Possible Strategies For The Bilski Machine-or-Transformation Test

by Brian Fletcher on April 17, 2009

I attended a recent BNA audioconference on patent prosecution and post-grant considerations in view of Bilski.  The conference offered the insights of practitioners Kenneth N. Nigon of RatnerPrestia, Valley Forge, Pa., and Stephen C. Durant of Duane Morris, San Francisco, on drafting claims and seeking to reissue patents put in doubt by the Federal Circuit’s Bilski decision.  What follows is a summary of the conference. 


Bilski Machine-or-Transformation Test


As determined by the Federal Circuit in In re Bilski, 545 F.3d 943, 88 USPQ2d 1385 (Fed. Cir. 2008) (en banc) (77 PTCJ 4, 11/7/08), “A claimed process is surely patent-eligible under § 101 if:  (1) it is tied to a particular machine or apparatus, or (2) it transforms a particular article into a different state or thing.”


To the “machine prong” of this test, Nigon concluded that a “particular machine” may be:

• a known apparatus;

• a special purpose computer;

• a general purpose computer programmed with particular software; or

• a computer-readable storage medium.


To qualify under the “transformation prong” of the Bilski test, Nigon said, the thing being transformed must be either tangible or data representing something tangible, what he called “meta-tangible.”  He added that the identified machine or transformation must be central to the invention and have specific utility and not just extra-solution activity.


Pre-Bilski Precedents Still Relevant


Before looking at each prong of the test in more detail, the presenters identified pertinent Federal Circuit precedents that relate to patent-eligible subject matter for processes and were not specifically invalidated by Bilski:


• The court found patentable a method claim that included both a computer and a digital plotting apparatus in In re Bernhart, 417 F.2d 1395, 163 USPQ 611 (C.C.P.A. 1969).  So, Nigon asked, “If a claimed process includes references to machines such that the process cannot be performed mentally, is that sufficient to satisfy the ‘particular machines’ branch of the test?”


• The allowed patent in In re Lowry, 32 F.3d 1579, 32 USPQ2d 1031 (Fed. Cir. 1994) read on data structures for an object-oriented database that the court found to be “physical entities that provide increased efficiency in computer operations,” which Nigon assumed to be still patentable as “transforming the memory of a particular machine.”


• The claims in In re Warmerdam, 33 F.3d 1354, 31 USPQ2d 1754 (Fed. Cir. 1994) (48 PTCJ 409, 8/18/94) would still be held to be non-patentable, he said, as the court found that all of the method claims recited only mathematical algorithms, with no machine recited and no tangible or meta-tangible data transformed.


• A diagnostic test-based claim in In re Grams, 888 F.2d 835, 12 USPQ2d 1824 (Fed. Cir. 1989) (39 PTCJ 23, 11/9/89) was not patent-eligible subject matter when the court determined that the only physical step in the method, “performing a plurality of clinical diagnostic tests,” was mere data gathering—i.e., extra-solution activity—for the algorithm defined in the remaining steps.


BPAI on Particular Machines


Durant then reviewed a number of post-Bilski decisions by the Board of Patent Appeals and Interferences with reference to the machine prong of the test.  He identified cases that appeared in conflict with the precedents cited above or with each other.


Though each claim step recited a different function performed “by a processor,” the board rejected under Section 101 claims in Ex parte Cornea-Hasegan, No. 2008-4742, 89 USPQ2d 1557 (B.P.A.I. Jan. 13, 2009) (77 PTCJ 291, 1/23/09).  Durant asked how that can be reconciled with Bernhart‘s acceptable claim of “printing using a plotter.”


In contrast, in Ex parte Kasper II, No. 2008-2297 (B.P.A.I. Feb. 2, 2009), he noted, the BPAI found statutory a method directed to processes that interact with a “local unit” to receive data through a “data link layer.”


He also compared e-commerce claims in Ex parte Borenstein, No. 2008-3475 (B.P.A.I. March 30, 2009) with Ex parte Harris, No. 2007-0325 (B.P.A.I. Jan. 13, 2009).  The former recited “path information” inherently implying that catalog information must be stored on a computer or database, and so was found to be patent-eligible.  The latter’s references to bidding “over a network to a server which collects said bids” was seen by the board as reciting “mere field-of-use limitations,” Durant noted.


Claiming Machine Ties


Nigon said that the disparity between the board’s decisions can be traced in part to the different panels of judges, but the “undercurrent is the use of the word ‘particular’ ” in the opinions.


He suggested that claim drafting that identifies functional steps “performed by something” are more likely to meet with BPAI approval.


To make the point, he built on the following example of a software method claim:  An encryption method comprising: (1) performing a first mathematical calculation on a data object to generate a first result; (2) performing a second mathematical calculation on the first result and an encryption key value to generate a second result; and (3) providing the second result as the output signal of the encryption method.  That claim would fail both prongs of the Bilski test, he said.  Add the phrase “on the computer” after each mention of the word “performing” in steps (1) and (2) and you have made it a little bit better for passing the machine prong, Nigon said.  But some panels would say that is a field of use limitation and still preempts the abstract idea of encryption.


Now consider identifying “a first computer process” as the operator in the first step and “a second computer process” in the second step, Nigon said.  Because you now have specific computer steps, you probably have properly limited the invention to meet the particular machine prong of the Bilski test, he said.


Of course, Nigon noted the trade-off in making that limitation, such that someone who could do the encryption in a single process would not be infringing.  He therefore suggested that a good approach would be to claim the computer in an independent claim and provide the additional processing limitations in a dependent claim.


Nigon had the same basic advice for business method patents, and also noted that flow charts may not be enough, unless a specific algorithm is claimed and the hardware is disclosed.  “A particular algorithm when combined with a general purpose computer is a new computer,” he said.


An alternative or addition would be to have a functional block diagram, Nigon said, describing the hardware system, in terms of the functions of the method performed.  If the business method uses databases, a search process and/or a user interface, those functions require machines in any case, he said.


BPAI on Transforming Articles


Durant identified a number of BPAI cases that failed to identify a Bilski-qualifying transformation.  For example:


Ex parte Barnes, No. 2007-4114 (B.P.A.I. Jan. 22, 2009) claimed the determination of “a planarity value” in seismic data.


• The “namespace” transformed in the steps claimed in Ex parte Zybura, No. 2008-2195 (B.P.A.I. Feb. 3, 2009) was determined not to be a physical object.


• The applicant in Ex parte Atkin, No. 2008-4352 (B.P.A.I. Jan. 30, 2009) attempted to claim a method to convert a unidirectional domain name to a bidirectional domain name, which the board found to be mere “data manipulation.”


Advice on Adding Transformation to Claims


Durant then considered how he might change the claim in Barnes, for example, which reads “A fault identification method that comprises: obtaining seismic data; and for each of multiple positions of an analysis window in the seismic data, determining a planarity value for discontinuities in the analysis window.”  He recalled how the Bilski opinion cited with approval a claim approved in In re Abele, 684 F.2d 902, 214 USPQ 682 (C.C.P.A. 1982), that recited x-ray attenuation data that “clearly represented physical and tangible objects, namely the structure of bones, organs, and other body tissues.”


Applying that lesson to Barnes, Durant asked: “Could this claim have been rewritten so that seismic data represents a physical object, and the planar data is produced by transforming that seismic data?”


He said there should be an emphasis in the specification that seismic data is physical, and an emphasis in the claim that the seismic data is transformed and represented differently in planar data.


In a second example related to diagnostic methods, Nigon looked at a claim rejected in Prometheus Laboratories Inc. v. Mayo Collaborative Services, No. 04cv1200, 86 USPQ2d 1705 (S.D. Cal. 2008) (75 PTCJ 642, 4/18/08), which he paraphrased as: “a diagnostic method comprising:  administering a drug at a dosage; measuring a level of a metabolite of the drug in blood; if the measured level is less than a first threshold, indicating a need to decrease the dosage; and if the measured level is greater than a second threshold, indicating a need to increase dosage.”


Nigon suggested that comparison is not processing and measuring is mere data gathering, so the only use of the method preempts a fundamental principle.  In looking for ways to make the claim patent-eligible, he asked whether normalizing the measured metabolite level to compensate for other drugs would be an appropriate limitation.


Reissue of Existing Patents


For both Bilski prongs, in any case, Nigon told the conference listeners to put in a range of claims because it will take at least a couple of years for the PTO to take up the application, and post-Bilski jurisprudence might be more definite by the time it is examined.


Then, for important patents issued before Bilski, with claims that may not meet either prong of the new machine-or-transformation, Nigon recommended the owners consider reissue, rather than waiting for the patent to be litigated.


Check the specification for a basis to add a general purpose computer to claims, at least, he said.  If you have a flow chart or algorithm in the specification, you may also be able to add a basic block of a computer.  Then, he said, add a range of claims that satisfy the Bilski test.

Previous post:

Next post: