 is a “dynamic provision designed to encompass new and unforeseen inventions.” Id. at 3239. “[T]he machine-or-transformation test would create uncertainty as to the patentability of software, advanced diagnostic medicine techniques, and inventions based on linear programming, data compression, and the manipulation of digital signals. … As a result, in deciding whether previously unforeseen inventions qualify as patentable ‘process[es],’ it may not make sense to require courts to confine themselves to asking the questions posed by the machine-or-transformation test.” Id. at 3228. Thus, the Supreme Court encouraged the CAFC to develop “other limiting criteria that further the purposes of the Patent Act and are not inconsistent with its text.” Id. at 3231.

On July 9, 2012, a divided panel of the CAFC attempting to apply the U.S. Supreme Court’s decision in Bilski decided that a computerized trading platform for exchanging obligations so as to eliminate “settlement risk,” defined patent-eligible subject matter under 35 U.S.C. Section 101. CLS Bank Int’l v. Alice Corp. Pty. Ltd., 685 F.3d 1341 (Fed. Cir. 2012). The panel in CLS Alice defined its task as determining “whether the claimed inventions fall within the ‘abstract ideas’ exception to patent eligibility,” Id. at 1348. Applying Bilski, the panel reasoned “that when—after taking all of the claim recitations into consideration—it is not manifestly evident that a claim is directed to a patent ineligible abstract idea, that claim must not be deemed for that reason to be inadequate under § 101.” Id. at 1352. Applying this new “manifestly evident” test, the panel concluded that: “it is difficult to conclude that the computer limitations here do not play a significant part in the performance of the invention or that the claims are not limited to a very specific application of the concept of using an intermediary to help consummate exchanges between parties,” id. at 1355, and were thus patent-eligible matter.

Two weeks after CLS Bank was decided, on July 26, 2012, a different panel of the CAFC unanimously decided that a computerized system and method for administering and tracking the value of life insurance policies in separate accounts did not define patent-eligible subject matter under 35 U.S.C. Section 101. Bancorp Services, L.L.C. v. Sun Life Assur. Co. of Canada, 2012 WL 3037176 (CAFC 2012). Like CLS Bank, the issue facing the panel was whether the claim was directed to an abstract idea. Although the panel did not employ the new “manifestly evident” test of CLS Bank, in finding the claims not patent eligible, it stated that “our conclusion is not inconsistent with CLS,” because: “[T]he computer limitations do not play a ‘significant part’ in the performance of the claimed invention. And unlike in CLS, the claims here are not directed to a ‘very specific application’ of the inventive concept; as noted, Bancorp seeks to broadly claim the unpatentable abstract concept of managing a stable value protected life insurance policy.” 2012 WL 3037176 at *12. The Bancorp panel further reasoned that: “Rather, the claims merely employ computers to track, reconcile, and administer a life insurance policy with a stable value component—i.e., the computer simply performs more efficiently what could otherwise be accomplished manually.” Id. at *11.

The State Street, Bilski, CLS Bank and Bancorp journey of decisions left many wanting for more. Why? Because a clear, repeatable test to determine when exactly a software (or computer-implemented) claim is patentable versus being simply an abstract idea “free to all men and reserved exclusively to none,” Funk Brothers Seed Co. v. Kalo Inoculant Co., 333 U.S. 127, 130 (1948), has yet to be articulated by the courts. This is why soon-to-be, former Chief Judge Rader has called the CAFC en banc decision in CLS Bank v. Alice Corp. “the greatest failure of my judicial career.” The fragmented, 125-page decision failed to provide the patent bar adequate guidance on the patentability of software.

As one of us has stated before, how can patent practitioners give “useful, concrete and tangible” advice to their clients about “manifestly evident” inventions that may relate to “machine[s]-or-transformation[s]?” Should such practitioners resign themselves to admitting that: “[P]erhaps [they] could never succeed in intelligibly doing so. But [they] know it when [they] see it”? Jacobellis v. Ohio, 378 U.S. 184, 197 (1964) (Stewart, J. concurring) (describing his threshold test for obscenity). Such a resignation is not ideal as uncertainty has never been the aim of the law. Yet, today, we find ourselves with a myriad of USPTO Section 101 guidelines, flowcharts and presentation slides – the latest of which is a March 4, 2014, 19-pager. Given the importance of intellectual property to the U.S. economy, we should aim to do better than that. Otherwise, as previously noted by IBM’s Chief Patent Counsel, Manny Schecter, patent law and practice could become as convoluted and voluminous as the tax code.

A Good Software Patent Application

Before setting forth our proposed test for software patentability, it is instructive to lay out what we believe constitutes a good software patent disclosure against which such a test would be applied. We feel that the specification should contain text to explain the following minimum set of figures:

  1. A system diagram explaining the overall environment in which the claimed software executes.
  2. At least one, top-level flowchart which explains the functionality of the claimed software, wherein: (a) each block of the flowchart should indicate which component of the system (in Figure 1) performs that step; and (b) the flowchart should contain at least one decisional block showing the “intelligence” included in the claimed software.
    • The disclosure should also include one or more second-level flowcharts set forth the functional steps of each block of the top-level flowchart where the inventive features lie.
  1. At least one graphical user interface (GUI) or user experience (UX) “screenshot” showing the claimed software in operation.
  2. A block diagram of a basic computer to support computer program product (i.e., In re Beauregard) type claims.

A Proposed Test

A discussion of software-related patent claim form, complexity and context only serves to obscure the real question – is the invention in question worthy of patent protection? Put another way, the issue (both in the U.S. and abroad) is really one of identifying the “intelligence” behind a software- or computer-implemented invention. By identifying that intelligence, one can determine whether the invention is merely abstract or covers a human-driven process – either not being patentable eligible – or involves some intelligent processing driven by the software that provides a new output – thus being eligible for patent protection. Phrased another way, out proposed test is:

Looking at the claim as a whole, does the potentially novel and non-obvious intelligence reside in the software (or the computer implementation) that is producing a tangible and concrete result, or does it simply reside with a human interacting with the software (or computer implementation)?

Yes, our test returns, in part, to the State Street test!

Examining “where’s the intelligence?” behind an invention should put to rest concerns over claiming an abstract idea, as well as the often-criticized token addition of the phrase “computer-implemented” in method claims or “a processor” to system claims. Further, an invention involving a computer or process to provide intelligent processing of information should elevate that claimed method and/or system above one that merely couches an old idea in the context of a computer.

Using the above test, easy cases would be:

Such an analysis can perhaps be further illustrated by looking at an example flow diagram that might be found in a patent application’s specification. In looking at the flow diagram, a straight line flow without any decision points seems to indicate a straightforward process flow in which the process itself is not relied upon for any decision-making (e.g., the automation of a previously manual process).

In considering a flow diagram with a decision point or a branch, however, this suggests that some decision has to be made that affects the outcome of the process. That decision point can be evaluated to determine whether the decision is: (i) a mental one, made by a human; or (ii) a decision made by the software or processor itself.

If the decision point is evaluated and a next path or branch in the claimed process flow is determined by the software process and/or system, this would seem to be strong evidence that the “intelligence” is in the claimed invention and therefore strong evidence of patent eligibility. The analysis should then continue to a 102/103 prior art analysis!

For example, let’s apply our “intelligence” test to the independent claim at issue in Bilski:

A method for managing the consumption risk costs of a commodity sold by a commodity provider at a fixed price comprising the steps of:

(a) initiating a series of transactions between said commodity provider and consumers of said commodity wherein said consumers purchase said commodity at a fixed rate based upon historical averages, said fixed rate corresponding to a risk position of said consumers;

(b) identifying market participants for said commodity having a counter-risk position to said consumers; and

(c) initiating a series of transactions between said commodity provider and said market participants at a second fixed rate such that said series of market participant transactions balances the risk position of said series of consumer transactions.

This method claims a series of sequential steps that (a) initiate, (b) identify and (c) initiate. Each initiation occurs without a test, comparison, or other decision point. Further, as made clear in the Bilski litigation, the claimed method did not necessarily involve a computer or processor. Thus, this claim fails to provide any software- or computer-implementation with sufficient “intelligence” to pass muster under our subject matter eligibility test. In contrast, the patent claims at issue in Diamond v. Diehr, 450 U.S. 175 (1981), involved several decision points around press status, a comparison of closure time to cure time and temperature status, which were driven by a processor; and the patent claims at issue in State Street, which involved computer-driven processing of data leading to a financial allocation and adjustment; would both pass our “intelligence” test consistent with the judicial outcomes in each case.


While no legal test is ever perfect, we believe the framework we propose above is workable, can at least provide stability to the software and computer-implemented invention field of technological arts and is consistent with over 30 years of U.S. Supreme Court jurisprudence. Further, our proposed test works well in light of the US Supreme Court’s April 29, 2014, unanimous decision in Octane Fitness, LLC v. Icon Health & Fitness, Inc. (making it easier for district courts to sanction plaintiffs for bringing meritless patent infringement suits). In this new and still emerging “loser pays” regime, software patent holders may do well to engage in some self-policing by applying our test, thus allowing the patent ecosystem to work itself out.


The Author

Raymond Millien & Christopher George

Raymond Millien & Christopher George

Warning & Disclaimer: The pages, articles and comments on do not constitute legal advice, nor do they create any attorney-client relationship. The articles published express the personal opinion and views of the author as of the time of publication and should not be attributed to the author’s employer, clients or the sponsors of Read more.

Discuss this

There are currently 5 Comments comments.

  1. Anon May 28, 2014 12:50 pm

    I have seen a simple (perhaps even more simple) solution:

    Take the claim as one who would have ordinary skill in the art to which the invention pertains would take the claim.

    Take judicial notice that infringement of such a claim cannot apply to any infringement that would occur totally in the mind of the alleged infringer.

    Any other infringement that reads on the claim – that involves more than something totally in the mind falls outside of the judicial notice.

  2. AP May 29, 2014 8:08 am

    You could rewrite step (b) in Bilski as a branch. Instead of “identifying market participants having…” you could write “does the market participant have…,” yes/no.

  3. Benny May 29, 2014 8:08 am

    Your test case isn’t as simple as it looks.
    Virtually all software involves a decision making process – in the example you gave above as non patentable, a real world computer would decide when to update the data and what fluctations to filter out. My personal opinion is that such a decision making process should be patentable if it is novel, but not patentable if an identical process was previously achieved by other, automated devices – a hardware implemented state machine, for example. Correct me if I’m wrong, but the innovation is not in being able to string together the code but in the decision process carried out – the algorithm, not the program.

  4. Gene May 29, 2014 11:43 am

    i’ll say it over and over again . . . what all software inventors invent is algorithms; algorithms, when compiled, construct a new machine made of transistors. disclose the algorithm and you disclose the invention (the machine). the algorithm can be analyzed under 112, 102 and 103. it automatically passes 101 because different algorithms that perform the same function are compiled into different machines. we can all then understand subsequent disputes over patentably distinct, or not, features of the machine. CAFC cases already require algorithms to be disclosed for M+F claims and method claims, else they are invalid (for lack of corresponding structure and abstractness, respectively). just make it an across-the-board rule to disclose the algorithm for enablement and adequate teaching.

  5. Fish sticks May 29, 2014 2:20 pm

    What steam was to the 18th century, electricity to the 19th, and hydrocarbons to the 20th, DATA will be to the 21st century,” IBM President Ginni Rometty recently said.
    Manipulations of steam, electricity and hydrocarbons are clearly patent eligible subject matter — and manipulations of data should be too IMHO.