This week’s intellectual property news is dominated by the Supreme Court oral argument in Alice Corp. v. CLS Bank, over the patent eligibility under 35 U.S.C. 101 of an invention in the field of financial transactions. The Supreme Court’s decision in this case will (or, at least, might) address several key questions of wide-ranging importance, including the definition of the term “abstract,” in the context of the utility requirement of patent eligibility.
The topic of this post is about the attitudes about software, as a technology, that appeared during the Alice Corp. v. CLS Bank oral argument, and that pervade the field of patent law.
The following exchange occurred between Justice Kagan and Mark Perry, the counsel for CLS Bank who argued that the invention described in U.S. Patent No. 6,912,510 (the ‘510 patent) is not eligible for patenting:
KAGAN: Mr. Perry, you said to Justice Scalia that if a patent sufficiently describes how a computer will implement an idea, then it’s patentable. So how sufficiently does one have to describe it?
PERRY: What the applicant or patentee must not do is simply describe the desired result. That would simply say: “I claim a magic box that buys high and sells low or vice versa; I claim a magic box for investing.” That’s what these patents do.
KAGAN: You’re not suggesting that specific code is necessary?
PERRY: No, Your Honor. The actual description of the programming is a (35 U.S.C.) 112 problem – the realm of the written description requirement. What is a 101 problem is: it is on the applicant to do more than simply describe the results, simply say: “A magic box that does intermediate settlement.” And we can tie that back to the ‘510 patent.
However, the claims of the ‘510 patent do not state: “Do intermediate settlement.” Here is claim 1:
A method of exchanging an obligation between parties, wherein an exchange obligation is administered by a supervisory institution, and wherein at least one credit record and one debit record is maintained with an exchange institution, the method comprising:
(a) maintaining a shadow credit record and a shadow debit record for a party to be held independently by the supervisory institution from the exchange institution;
(b) for every transaction resulting in an exchange obligation, the supervisory institution electronically adjusting said shadow credit record and/or shadow debit record, allowing only those transactions that do not result in a value of said shadow debit record being less than a value of said shadow credit record; and
(c) at the end of a period of time, the supervisory institution providing an instruction to the exchange institution to credit and/or debit in accordance with said adjustments of said allowed transactions, wherein said instruction being an irrevocable, time invariant obligation placed on the exchange institution.
That’s the broadest and most general claim. The ‘510 patent has 74 other claims that provide additional detail about variations and other embodiments of this process. The specification is even more detailed: the figures show a flowchart spanning Figures 8 through 16, providing an extremely detailed logical process expressing a sophisticated technique.
And yet, Perry summarizes this invention as: “a magic box that does intermediate settlement.”
Perry’s argument reflects an attitude of dismissiveness – or even disrespect – of the details of the invention. It asserts that the patent uses a lot of words to describe a comparatively simple idea – and that the simple idea, when generalized away from the specifically claimed details, is too simple, broad, and “abstract” for patent eligibility.
This dismissive attitude is common in discussions of patent law, particularly when the invention involves software:
- In the majority opinion of Bilski v. Kappos, the claimed invention is held ineligible with the following understanding of the claims:
Petitioners seek to patent both the concept of hedging risk, and the application of that concept to energy markets.
However, a review of the published portions of Bilski’s application reflect much more detail than simply “the concept of hedging risk” – very specific, highly detailed mathematical models and formulae are included.
It may have been legitimate to challenge this technique on grounds of novelty, or even rigorously scrutinizing the claim to determine whether it only recites a “law of nature.” The Supreme Court did neither: it overgeneralized the specifically disclosed and claimed techniques as “the concept of hedging risk,” and refused to consider the merit of the claimed invention in its full and precisely described detail.
