Sunday, December 19, 2010

First post-Bilski Federal Circuit decision upholds software method claims

Takeaway: In the first Federal Circuit decision to apply Bilski v. Kappos, the court held that two software method claims for halftoning gray scale images were not abstract, and so reversed the district court's invalidation under § 101. (Research Corp. Tech. v. Microsoft Corp., 2010-1037, Fed. Cir. 2010.) The claims did not recite a computer, but the court found that the "invention presents functional and palpable applications in the field of computer technology."

Research Corp. Tech. v. Microsoft Corp.
Fed. Cir. 2010-1037
December 8, 2010

The district court invalidated two method claims under § 101. A sample claim reads as follows:
11.  A method for the halftoning of color images, comprising the steps of
utilizing, in turn, a pixel-by-pixel comparison of each of a plurality of color planes of said color image against a blue noise mask in which the blue noise mask is comprised of a random non-deterministic, non-white noise single valued function which is designed to provide visually pleasing dot profiles when thresholded at any level of said color images,
wherein a plurality of blue noise masks are separately utilized to per-form said pixel-by-pixel comparison and in which at least one of said blue noise masks is independent and uncorrelated with the other blue noise masks.

The Federal Circuit's analysis first noted that "in this case, the subject matter is a 'process' for rendering a halftone image." The court found that this subject matter qualifies as patentable subject matter under both the explicit category list in § 101 and the definition of process under § 100 ("The term 'process' means process art, or method ..." ).

The court then proceeded to examine the three patentable subject matter exceptions in the Supreme Court's jurisprudence: laws of nature; physical phenomena; and abstract ideas. Since the parties agreed that the claims were not directed to laws of nature or physical phenomena, the court then considered whether the subject matter amounted to an abstract idea. The court decided that the subject matter was not abstract:
   The court perceives nothing abstract in the subject matter of the processes claimed in the ’310 and ’228 patents. The ’310 and ’228 patents claim methods (statutory “processes”) for rendering a halftone image of a digital image by comparing, pixel by pixel, the digital image against a blue noise mask. The invention presents functional and palpable applications in the field of computer technology.
... The fact that some claims in the ’310 and ’228 patents require a “high contrast film,” “a film printer,” “a memory,” and “printer and display devices” also confirm this court’s holding that the invention is not abstract. 

The court acknowledged that the methods included algorithms, and even that those algorithms were a "significant part of the claimed combination." Even so, the algorithms "do not bring this invention even close to abstractness that would override the statutory categories and context." Using language from Diamond v. Diehr, the court found that "the patentees in this case 'do not seek to patent a mathematical formula' " but instead " 'seek patent protection for a process of' halftoning in computer applications."

My two cents: Definitely a pro-software decision. I would expect district courts to uphold software claims when applying this decision.

The decision is consistent with my understanding of "abstract." And more specifically, my position that claims that cover a computer implemented method cannot, by definition, be abstract – i.e., the computer takes the claims out of the realm of the abstract idea and into the physical world.

I was mystified by all the talk about "the invention" and "the subject matter of the invention." I'm used to thinking about claims, and considering whether claims cover patentable subject matter. I wonder if the talk about the "subject matter" relates to reading claims in the context of the specification. Certainly these claims are limited to a computer implementation unless you ignore the entire specification and what it means to a person of ordinary skill in the art.

I'm glad to have some post-Bilski guidance from the Federal Circuit. But does RCT v. Microsoft provide much guidance for evaluating tougher cases? Does the decision amount to "we know abstract when we see it, and this ain't it" ?

So far we have two guideposts. Bilski's invention, which didn't mention computers in the claims or the spec, is abstract. RCT's invention – which under a reasonable reading clearly ties to a computer even though a computer isn't positively recited – is not abstract. I see those guideposts as being pretty far apart, and we'll have to wait for other cases to fill out the stuff in between.


  1. Karen,

    So many comments to make.

    1. I have no idea how this method is "palpable."
    2. Computer-implemented methods can be abstract. Check out Benson and Flook, which created the 101 doctrines of preemption and "insignificant postsolution activity."
    3. In general, the "invention" can refer to the invention of a claim. Claims define inventions. See Boundy's comments here:

    4. Still, Boundy's comments conflict with the establish Restriction Requirement of defining inventions in terms of something other than claims. And, of course, CAFC case law holds that references to the invention in the spec can insert further limitations into the claims.

    5. Moreover, this case introduces the idea of analyzing an "invention" in terms of the disclosure as a whole. At one point, the court suggests that method steps to the same invention will generally survive 101 if the apparatus steps do. I think this case is ripe for SCOTUS reversal on that ground alone.

    6. I completely agree that the decision offers only a "know it when you see it standard." It's not helpful to applicants. Then again, why would the CAFC come up with another bright line test just so that the SCOTUS could shut it down, even though it knows 1/100th as much patent law.

  2. Benson and Flook are bad examples to cite. The algorithm in Benson was about as abstract an equation as one can find. As to Flook, although commonly known as a case based on 101, the holding was really based upon the Court's belief that there was no claim to a patentable invention. Either way, both decisions were made during a time in which 99.9% of the people didn't understand the technology or its ultimate ramifications of this technology to everyday life.

    The Research Corp case is very from the facts of either Benson and Flook. These claims are nowhere near as abstract as Benson and there has been no allegation of using a known formula in the context of a computer (as in Flook).

    This is a case that the Patent Office is going to hate (or love) depending upon whether or not they are predisposed one way or another with regard to statutory subject matter.

    It should not be that difficult to draw the line between what is patentable under 35 USC 101 and what is not. The claimed process is a practical application of technology. Too much time and effort has been wasted at the USPTO and the Courts splitting hairs over these things. There may be things out there that could be considered "abstract," but this isn't one of them.

  3. Karen,

    There is literally nothing you can point to in this opinion to help make a case that a patent fails under 101. It seems as if Rader chose his words carefully to discourage parties from making 101 arguments in the future.

    Even Prometheus v Mayo consists mostly of conclusory statements supporting their, dare I say, predetermined conclusion.

    The real test will be the Classen remand.

  4. FYI -- 2 weeks after the decision and not a single BPAI decision relies upon this case to reverse a 101 rejection -- (at least based upon publicly available information).

    Methinks that the BPAI wants to pretend this decision never happened.

  5. >not a single BPAI decision relies upon this
    >case to reverse a 101 rejection

    Someday an Applicant with deep pockets will appeal an affirmed 101 to the Fed Cir, brief it properly, and win under RCT v. Microsoft.

  6. "Someday an Applicant with deep pockets will appeal an affirmed 101 to the Fed Cir, brief it properly, and win under RCT v. Microsoft."

    Until then, the BPAI will continue to make it up as they go along -- almost always to the detriment of Appellants.

  7. Question for you patent law experts: Suppose a patent describes a process that would, if performed by a human being, be bloody obvious (for example, there would be plenty of prior art one could dredge up of people doing this same basic thing by hand--if you see what I mean), and the patent really just adds in the further specification that all this should be performed with a computer (without contributing anything innovative about the implementation). So, have the courts clearly held (I hope :-) ) that something like this should not satisfy the nonobviousness criterion for patentability? Thanks! (from a nonlawyer inventor)

  8. >patent really just adds in the further
    >specification that all this should be performed
    >with a computer (without contributing anything
    >innovative about the implementation)
    >So, have the courts clearly held [this is obvious]

    While obviousness is fact dependent, and there is no bright line rule ... Yes, several important cases have held this sort of thing to be obvious. For example:

    Leapfrog Enters., Inc. v. Fisher-Price, Inc., 485 F.3d 1157, 1161 (Fed.Cir.2007) ("Accommodating a prior art . . . device . . . to modern electronics would have been reasonably obvious to one of ordinary skill in [the art]" because "[a]pplying modern electronics to older . . . devices has been commonplace in recent years.");

    KSR Int'l Co. v. Teleflex Inc.,127 S. Ct. 1727 (2007) (addition of a well-known electronic sensor to a well-known mechanical adjustable pedal would have been obvious);

    Dann v. Johnston, 425 U.S. 219, 220, 230, 96 S.Ct. 1393, 47 L.Ed.2d 692 (1976) (finding it obvious to combine the modern computer program described in the patent with "existing machine systems in the banking industry").

  9. Thanks, Karen, for your super-helpful response.