Life after Bilski.

• Author: Lemley, Mark A.
• Position: Symposium

INTRODUCTION I. THE RETURN OF MACHINE-OR-TRANSFORMATION A. The History of Patenting Software B. The Persistence of the Machine-or-Transformation Test C. The Perils of the Machine-or-Transformation Test II. THE THEORY OF SUBJECT MATTER RESTRICTIONS A. Prior Subject Matter Theories B. Abstract Ideas as Scope Limitation C. Distinguishing the Scope Limitations in [section] 101 and [section] 112 D. Scope Limitation in the Case Law 1. Historic cases 2. Modern Supreme Court cases III. A PROPER APPROACH TO IDENTIFYING CLAIMS THAT IMPERMISSIBLY COVER ABSTRACT IDEAS A. The Failure of the Machine-or-Transformation Test B. Identifying Abstract Ideas C. Overclaiming and the Decision Process D. Applying the Scope-Based Patentable Subject Matter Analysis CONCLUSION INTRODUCTION

In Bilski v. Kappos, the Supreme Court rejected calls to categorically exclude business methods--or any technology--from the scope of patent law. (1) It also rejected as the sole test of subject matter eligibility the Federal Circuit's deeply-flawed machine-or-transformation test, under which no process, and perhaps no invention of any type, is patentable unless it is tied to a particular machine or transforms an article to another state or thing. (2) Nonetheless, the Court held that the machine-or-transformation test still "is a useful and important clue, an investigative tool, for determining whether some claimed inventions" are patentable subject matter. (3) The result was a (narrow) victory for inventors, as well as for context-specific standards over formal rules.

Subsequent developments threaten to undo that win, however. Relying on the Court's "useful and important clue" language, the U.S. Patent and Trademark Office (PTO), patent litigants, and district courts have all continued to rely on the machine-or-transformation test in the wake of Bilski: no longer as the sole rule, but as a presumptive starting point that threatens to become effectively mandatory. Put simply, the problem is that no one understands what makes an idea "abstract," (4) and hence ineligible for patent protection, so decision-makers fall back on the one test that has been articulated.

In this Article, we suggest a new way to understand the exclusion of abstract ideas from patentable subject matter. No class of invention is inherently too abstract for patenting. Rather, the rule against patenting abstract ideas is best understood as an effort to prevent inventors from claiming their ideas too broadly. By requiring that patent claims be limited to a specific set of practical applications of an idea, the abstract ideas doctrine both makes the scope of the resulting patent clearer and leaves room for subsequent inventors to improve upon--and patent new applications of--the same basic principle.

Understanding the abstract ideas doctrine as an overclaiming test eliminates the constraints of the artificial machine-or-transformation rule, as well as the pointless effort to fit inventions into permissible and impermissible categories. It also helps understand some otherwise-inexplicable distinctions in the case law. A focus on overclaiming allows courts to focus on what really matters: whether the scope of the patentee's claims is commensurate with the invention's practical, real-world contribution, rather than asserting coverage over general ideas unmoored to a specific application. This inquiry, we suggest, is the touchstone of the abstract ideas analysis, and the way out of the post-Bilski confusion.

In Part I, we discuss the rise, fall, and apparent rebirth of the Federal Circuit's machine-or-transformation test, and explain why it is so problematic. In Part II, we offer a new way of thinking about abstract ideas, one focused on the scope of the patentee's claims rather than the category of the invention. In Part III, we apply our test to a variety of current cases and problems.

1. THE RETURN OF MACHINE-OR-TRANSFORMATION

   A. The History of Patenting Software

   The patentability of software and business methods has a long and tortured history. Put extremely briefly: The Supreme Court in the 1970s struggled with mathematical algorithm and software patents, first refusing to patent a mathematical algorithm, (5) but later holding that a process could be patented even though its novel elements included software so long as that software was tied to a particular machine. (6) After these cases, the Supreme Court got out of the business of patentable subject matter for nearly thirty years. (7)

   Through the 1980s and 1990s, courts gradually eroded the requirement that a software invention be tied to a particular machine. (8) Finally, in 1998, in State Street, the Federal Circuit eliminated any such requirement, holding that both software alone and business methods were patentable regardless of the form in which they were implemented, if the invention produced a "useful, concrete, and tangible result," even one expressed in numbers. (9)