- The CLS Bank v. Alice Corp. per curiam opinion of the Court of Appeals for the Federal Circuit critically turned on what the claimed invention “basically” provides:
Claim 33 plainly recites a process. The methods claimed here draw on the abstract idea of reducing settlement risk by effective trades through a third-party intermediary empowered to verify that both parties can fulfill their obligations before allowing the exchange – i.e., a form of escrow. The concept of reducing settlement risk by facilitating a trade through third-party intermediation is an abstract idea because it is a “disembodied” concept. 1
- In 2011 and 2013, National Public Radio ran two episodes of a radio series called When Patents Attack, which discussed a patent that covers “a lot of the most common things we do on the Internet.” The segments provided:
David Martin runs a company called M-Cam. It’s hired by governments, banks and business to assess patent quality, which the company does with a fancy piece of software. We asked Martin to assess Chris Crawford’s patent.
Crawford’s patent was for “an online backup system.” Another patent from the same time was for “efficiently backing up files using multiple computer systems.” Yet another was for “mirroring data in a remote data storage system.”
At the same time Crawford’s patent was being prosecuted, more than 5,000 other patents were issued for “the same thing,” Martin says.
This conclusion asserts this methodology of looking at the patent and saying, “it’s ‘basically’ this simple idea.” Martin’s conclusion is based on a comparison of patent titles, which are meaningless to the scope of the patent.
Amazon.com has 6,500 books with the word “Pirate” in the title – no one would presume that all 6,500 books tell the same story. Similarly, 3,700 patents feature the word “refrigerator” in the title – but no one (except, presumably, David Martin) would claim that all 3,700 patents claim “the same thing,” i.e., “basically just the concept of a refrigerator.” 2
- The Electronic Frontier Foundation runs a “Bust Patents” program that asks the public to find prior art for pending patents – which is a perfectly legitimate and valuable pursuit. However, to this end, the group routinely overgeneralizes software patents for dramatic effect.
For example, the EFF references U.S. Patent No. 6,026,368, which claims:
1. A system comprising:
a queue builder for generating priority queues;
a computer mediated communications network for sending content data and subscriber data to the queue builder;
an on-line queue manager for receiving the priority queues from the queue builder and for sending content segment play lists to the computer mediated communications network; and
an exposure accounting module for receiving exposure data from the on-line queue manager and computing and outputting exposure billing data,
wherein the queue builder is provided with profile editors through which rule developers create definitions by building data base selection criteria for extracting records with a common profile,
wherein the queue builder accesses a disk storage that stores targeting data and rules for generating the priority queues, and
wherein the disk storage is also accessed by the exposure accounting module to adjust priorities in order to meet exposure goals.
…which the EFF describes as a patent for “targeted banner ads.”
1. A digital information storage, retrieval and display system comprising:
a central computer means in which plural blocks of information are stored at respectively corresponding locations, each of which locations is designated by a predetermined address therein by means of which a block can be selected, each of said blocks comprising a first portion containing information for display and a second portion containing information not for display but including the complete address for each of plural other blocks of information;
plural remote terminal means, each including (a) modem means for effecting input/output digital data communication with said central computer means via the telephone lines of a telephone network, (b) local memory for locally storing digital data representing at least the first portion of the selected block of information received via said modem means from the central computer, (c) display means for visually displaying such a locally stored first portion of a block of information and (d) key pad means connected to communicate data to at least said modem means for manual entry of keyed digital data; and
further memory means being provided as a part of said central computer means for receiving and storing said second portion of the block of information selected by a particular terminal means in response to the selection of the block and when its respective first portion is transmitted to that terminal means for display, said central computer means utilizing keyed digital data from that particular terminal means of less extent than any one of said complete addresses for another block of information but nevertheless uniquely indicative of one of the complete addresses contained in said portion of the block of information which contains the first portion then being displayed by that particular terminal means for selectively accessing the part of said further memory means associated with that particular terminal means and for supplying the complete address of the next block of information which is to be retrieved for that particular terminal means and utilized for display purposes at that terminal means.
…which the EFF describes as a patent for “the hyperlink.”
This type of dishonesty on the part of the EFF undercuts the purpose of the program and the credibility of the organization.
- The software development community is both hostile to software patents, and misinformed about patents in general, with an abundance of misunderstanding in the same manner exhibited by National Public Radio and the Electronic Frontier Foundation.
