As yesterday's post explained, 101 rejections at the BPAI are at an all-time high. Particularly troubling are the rationales supporting many of the decisions. Over the last year, the BAPI has regressed 35 U.S.C. 101 back to the Warmerdam days, where methods describing algorithmic processes are presumed unpatentable, while machines performing algorithmic processes will be allowed.
This, of course, was done away with in State Street, but you couldn't tell this from reading many of the BPAI's opinions. In fact, it is apparent that the BPAI doesn't favor State Street at all, since the case (as well as AT&T) is typically cited only in passing. Instead, the BPAI starts with the Supreme Court cases - Gottschalk v. Benson (decided in 1972) and Diamond v. Diehr (decided in 1981) - and then proceeds to cite many of the pre-State Street CAFC cases to support their decision to find the claim(s) unpatentable.
This approach is played out in numerous BPAI opinions:
Appeal 2007-1251 (U.S. Application 09/451,097):
Claim 37 recites two steps, (1) calculating statistics of motion vector information, and (2) generating a frame feature value comprising numerical information representing a quantity of a feature contained in a frame of image data using the calculated statistics. Both steps are mathematical functions, and the result is a mathematical value. Further, the claimed method includes no recitation of a computer. Thus, the method appears to be a disembodied concept.
[W]e find no physical subject matter being transformed, just numerical values being manipulated. Further, though the preamble recites "[a] method of associating frame feature values with a plurality of frames of image data," the method steps merely calculate statistics and generate a numerical value from the statistics. Thus, we find no physical subject matter being transformed.
We also find that the method of claim 37 fails to produce a useful, concrete, and tangible result. Specifically, the result of claim 37 is a numerical representation of a quantity of a feature. However, a number is neither concrete nor tangible. Thus, claim 37 is an abstract idea that is nonstatutory under 35 U.S.C. § 101.
Appeal 2007-0759 (U.S. Application 10/177,732):
Appellant’s disclosed and claimed invention is completely silent as to what is used to solve the algorithms and mathematical equations. The Examiner mentions a computer, but we find no mention in the disclosure of the use of such a device to solve the algorithms and equations (Answer 3). Thus, the disclosure does not rule out the use of a pencil and paper to solve the algorithms and equations. Even if a computer was used by Appellant, we find that the solved result of the algorithms and equations is merely a solved mathematical result that is never applied to something to produce some kind of response or result.
Appeal 2007-1089 (U.S. Application 10/348,277):
This is probably one of the most controversial decisions from the BPAI in 2007. Despite finding that a computer process is involved, and despite the finding that the process was useful, the BPAI nevertheless found the claims directed to nonstatutory subject matter. In an odd twist of logic, the panel looked to the specification (which described the computers, software, and "media objects") and concluded that this alone described data structures per se. Stranger still, the use of the term "media objects" were interpreted as "nonfunctional descriptive material":
[W]e find the scope of the instant claimed system and associated components broadly encompasses software, and/or data structures per se. We further find the scope of the instant claimed media objects broadly encompasses nonfunctional descriptive material (i.e., pictures, photographs, music, sounds, text, e-mail, movies, video, messages, documents, slides, movie or video stills, streaming video and/or audio and/or any combination thereof) .
Also, according to this panel, the key to finding patentable subject matter (where a machine is not explicitly recited) lies with In re Schrader. Moreover, the death of the Freeman-Walter-Abele test (FWA) at the BPAI appears to have been greatly exaggerated:
Although the FWA test is no longer considered particularly probative in the context of computer-implemented process inventions in view of Diehr . . . the erosion of FWA provides no support for the position that a non-machine implemented process, not
involving any transformation, might be patentable. The answer to that question is still provided by Schrader, and that answer, so far, is negative. While AT&T indicated that Schrader is "unhelpful" because it did not reach the question whether a "useful, concrete, and tangible result" occurred, the reason that case did not need to reach that question was because it found that Schrader’s method claims were unpatentable for lack of any transformation. In addition, Schrader’s claims did not require machine-implementation, unlike AT&T’s claims.
You can sample these, and numerous other BPAI opinions on patentable subject matter here (link).