New York Times Op-Ed Argues Law Takes Misguided Approach to Software Patents

By Michael Borella --

On March 28, Professor Robin Feldman of the University of California Hastings College of Law wrote an op-ed piece in the New York Times entitled "Slowing the Patent Trolls".  Unfortunately, like so many articles of its ilk, Professor Feldman's offering uses misleading hypotheticals and unsupported assertions to allege that the mere existence of software patents is allowing the so-called trolls to harm our economy.

Ostensibly written in anticipation of Monday's Alice v. CLS Bank oral arguments scheduled to be heard by the Supreme Court, Professor Feldman writes that the case "offers the court an opportunity to resolve two decades of economically harmful confusion over how the law grants patent protection to computer software."  Professor Feldman, however, seems to be focused not on the specific facts that lead to the Court's review of Alice's patents, but instead on her erroneous belief that software patents are, by their very nature, too broad.

She immediately starts off on the wrong foot by stating that "[s]ince the mid-1990s the software patent system has operated by its own rules . . . compared with patents for other innovations, those for software are granted using a very broad and lax standard of invention."  This is untrue, even as an opinion.  There certainly are specific rules regarding computer-implemented inventions, particularly with regard to subject-matter eligibility (see, e.g., the USPTO's post-Bilski and post-Mayo examination guidelines).  Additionally, several well-known Federal Circuit cases have resulted in the requirement that, for inventions that operate on a general-purpose computer claimed in means-plus-function format, the specification must provide an algorithm that performs the claimed functions (see, e.g., M.P.E.P. 2181(II)(B) for an overview of the case law).

If anything, computer-implemented inventions are held to a higher standard than most other types of inventions due to these additional requirements.  Rather than acknowledging that the computing arts have been singled out in this fashion, Professor Feldman contends that "[o]rdinarily, the law requires inventors to explain not just the result of an invention, but also how the invention actually works.  If you invent a car that drives on water, you have to explain exactly how you get it to stay afloat.  Not so for software: the mere idea of a floating car is enough."

This statement is a misrepresentation of the law.  If Professor Feldman can cite to an issued patent that claims software for a flying car but does not disclosed how the invention works, I will happily join her in condemning that specific patent for being vague and overly broad.  But even if such a patent comes to light, Professor Feldman's statement is yet another example of the hyperbole that is drowning out reason in an otherwise important debate.

Surely, not all patent applications that claim software should be allowed.  Some claims will be anticipated, obvious, or vague.  But this analysis is currently carried out on a case-by-case basis by USPTO examiners who apply the guidelines cited above, as well as a comparison of the claims to prior art.  Professor Feldman's flying car hypothetical would simply not pass muster under the existing examination standards.

Professor Feldman provides another hypothetical about a "software program to assess the riskiness of a driver for car-insurance purposes . . . based on how much the person texts while driving, combined with other risk indicators."  She adds that "[u]nder the current software patenting paradigm, providing the description above without much more could be enough to merit protection."  As discussed above, this notion is unsupported by the facts.

Regardless, she proposes that if one were to "specify the particular inputs used -- texting frequency, credit score, hobbies, driving history -- along with the weights and multipliers and software approach used to produce the risk score" the resulting invention would be patent-eligible.  Maybe.  The problem with approaching the topic of patentable subject matter at such a high level of abstraction is that patentability decisions are made based on the specific language of the claims when read in light of the invention as disclosed in the specification.  Hypotheticals may be useful thought exercises, but patent eligibility decisions must be made based on facts and the law.

Now turning briefly to her patent troll contentions -- Professor Feldman fails to acknowledge that the United States' patent system purposefully encourages the existence of non-practicing entities by making patent property rights freely assignable and eschewing a "working" requirement.  This is thought to help make the patent system accessible to those without abundant resources.  Indeed, public universities, such as the one that employs Professor Feldman, are permitted to acquire and hold patents related to the inventions fostered by research conducted under their auspices.  Further, as we have pointed out recently, calculations of the supposed harm due to such non-practicing entities are questionable at best.

The issues of patent-eligible subject matter and patent "trolls" are neither simple nor black-and-white.  There is a need for a rich, informed public discourse regarding both.  But this discourse is not aided by drawing questionable conclusions from an incomplete view of patent law.

Comments

[EG]

Michael,

As you suggest, Feldman's op-ed piece in the NYT is intellectually dishonest in the extreme. And it's not a surprise given what clearly appears to be someone with no undergraduate background or work experience showing any knowledge of science or technology: “Professor Robin Feldman received a bachelor’s degree from Stanford University and a J.D. from Stanford Law School.” There is no reference to any undergraduate degree in science or engineering in the cv on her web site http://robinfeldman.com/cv.htm , which lists a BA undergraduate degree. That doesn’t necessarily mean a non-science degree as I’ve got a BA degree in chemistry, but if you look at the articles she’s written, I would presume she has none, given her obvious disdain for science (something she shares in common with Justice Breyer amongst others). In other words, an IP law professor with no suitable undergraduate or work background to be teaching IP law, especially be teaching patent law. As others have said, I frankly pity the students who learn IP law, in particular patent law, from this charlatan.

[Mike Borella]

EG,

Thanks. There are valid points to be made on both sides in the software patent debate, but articles like this aren't making any. Even worse, given the weight that a New York Times op-ed holds in some circles, it does a gross disservice to the public to scope the issues so incorrectly.

Mike

[Skeptical]

All is fair in love and war, Mike.

Make no mistake, there is a revolution underway, and not all battles are fought in blood. Many modern day battles are fought in ink.

[Igor Faynshteyn]

I partially agree and disagree with this post. Professor Robin Feldman makes important points about the software patent debate. She is hardly the only one. As you may know, Judge Richard Posner is very skeptical of software patenting too. So are many other academics, judges, attorneys and scientists.

To be fair, it is a NYT Op-Ed piece, which is not expected to go into details, and nuts & bolts of software patenting. And it is a legitimate exercise of academic criticism to look at the entire system as a whole and make generalizations about it. There is no conflict between making generalized statements and the notion that software patenting is decided on a case by case basis.

Nor is there a conflict between saying, that "patent system purposefully encourages the existence of non-practicing entities by making patent property rights freely assignable and eschewing a 'working' requirement," and that the system may be malfunctioning due to so many lawsuits being filed and litigated over software patents, most of which are commenced by NPE's.

A system may be designed to encourage a particular behavior, but it may very well result in encouraging a whole array of other behaviors, which may be harmful. Although the analogy to financial securitizations (and other financial instruments) may not be apt on many levels, in a broad way it demonstrates my above point, as the financial crash of 2008 showed.

In my opinion, a Supreme Court opinion, unless sweeping in scope, will not accomplish much in terms of allocating incentives differently as to effect behavior on the market that may produce more positive results (however it is defined). It seems to me that providing better training to PTO examiners, hiring more talented individuals and generally providing more resources to the PTO will likely accomplish more in the way of repairing the damage to the system than any new legal tests, either from CAFC or the Supreme Court.

Lastly, I agree that it is debatable whether NPE's are really harming the system or not, since it hinges on how one designs empirical studies in this area. And it may very well be too early to tell.

[TBP]

Michael: "She immediately starts off on the wrong foot by stating that "[s]ince the mid-1990s the software patent system has operated by its own rules . . . compared with patents for other innovations, those for software are granted using a very broad and lax standard of invention." This is untrue,"

No, it's not untrue at all. In other fields, you aren't allowed to distinguish the prior art from your invention solely on functional terms. The so-called "antibody exception" is the only other area where this is permitted, and I doubt the Supreme Court would approve of that exception if it were ever put to the test (I'm not even sure en banc Federal Circuit would approve).

"If Professor Feldman can cite to an issued patent that claims software for a flying car but does not disclosed how the invention works, I will happily join her in condemning that specific patent for being vague and overly broad. But even if such a patent comes to light, Professor Feldman's statement is yet another example of the hyperbole"

Seems you're the one engaging in the "hyperbole", Michael. Professor Feldman was making an argument by analogy. You didn't realize that? The point is that computer-implementers need not describe in detail a single working embodiment of their alleged "innovation." They can (and do) get away with simply describing the desired function for a computer which, in nearly every case, has never been reduced to practice, constructively or otherwise. The fiction is that the mere description of an information-processing functionality is sufficient to justify ownership of all computing devices that are subsequently programmed to carry out those functions, regardless of the operating system and regardless of any practical hurdles that are encountered.

Conviently for those invested in the status quo, however, this fiction operates as a one-way street. Although the usefulness of computers for carrying out any and every information processing function has been well-known for decades, the prior art is invariably read narrowly such that a reference teaching the delivery of information about a movie upon a request (or in response to some "user action") is frequently deemed irrelevant to a claim describing the delivery of information about a heart condition. Everybody knows the technology is identical. All that has changed is the information going in and out.

"The issues of patent-eligible subject matter and patent "trolls" are neither simple nor black-and-white. There is a need for a rich, informed public discourse regarding both."

Agreed. As you can imagine, it's very difficult to have the discussions with folks who are deeply invested in the status quo.

[EG]

Michael/Skeptical,

I suspect that "TBP" is Malcolm Mooney (aka "MM") in disguise. Again, "guilty" until proven otherwise.

[Mike Borella]

"The point is that computer-implementers need not describe in detail a single working embodiment of their alleged innovation.'"

If this were the case, the patent would be invalid under section 112.

"They can (and do) get away with simply describing the desired function for a computer which, in nearly every case, has never been reduced to practice, constructively or otherwise."

Constructive reduction to practice of an invention is, by definition, filing a patent application for the invention. But in order for the constructive reduction to practice to be successful, the application must meet the requirements of section 112.

Neither of these points invoke section 101, and as many of us have been saying for quite some time, issues of written description and vagueness should be determined under the proper provision of 35 U.S.C. Hint: it isn't 101.