However, this combination of traits causes an “echo chamber” effect: reports of newly issued software patents are portrayed in a factually incorrect manner that reinforces hostility about software patents, which, in turn, enables readers to disregard responses to such reports that provide correct information.
- U.S. Patent No. 7,415,666: It’s basically PageUp/PageDown.
- U.S. Patent No. 7,650,331: It’s basically MapReduce.
- U.S. Patent No. 7,669,123: It’s basically the Facebook news feed.
- U.S. Patent No. 8,139,822: It’s basically motion analysis.
- U.S. Patent No. 8,223,134: It’s basically scrolling.
- U.S. Patent No. 8,615,473: It’s basically anticipatory shipping.
- U.S. Patent No. 8,676,910: It’s basically Twitter.
- U.S. Patent No. 8,689,141: It’s basically geo-targeting.
- U.S. Patent No. 8,223,134: It’s basically scrolling.
Every single one of these summaries is demonstrably false: a glance at the claims reveals additional detail and limitations reciting a specific technique. But the software community is more inclined to cling to a skewed perception of software patents than a more nuanced but accurate one.
- U.S. Patent No. 7,415,666: It’s basically PageUp/PageDown.
- This attitude is also common among USPTO examiners in software fields.
I have encountered many examiners in the software arts who leave the impression that they do not care about the details of software technology. When presented with a portion of a claim that provides a specific technique, these examiners respond: “this is basically just the concept of encryption,” or “this is basically just a user interface,” or “this is basically just a data structure.” These examiners do not acknowledge that the claims recite a specific technique, or how it differs from their references, or why the difference matters. They do not understand or care about these details, and feel empowered to ignore critical portions of patent claims – because, well, it’s “just” software.
The disregard of software inventions by the patent community is noteworthy for at least three reasons.
First: The central feature of modern patent law – the patent claim – is intended to prohibit overgeneralizing and dismissiveness.
Before the 1952 Patent Act (which provides the basis for our current patent law), the U.S. patent system was based on a “central” claiming system – a patent application only recited a basic invention, and the courts determined how much breadth to give it. However, this earlier system proved unsustainable: the decisions of examiners and judges were subjective, arbitrary, unpredictable, and inconsistent. In response, the 1952 Patent Act implemented a “peripheral” claiming system: applicants are required to present a set of claims, each reciting the “metes and bounds” of the invention. The breadth of the patent is determined by the amount of detail that the applicant included in the claim. This system provides more objectivity: either every word in the claim describes an accused infringer’s product, and thus infringes the patent; or it doesn’t.
The role of the patent claim is reflected in Diamond v. Diehr. In deciding the patent eligibility of a method of controlling a boiler using a computer prograamined with the (admittedly centuries-old) Arrhenius equation, the Supreme Court held:
Respondents’ claims must be considered as a whole, it being inappropriate to dissect the claims into old and new elements and then to ignore the presence of the old elements in the analysis. The questions of whether a particular invention meets the “novelty” requirements of 35 U.S.C. 102 or the “nonobviousness” requirements of 103 do not affect the determination of whether the invention falls into a category of subject matter that is eligible for patent protection under 101.
Diehr is routinely cited with continued approval in every major patent law case (up to and including Bilski v. Kappos). However, recent case law often turns on extracting the “essence” of the claim – the core idea of the invention – from the immaterial details surrounding it, and deciding whether the essence, considered by itself, is patent-eligible – directly contrary to the central holding of Diehr.
Second: Decisions that many software inventions are “abstract” appear unreasonable when considering that “abstract” is used to mean “not useful,” as required by 35 USC 101.
The decades-long argument over the term “abstract” originated in the 1972 Supreme Court case of Gottschalk v. Benson. That case involved quintessentially “abstract” claims: the conversion of numeric data from binary-coded decimal to binary, without any suggested motivation for doing so. The Supreme Court held that these claims were “abstract” because, to the Court, the conversion appeared to be a bare mathematical formula without any recited or apparent utility.
