Klemens: Bring Back Freeman-Walter-Abele For Software Patents!

Ben Klemens, author of "Math You Can't Use," is a guest scholar at the Center on Social and Economic Dynamics at the Brookings Institution, where he writes programs to perform quantitative analyses and policy-oriented simulations. He also consults for the World Bank on intellectual property in the developing world and computer-based simulations of immigration policy.

Ben is also an aggressive critic of software patents (who has previously butted heads with the 271 blog on this topic), and has just published a working paper titled "Drawing the Line: The Rise of the Information Processing Patent." The paper highlights many of the points brought up in his book, and provides some updates on recent developments in software patents.

Ben's work is important because he is one of the few critics that does a decent job at times in dissecting many of the problems facing the software industry in light of patents (as opposed to IP anarchists Eben Moglen and Richard Stallman, who propose banning software patents and other IP entirely). While I do not agree with many of his conclusions (e.g., software is math and is therefore unpatentable), his papers are often thought-provoking and well-written.

The timing of Ben's most recent article is interesting, since the criticism of software patents has now served as a catalyst for other industry and business members to call for sweeping reforms that would limit patentable subject matter and restrict patent rights across the board. While people have reflexively equated patents with a "pro-business" outlook, this has started to change. As Professor Wegner stated in his recent post on Patently-O, in today's business climate, being "pro-business" means to some people that you are "anti-patent."

And this sentiment is showing support in the USPTO as well. Setting backlog issues aside, the PTO has always been resistant to expanding the scope of patentable subject matter, and continues to be vexed over the proper handling of software and business-method patents. Recently, the undeniable fact exists that the PTO has renewed efforts to find any way to limit patentable subject matter (sua sponte, I might add), and has demonstrated unusual eagerness in aggressively attacking software patents. After reading Ben's paper, I was surprised how much Ben's recommendations had in common with the PTO's Interim Guidelines for Examination of Patent Applications for Patent Subject Matter Eligibility.

Critics contend that the primary problem with software patents is that they do not apply singularly to a specific "software" industry. While certain "centralized" technologies (such as biotech) have an inherently limited base of patents and patent owners, Klemens argues that the cost of R&D and compliance on account of patents is relatively straightforward, since most companies know who the developers are, and the technologies they normally develop in. However, since software is "decentralized," patents related to software have the potential to spill over to all industries, making patent compliance extremely difficult, if not impossible to accomplish.

While companies fabricating semiconductors or synthesizing drugs know where most of the patent threats lie, the same cannot be said of companies trying to assess a patent threat on software being used within the company, regardless of whether the software was developed there or not. As such, situations arise where a telephone company (e.g., Southwestern Bell Corp.) can send cease-and desist letters to a toy company (e.g., Museum Tour) on software patents related to a web-based structured document browser. As Klemens points out, "[i]t is unlikely that the owners of Museum Tour thought of themselves as a computer software firm, but by having a web site, they effectively were."

Accordingly, Klemens argues that

[i]t is wrong to think of software patents as only affecting companies that vend software. The majority of software designers and authors are at companies where software is simply necessary for daily operation. Whereas the set of people who have the know-how, equipment, and inclination to work on a mechanical device may be limited to a few hundred organizations, the set of people who work full-time on writing new and potentially innovative software is distributed among millions of organizations.

The paper goes further to posit that one of the main reasons why software patents permeate through multiple industries and cause problems is due to the elimination of the “insignificant postsolution activity” test put forward in Freeman-Walter-Abele:

The size of the software and business method pseudo-industries derives directly from the elimination of the concept of insignificant postsolution activity. Information and the tools to manage information are truly ubiquitous, but the equipment needed to engage in significant postsolution activity, such as testing a chemical’s properties or verifying that a machine actually runs, are available to a more limited range of people. Therefore, when inventions with no significant postsolution activity are allowed, the number of potential innovators grows exponentially.

* * *

The problems with patents in a massively decentralized industry stem from the elimination of the “insignificant postsolution activity” stipulation on patentable subject matter. Virtually every company in the country has a computer on hand, and those of a certain scale have full-time employees actively writing code. Yet only a handful of companies have employees actively synthesizing new drugs. This is closely related to the fact that information processing is a general requirement for anyone with information, and that information processing tools are considered to be stock equipment.

Thus, the recommendation of this paper is not an innovation, but a regression. A great many of the problems with patents that fill the newspapers and vex businessmen can be solved by reinstating the distinction from Diehr and its predecessors that indicate a device is patentable only if it is based on steps that are simultaneously novel and non-trivially physical.

Of course, at this point, most of Ben's suggestions are purely academic, since CAFC caselaw has expressly overruled Freeman-Walter-Abele. However, in light of the current climate perceived at the USPTO and Supreme Court, these is the distinct possibility that these questions will be raised again sooner than we think.

Download/read Ben's paper here.

Read the transcript from Brookings Institution symposium "Software and Law: Is Regulation Fostering or Inhibiting Innovation?"

• Patent Trolls: The paper addresses "patent trolls" and also provides a revised definition of "patent troll," which, in my opinion, is more accurate: "a patent troll is one who unfairly takes advantage of informational asymmetries by suing independent inventors who are ignorant of the field of patents in which the troll works."
• "The Rule of 25": Chip Lutton, Chief Counsel of IP at Apple testified before the House Subcommittee on Courts, the Internet, and Intellectual Property on April 20, 2005, where he offered a “rule of 25” on patent litigation: for every case litigated, there are 25 that are filed, and for every 25 that are filed, there are 25 notice letters claiming a patent
  infringement sent out. For more info, see here and here.