   For a decade after 1998, patentable subject matter was effectively a dead letter. That changed dramatically in 2008 when the Federal Circuit decided In re Bilski en banc. The case involved a patent application on a method of hedging risk in the sale of commodities. The patent application also included dependent claims that limit the method to use in the energy industry and one dependent claim that offers a formula for calculating risk-adjusted prices. The Feder-Federal Circuit ruled, eleven to one, that Bilski's claimed method of hedging financial risk was unpatentable. (10)

   The Federal Circuit crafted what has come to be known as the machine-or-transformation test: "A claimed process is ... patent-eligible under [section] 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." (11)

   Moreover, even if a process meets this test, it is unpatentable if the machine or transformation is merely incidental extra-solution activity. (12) And an invention that preempts all use of a law of nature or algorithm, even in a particular field of use, is not patentable even if it would otherwise survive the test. Purporting to derive this test from Supreme Court precedent, the Federal Circuit displaced all prior tests for patentable subject matter, including State Street's "useful, concrete, and tangible result" analysis. (13)

   The Supreme Court granted certiorari and affirmed that Bilski's invention was unpatentable as an abstract idea. (14) However, the Court rejected the machine-or-transformation analysis as the exclusive test of patentability, (15) though all the Justices thought that the test was an "important clue" to patentability. (16) Alas, the Court's majority did not provide additional guidance on why Bilski's invention was unpatentable. (17)

   B. The Persistence of the Machine-or-Transformation Test

   Patent owners narrowly dodged a bullet in Bilski. No one except perhaps Bilski himself thought that he would get a patent. (18) However, four Justices would have held business methods unpatentable altogether, (19) and in the course of doing so would have substantially limited patentable subject matter. Thus, the fact that Justice Kennedy's opinion for the Court rejected strict application of the machine-or-transformation test and left open the prospect of patenting software and business methods led early commentators to declare it a victory for patent owners in those industries. (20)

   In fact, the PTO and the few district courts to have decided cases since Bilski have continued to apply the machine-or-transformation test. For example, in Ultramercial, LLC v. Hulu, LLC, the claims related to the process of displaying copyrighted media in exchange for watching certain advertising. (21) The district court held that "even though the machine or transformation is no longer the litmus test for patentability, the Court will use it here as a key indicator of patentability." (22)

   The PTO Board of Patent Appeals and Interferences (BPAI) has similarly focused on the machine-or-transformation test in analyzing post-Bilski claims. (23) In Ex parte Tse-Huong Choo, the BPAI rejected a claim that it found "can exist solely in software and data structures." (24) More importantly, the Board looked at a Beauregard claim (25) and found that "merely reciting data or instructions on a stored computer readable medium does not make a claim statutory [subject matter] under [section] 101. Similarly, merely placing instructions or code on a computer readable medium does not render claim 24 statutory." (26) Thus, the Board concluded that Bilski bars not just process but also article-of-manufacture claims to software standing alone. (27)

   The Board has been equally strict in applying the transformation step of the machine-or-transformation test to software. In Ex parte Heuer, the Board rejected a claim to a method for decoding a binary representation of a document. (28) The BPAI found that the claims failed both prongs of the machine-or-transformation test. It specifically looked at an "XML-based document," "a simplified schema," "a schema," "correction code," and "at least one decoder" and found that nothing in the specification indicating that they "must be hardware" and that there was therefore no machine. (29) On the transformation side, the BPAI found that "[t]he fact that Appellants' XML-based document is decoded using the claimed 'namespace' or the claimed 'simplified schema' fails to qualify as a transformation of an article of manufacture consistent with the second prong of the 'machine-or-transformation' test." (30) The latter ruling seems to be more restrictive than the Federal Circuit decision in In re Bilski, since it doesn't permit transformation of anything other than an article of manufacture, while the Federal Circuit had appeared willing in Bilski to allow some transformations of data to meet the machine-or-transformation test. (31)

   The PTO issued guidance in interpreting Bilski in July 2010...