By contrast, the inventions claimed by Alice and Bilski, among others, have clearly recited utility stated plainly in the specification. From Bilski’s application:
The invention relates to method for managing the consumption risk costs of a commodity sold at a fixed price and, more particularly, managing the weather-related risks associated with energy pricing.
From Alice’s patent:
The invention encompasses methods and apparatus enabling the management of risk relating to specified, yet unknown, future events by constructing compensatiory claim contract orders. … The invention enables parties to manage perceived risk in respect to known, yet non-predictable possible future events.
Moreover, in both cases, the utility was recited in the claim preambles (“A method for managing the consumption risk costs of a commodity,” and “A method of exchanging an obligation between parties”). Both claimed techniques certainly had demonstrable, significant value for the owners, and for other financing companies using the same techniques. Yet, both courts decided that such inventions are not eligible for patenting on the grounds of “abstract,” i.e., “not useful.” Such decisions require ignoring the context of each invention – as presented in extensive detail in the specification, as well as in the claims.
Third: Inventions in fields other than software are rarely subjected to the level of dismissiveness that is characteristic of discussions of software inventions. 3
Consider U.S. Patent No. 8,479,619 (“JacLock screw and JacLock screwdriver”), and the following claim:
A tool driver for removing a fastener comprising: the tool driver comprising a body; a handle connected at one end of said body; an elongated shaft extending along a longitudinal center line; a blade means located at another end of said body and extending from an end of said elongated shaft; said blade means comprising a plate plane containing the longitudinal center line and two latitudinal plates extending from said longitudinal center line plate plane spaced equally apart from the center of the longitudinal center line; and said longitudinal center line extending from outer sides of said two latitudinal plates, thereby forming a H shape with lateral extensions.
Now imagine the following argument: “This claim is basically just the concept of a screwdriver. All screwdrivers have a handle, an elongated shaft, and a blade. The other details are trivial variations that don’t matter.”
Consider U.S. Patent No. 8,431,537 (“Glue composition for lung volume reduction”), and the following claim:
A method of reducing lung volume in a lung of a subject in need thereof, comprising: administering an effective amount of a glue composition to a localized region of damaged lung tissue of the subject, wherein the glue composition comprises a polyvinylalcohol polymer and a cross-linker, such that the polymer and cross-linker form a hydrogel in the lung, thereby reducing lung volume in the lung of the subject.
Now consider the following argument: “This claim is basically just the use of a hydrogel as a biologically inert filler.”
Consider U.S. Patent No. 8,362,990 (“Video processing system”), and the following claim:
A video processing system comprising: a video input; a video output for outputting at least one manipulated video image; a video processing chain connecting said video input to said video output, said chain including a series connection of at least two video processing components, each including: a component input for receiving at least one video image; a component output for outputting at least one processed video image; a control input for controlling the video processing component to be in an enabled state or a disabled state, said video processing component being arranged to obtain in said enabled state the processed video image by performing a respective processing operation on the received video images, in which disabled state the processed video image is inputted to a node downstream of the video processing component in the disabled state without being subjected to said processing operation.
Now imagine the following argument: “This claim is basically just the concept of a circuit that renders video.”
These types of arguments would never be made in these fields of technology, in the context of patent eligibility or novelty. An examiner who refused to look at the details, to give meaning to every word in each claim, would be chastised for performing an inadequate examination. An accused infringer who walked into court with that argument would be soundly defeated. Yet, such dismissiveness abounds in discussions of software inventions.
Do bad software patents exist? Of course – as do bad mechanical, electronic, and chemical patents. Patent examination is a difficult challenge, and the public serves an important role in finding prior art that invalidates incorrectly issued patents. But there is a strong public perception difference: instances of bad patents issuing in fields other than software prompt hostility toward the applicant, but bad software patents evoke hostility toward the entire field of software patenting.
Finally, this trend raises the question: Why is this dismissive attitude prevalent in discussions of the patent eligibility of software, and almost inapplicable to mechanical, electrical, chemical, and biotechnology fields?
I believe that these attitudes are a result of Parkinson’s Law of Triviality:
Parkinson dramatises his Law of Triviality with a committee’s deliberations on a nuclear power plant, compared to deliberation on a bicycle shed. While discussing the bikeshed, debate emerges over whether the best choice of roofing is aluminium, asbestos or galvanised iron, then over whether the shed is a good idea or not.
A nuclear reactor is so vastly expensive and complicated that people cannot understand it, so they assume that those working on it understand it. Even those with strong opinions might withhold them for fear of being shown to be insufficiently informed. On the other hand, everyone understands a bicycle shed (or thinks he does), so building one can result in endless discussions: everyone involved wants to add his touch and show that he is there.
Our culture is awash in software: computers, the internet, email, mobile phones, tablets, ATMs, checkout lanes at the grocery store – it is everywhere. It is impossible to participate in contemporary society without interacting with software. Its familiarity leads to a presumption that the underlying technology is easy. Significant functional differences can therefore be dismissed as mere design choices – while in other fields, even miniscule differences are presumed to be meaningful.
Returning to Alice Corp. v. CLS Bank, this attitude is reflected in Kennedy’s remarks:
KENNEDY: If you describe (the patented invention) to a second-year college class in engineering and said: here’s my idea, now you go home and you program over this weekend – my guess is that it would be fairly easy to program.
Again, this argument would simply not arise in the context of other fields of technology. The screwdriver, the hydrogel method, and the circuit claimed above could be “fairly easily” implemented by a tool craftsman, a biomaterials engineer, and an electrical engineer. Why is that factor relevant to patent eligibility for software?
Indeed, “fairly easy to implement” is a requirement of patent law: “The test of enablement is whether one reasonably skilled in the art could make or use the invention from the disclosures in the patent coupled with information known in the art without undue experimentation.”
The fact that Alice’s specification describes the invention in a manner that is “fairly easy to program” should be considered good patent practice! Instead, it is being held against Alice, as evidence that the invention is too “basic” for patent protection.
In conclusion: Currently, within the patent community – including the examining corps of the USPTO – software, as a technology, is evaluated from a fundamentally different perspective than other technologies. According to that perspective, software techniques can be oversimplified as “basically just” a concept; details can be disregarded as trivial; claim elements can be ignored as immaterial.
Is this peculiar perspective about software warranted? Certainly, it has no basis in patent law: neither 35 USC nor 37 CFR makes any provision for evaluating software differently than any other technology.
But before we can discuss the merit of this perspective, the patent community must first acknowledge that this peculiarly dismissive attitude toward software exists; that it is pervasive at all stages of the patent process, including examination, court review, and popular culture; and that it has far-reaching implications for patent law and the software industry.
- While claim 33 of Alice’s ‘479 patent did not recite any type of machine but simply a method, that factor was also not relevant to the “abstractness” of the claim. Claims 35-39 recited variations on a “data processing system,” but were held patent-ineligible by the per curiam opinion, for the same reasons as claim 33. ↩
- This aspect of the NPR segments reached a height of absurdity in the following exchange:
Alex Blumberg: 30% of US patents are essentially for things that have already been invented.
David Martin: So, for example, toast becomes the thermal refreshening of a bread product.
Alex Blumberg: There’s a patent on toast?
David Martin: Yes, thermally refreshened bread, not on toast.
Ladies and gentlemen, patent number 6,080,436, bread refreshing method, issued in 2000.
It’s clear that Martin and NPR failed to read the patent – which, incidentally, has only five figures and two pages – because U.S. Patent No. 6,080,436 actually claims:
A method of refreshing bread products, comprising:
placing a bread product in an oven having at least one heating element,
setting the temperature of the heating elements between 2500 F. and 4500 F., and
ceasing exposure of the bread product to the at least one heating element after a period of 3 sec. to 90 sec.
Seems a little different than “a patent on toast,” but such is the typical consequence of saying what a patent is “basically” about. ↩
- To be fair, the Mayo v. Prometheus majority opinion evoked a dismissive perspective of a biotech claim, which was widely characterized as “basically just the concept of differential diagnosis.” However, in other fields, such instances are regarded as individual aberrations, rather than an endemic problem with the entire field of technology and patenting. ↩