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.