Machine Inconceivable

What’s black and white and more mystifying than a book on quantum physics written in cuneiform? That’s right, the Federal Circuit is out with another opinion on patentable subject matter.

Dealertrack, Inc. v. Huber, No. 2009–1566, –1588 (Fed. Cir. Jan. 20, 2012) (Judges Linn, Plager, and Dyk)

Let’s start this post off with a quiz. I’ll give you the start of four patent claims, and you tell me which ones (if any) require a computer. Here’s a hint: none of the claims mentions a computer in any of its method steps.

Four claims – which one(s) is (are) patentable subject matter?

Claim A – A method for verifying the validity of a credit card transaction over the Internet
Claim B – A method for distribution of products over the Internet via a facilitator
Claim C – A method for the halftoning of gray scale images
Claim D – A computer-aided method of managing a credit application

If you said claim A, unfortunately you’re wrong, because the Federal Circuit held in CyberSource Corp. v. Retail Decisions, Inc. that the Internet isn’t a machine, so the method doesn’t require a computer.

On the other hand, if you said claim B, you’re right, because the Federal Circuit held in Ultramercial, Inc. v. Hulu, Inc. that using the Internet requires complex computer programming, so it does require a computer.

Claim C looks easy, because it doesn’t even mention a computer in the preamble or in the claim. Oh, wait, that does require a computer, according to the Federal Circuit’s decision in Research Corp. Technologies v Microsoft Corp.

Well, claim D must need a computer, because it’s a “computer-aided method,” right? Nope, sorry. The Federal Circuit held in this case, Dealertrack, that only being “aided” by a computer isn’t enough.

If you even tried my little quiz, you fell victim to one of the classic blunders, the second most famous of which is never go in against a Sicilian when death is on the line. Only slightly less well known is this: never try to predict which way the Federal Circuit will decide when it comes to patentable subject matter!

In this case, Dealertrack sued for infringement of several patents for processing credit card applications over computer networks:

Dealertrack Patent Process Illustration

The decision is fairly long (about 40 pages) and deals with a number of issues, but I think that the patentable subject matter question is the most interesting part. The district court found that several claims in one of Dealertrack’s patents were invalid for failing to claim patentable subject matter, although the ruling was pre-Bilski and used the old “machine or transformation” test.

The majority agreed with the trial judge that the claims are not directed to patentable subject matter. They said that the method itself is abstract and just adding the words “computer-aided” in the preamble doesn’t save it. The claims aren’t tied to any particular implementation or machine, so they’re still abstract.

The court distinguished this case from Ultramercial, saying that:

Unlike in Ultramercial, where this court found that the patent claimed a practical application with concrete steps requiring an extensive computer interface, id. slip op. at 11, the claims here recite only that the method is “computer-aided” without specifying any level of involvement or detail. The fact that certain algorithms are disclosed in the specification does not change the outcome. (Page 36 of the opinion.)

I have to be honest about this: I don’t think that explanation holds up. The claim at issue in Ultramercial didn’t recite any particular computer-implemented steps. The court said in Ultramercial that, “Many of these steps are likely to require intricate and complex computer programming” (Ultramercial slip opinion at 11), but there was no specific algorithm in the claim at issue.

The difference is the composition of the panel. The majority here was Judges Linn and Dyk (Judge Plager dissented from considering the patentable subject matter question at all). Judge Dyk wrote the opinion in CyberSource, which came to a similar conclusion. I’ve already written about this split on the court a couple of times.

I’ve tried to find some way to reconcile the various subject matter patentability of software cases, but I don’t think there is one.

Until we get a rehearing en banc, we’re stuck with essentially contradictory precedent, with CyberSource and Dealertrack on one side and Ultramercial and RCT on the other. This is a mess for everyone, district court judges in particular. I have no idea what they’re supposed to do now. If a question of subject matter patentability of software comes up, each side will cite Federal Circuit cases in its favor, and the judge will have no useful guidance.

Maybe the judge’s best option is to follow Judge Plager’s opinion in his dissent here and just avoid reaching the issue. That’s not exactly satisfying, though.

You can read the Dealertrack opinion for yourself and see what you think.