Wednesday, September 30, 2009

Another computer-implemented business method invalid under 101 (Ex parte Myr)

I haven't yet covered §101 issues in any BPAI cases because the BPAI Watchdog blog and the 12:01 Tuesday blog do a great job on this topic. However, this post will discuss Ex parte Myr because it provides an in-depth analysis (7 pages worth) as to why a "computer-implemented method for appraising real estate property" does not meet either prong of the Bilski test. Unfortunately, it's possible that the same reasoning will be used to reject many other computer-implemented but non-business methods.

Ex parte Myr
Decided September 18, 2009
(Appeal 2009-005949; Appl. No. 10/610,955; Tech. Center 3600)

 The claim at issue is:

1. A computer-implemented method for appraising a real estate property, the method comprising the steps of:
a) storing influence factors and a range of influence factor values for each of different types of appraisal approaches;
b) performing nonlinear programming with a predetermined nonlinear objective function that uses each of the different types of appraisal approaches according to the stored influence factors and the stored range of influence factor values; and
c) providing signals indicative of an optimal range of appraisal values for the real estate property from the performed nonlinear programming according to each of the different types of appraisal approaches, wherein each of the different types of appraisal approaches are a sales comparison approach, an income capitalization approach and a cost approach.

The Board spends most of the first half of the analysis reducing the 3 steps of the method to their essence, which it finds to be: "a series of 'computer-implemented' steps for gathering data, manipulating a mathematical formula, and communicating a result." (pp. 14-15.) Such a reduction leaves out all specifics of the data being manipulated by the steps, but the Board says that's just intended use. ("The field of use for the formula in the claimed process is real estate appraisal."; p. 15.)

Having reduced the steps to their essence, the Board then notes that the claim is nominally drawn to a process, and therefore turns to the Bilski test to determine whether the claim passes muster under § 101. (p. 15.) The Board finds that the phrase "computer-implemented" in the preamble is not enough to satisfy Bilski. (p. 16.) Therefore, the Board moves on to the body of the claim. Quoting a couple of passages from Bilski that deal with data-gathering and insignificant post-solution activity, the Board then finds that the first ("gathering data") and last ("communicating a result") steps "do not impose meaningful limits on the claim's scope." (p. 16.) The Board seems to have arrived at this conclusion because these steps are not "limited by structure or apparatus." (p. 12 and p. 14.)

Having quickly disposed of the first and last steps, the Board turns to the remaining step, which it earlier described as "manipulating a mathematical formula". Since the claim involves a formula, the Board then applies the "transformation" prong of the Bilski test, which specifically references the application of mathematical formulas. Under the transformation prong, § 101 is satisfied when a claim applies a formula in a process which transforms an article to a different state or thing. (pp. 16-17.) However, the Board finds that for this claim, the answer to the transformation question is No because while the property being appraised is physical, the appraisal doesn't transform the land, but rather transforms the rights in the land. (p. 18.)

The Board then moves on to the "particular machine" prong of the Bilski test. The Board finds that the "computer" referred to in the preamble is a general purpose computer, and "thus, the claim covers implementing a particular algorithm in a general purpose computer for the appraisal of a real estate property". (p. 17.) An algorithm is an abstract idea under State Street, and Diehr says an abstract idea is unpatentable under § 101. (p. 17.) Finally, the computer in the preamble does not satisfy the "particular machine" because it's merely a "token recitation", which fails to to qualify a method as a statutory process under Benson. (p. 18.)

It appears this reasoning could be used to reject many computer-implemented claims. Not the small percentage that result in a physical transformation of an article, such as controlling a rubber curing machine or fuel injection in an engine. And not those that are tied to a "particular machine" — I'm thinking of something like a DSL modem or an IP router.

But at some level, can't all computer-implemented claims be characterized as gathering data, performing an algorithm, and communicating a result? After all, that's basically what computers do, right? And if that's all the claim does — if it doesn't transform an article and isn't tied to a particular machine — then it seems the claim doesn't pass § 101 under this reasoning.

Maybe the area to fight about is "performing an algorithm." The claim in Myr seems particular vulnerable because it was tied to a formula, and formulas remind us of "mathematics" and "abstract," which § 101 prohibits.

But what about the following sets of verbs/verb phrases? I think these verbs definitely take the claim out of the realm of "formula." But is a claim using each of these sets of these terms still an "algorithm"?
  1. selecting; identifying; determining; comparing
  2. applying weighting factors; generating results
  3. assembling a data set; assessing an impact 
I think that #1 sounds more like an algorithm than #2, which sounds more like an algorithm than #3. That is, I think people have a certain perception of what an algorithm is, and some words trigger this more than others. So maybe the trick is to choose words that don't sound like an algorithm.

No comments: