Software Patents Before Bilski
For more information on this subject, see Bitlaw's history of software patents in the United States.
Statutory (or "Patentable") Subject Matter. Under the U.S. Patent Act, not all inventions are patentable. Patent attorneys use the phrase "patentable subject matter" or "statutory subject matter" to describe that category of inventions that is eligible for patent protection. Unfortunately, the relevant statute language (35 U.S.C. 101, or "Section 101") is purposefully ambiguous. Rather than explicitly listing which types of inventions are patentable, the statute states that anyone who "invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent." The statute clearly does not mention computer software, but it also does not mention gene sequences, or new types of genetically modified bacteria. Nonetheless, these latter types of inventions have been determined to be statutory subject matter.
Because there is an ambiguity in the statute, its meaning must be interpreted by the courts. When analyzing this issue, the Supreme Court has stated that, when considering whether a new type of invention should be considered statutory subject matter, the presumption should be that "anything under the sun that is made by man" should qualify. This language originally came from Congress, which used this language to describe their intent when the patent statute was enacted in 1952. Although this intent shows that the patent statute should be interpreted broadly, we must remember that not everything that is discovered by humans was actually made by humans. Laws of nature and natural phenomenon are not made by anyone, but are rather discovered. Based on this rationale, courts have consistently stated that laws of nature, natural phenomenon, and abstract ideas cannot be considered statutory subject matter. Abstract mathematics is also generally considered unpatentable, either as an expression of a law of nature or as an abstract idea.
Software-Related Inventions as Statutory Subject Matter. The question that courts have struggled with is whether a computer software program should be considered like a mathematical formula and therefore unpatentable subject matter, or more like a modification to an electrical device and therefore eligible for patent protection. In the 1970s, the Supreme Court twice analyzed inventions including computer software, and both times held the inventions unpatentable. In the 1972 case of Gottschalk v. Benson, the Supreme Court considered whether a method that operated on a computer's "shift register" to convert binary-coded decimal numbers into true binary numbers was a patentable invention. The Court held that granting a patent for this algorithm on a shift register (the court defined the term "algorithm" as a procedure for solving a given type of mathematical problem) would pre-empt the entire mathematical algorithm. Since the court felt it inappropriate for any patent to preempt a mathematical formula, this invention was not statutory subject matter. In its next case to consider this issue, namely Parker v. Flook, the Supreme Court decided that a method for updating an "alarm limit" that was used to signal abnormal conditions in a catalytic conversion process was also unpatentable subject matter. The Court felt that since the only new element in Flook's invention was the mathematical formula used to calculate the alarm limit, the invention was not patentable subject matter even though the invention involved steps beyond mathematical calculations.
After these two cases, most observers felt that computer software would be treated like mathematical algorithms, and therefore would be considered unpatentable. This thinking changed in 1981 when the Supreme Court decided for the first time in Diamond v. Diehr that an invention containing computer software should be considered statutory subject matter. The invention in this case related to improved computer software that was used to operate a rubber curing process. The Court held that the invention should be considered an improved process for curing rubber and therefore patentable, even though the only new feature of this invention was the timing process controlled in software by a computer. The Diamond v. Diehr decision rejected the implication found in the Flook decision that one should determine where the new or novel aspect of the invention is found to determine whether an invention is patentable.
The Role of Claims in Evaluating the Patentability of Software. Once Diehr opened the door for some software patents, inventors and patent attorneys filed many more applications for software related inventions. By the early to mid 1990s, most patent attorneys considered software to be patentable, but only if the software was properly claimed. Patent claims specifically define the scope of the invention (see the patent application section of BitLaw for more information on claims). For instance, the court in In re Lowry held that while claims that recite a particular data structure are unpatentable subject matter, the same data structure would be patentable if it was claimed as a memory device that contained that data structure. Since all computer data structures must be stored in a memory to be of any use in the real world, this distinction did not limit the type of inventions that could be patentable, but instead informed patent attorneys about the proper way to draft claims for this type of invention. This form over substance approach continued in In re Beauregard. In this 1996 case, the patent office withdrew a statutory subject matter rejection during the appeal of a patent application where the claims described computer software residing on a tangible medium such as floppy diskettes. Even though the Federal Circuit never issued an opinion in this case, patent attorneys learned that while "pure software" patent claims might be considered unpatentable by the patent office, the patent office would accept those same claims if the software was described as residing on a tangible memory device. In memory of this abandoned appeal, patent attorneys still refer to patent claims that cover software stored on a computer readable media as "Beauregard Claims."
State Street Bank. The importance of carefully drafting patent claims to meet the current case law largely diminished with the 1998 decision of the Federal Circuit in State Street Bank (which is often mistakenly believed to be the beginning of computer software patents in the United States). In this case, the Federal Circuit made it clear that software was to be considered patentable as long as the software yields "a useful, concrete, and tangible result." This case explicitly stated that business methods can be considered patentable subject matter as long as they meet the other requirements for patentability. After the State Street Bank decision, many more patent applications were filed on software and business method inventions. In fact, due to the permissive language of State Street, some of Internet-related business method patent applications were filed containing claims that did not even mention the role of computers in the implementation of these inventions. In other cases, business method patent applications were filed on methods that did not involve computers at all, such as on tax avoidance schemes. Because of the negative publicity surrounding many of these patents, and the fact that many bad software patents were issued by the patent office, the Patent and Trademark Office has recently made it much more difficult for these patents to issue. For instance, it is not unusual for a patent applicant in the software field to wait four or more years before the patent office first examines their patent application. In addition, the rate of allowance by the patent office for these applications is rumored to be under ten percent, which is quite low considering that the over-all allowance rate for patent applications is more than fifty percent.
Bilski and the Federal Circuit
On October 30, 2008, the Federal Circuit came down with its In re Bilski decision. In this decision, the Federal Circuit rejected its holding in State Street Bank that software and business methods need only produce a "useful, concrete, and tangible result" in order to be patentable subject matter. However, the court refused to state that all business methods were unpatentable. Instead, the court replaced the State Street bank "useful, concrete, and tangible result" test with a new test that applies only to process/method claims, namely the "machine-or-transformation" test.
Bilski's invention related to a "method of managing the consumption of risk costs of a commodity sold." There were no apparatus claims (such as claims including the computer that runs the software or "Beauregard claims" that describe the software as being stored on a memory device) in Bilski's patent application. In addition, there were no claims for "a computerized process" (such as a computer implemented method). Consequently, the court did not directly rule on whether these types of claims were patentable. Instead, the majority ruling was limited to the patentability of processes where the process steps are not necessarily performed on a computer. After rejecting the State Street Bank test, the majority set forth a single test for determining the patentability of processes. This test holds that a process is patentable if "(1) it is tied to a particular machine or apparatus, or (2) it transforms a particular article into a different state or thing."
The majority opinion decided not to determine whether a software process operating on a general purpose computer is considered to be tied to a "particular machine or apparatus." In other words, the decision as to whether a general purpose computer is a "particular machine" was left for another day. If this test does hold that a general purpose computer is not a "particular machine," then many software implemented process patent claims could be invalid. In addition, the Federal Circuit did not explicitly determine how this new "machine-or-transformation" test applied to apparatus claims such as computer implemented software claims or Beauregard claims. After performing its own analysis on the Federal Circuit's In re Bilski decision, the U.S. Patent and Trademark Office determined (I think correctly) that In re Bilski is not directly relevant to these types of apparatus claims.
An analysis of the Federal Circuit decision reveals that it has several inherent flaws that led many observers to expect a reversal by the Supreme Court. One primary flaw is that it does not appear that the Federal Circuit correctly interpreted the Supreme Court's Gottschalk v. Benson, Parker v. Flook, and Diamond v. Diehr decisions. Each of these cases emphasized that the determination of whether an invention is patentable subject matter should focus on whether a particular "fundamental principal" (such as an abstract idea, a mathematical algorithm, or a law of nature) has been preempted by a patented claim. In contrast, under the Federal Circuit's In re Bilski decision, it is easy to contemplate scenarios where inventions that might preempt a law of nature would be patentable because the claim transforms an article into a different state or thing (such as a patent claim covering all nuclear fission), or where inventions that would have no risk of preempting a fundamental principal would fail to pass the Federal Circuit's test.
The Supreme Court's Analysis in Bilski v. Kappos
In June 2010, the U.S. Supreme Court in Bilski v. Kappos partially overturned the Federal Circuit's In re Bilski. The Supreme Court rejected the Federal Circuit's believe that the sole test for determining the patent eligibility of processes was the machine or transformation test. While the Supreme Court did not completely throw out this test, and in fact called the test useful and important, the Court held that this is not the sole test for determining patent eligibility for processes.
This decision is unusual in that there is no complete decision that has been adopted by the majority of the court. While a portion of the Kennedy opinion does constitute a five Justice majority, Justice Scalia refused to join two sections of this opinion. The portion of this opinion that did win the support of five justices held that the machine-or-transformation test is not the sole test for determining patent eligibility for processes, but is instead "a useful and important clue, an investigative tool, for determining whether some claimed inventions are processes under §101." (Section II-B-1). The majority opinion further stated that Bilski's invention is not patent eligible because the invention is preempting an abstract idea (Section III), referring back the the Supreme Court decisions in Gottschalk v. Benson and Parker v. Flook. This majority portion stated that granting a patent on Bilski's new method of hedging risks would preempt use of this concept in all fields, and effectively grant Bilski a monopoly in this abstract idea. The majority opinion also stated that the term "process" in § 101 should be interpreted broadly and therefore should be interpreted in light of §100(b) to include "methods." Since the "ordinary, contemporary, common meaning" of the term "methods" includes business methods, the majority believed that business methods should not be excluded from patent protection as a whole. Rather, these types of methods should be analyzed as to whether they preempt an abstract idea.
The Kennedy opinion loses its majority in sections II-B-2 and II-C-2. In II-B-2, Kennedy argues that the machine-or-transformation test is less important for certain types of inventions. Kennedy argues while the test may have been useful for inventions in the "industrial age," it "may not make sense" to require a judge to inquire as to whether a process meets the machine-or-transformation test for inventions in the "information age." As Kennedy and three other Justices state, "Section 101's terms suggest that new technologies may call for new inquiries."
In II-C-2, the opinion acknowledges that "some business methods patents raise special problems in terms of vagueness and suspect validity. "The court believed that "[i]f a high enough bar is not set when considering patent applications of this sort, patent examiners and courts could be flooded with claims that would put a chill on creative endeavor and dynamic change." It is clear, however, that the exclusion of these undesirable (yet undefined) business methods should not be attempted by defining the term "process" to exclude these applications. Instead, the court explicitly stated that it might be permissible to permanently exclude some subset of business methods from the patent system as attempts to patent an abstract idea.
In the concurrence written by Justice Stevens, four justices agreed with the decision that the Bilski invention was unpatentable. However, rather than rejecting the invention for preempting an abstract idea, these four justices would exclude all business methods from patent protection. Stevens believed that when properly interpreted, the protection provided under the patent statute to "processes" does not include any business methods whatsoever.
Justice Breyer wrote a second concurrence for the purpose of "highlight[ing] the substantial agreement among many Members of the Court on many of the fundamental issues of patent law raised by this case." Justice Breyer identified four points that are "consistent" with both the majority opinion and the Stevens concurrence:
- There is a limit to patentable subject matter. In particular, "phenomena of nature, mental processes, and abstract intellectual concepts" are not patentable.
- The machine-or-transformation test is "the clue" to the patentability of process claims.
- The Federal Circuit was wrong in holding that the machine-or-transformation test was the "sole test" for determining patentability for process claims.
- The Supreme Court is not resurrecting the "useful, concrete, and tangible result" test of the Federal Circuit's State Street Bank decision.
Where we stand now
After the Supreme Court has spoken in Bilski v. Kappos, some things have become clearer while others have become more murky. It is clear that the machine-or-transformation test of In re Bilski remains "a useful and important clue" and "an investigatory tool" for analyzing whether processes meet the patentable subject matter requirement of Section 101. It is also clear that this test is not definitive, as it is not to be considered the sole test for determining patent eligibility of processes.
Once crucial question is, if the machine-or-transformation test is an important investigatory tool, what exactly is being determined by the machine-or-transformation test. Before the Supreme Court's decision, it could be strongly argued that the Federal Circuit's machine-or-transformation test was being used to determine whether or not a method fell within the statutory definition of a "process." (See Diamond v. Diehr, which explained that Gottschalk v. Benson cited this machine-or-transformation test as an addition to other definitions of the word "process"). If the machine-or-transformation test was part of the definition of "process," patent attorneys could (and did) limit the test to process claims only. However, the role of the machine-or-transformation test is not so clear after the Supreme Court decided this case. The five Justice majority in Bilski v. Kappos makes it plain that the term "process" is to be interpreted very broadly. One can no longer say that the machine-or-transformation test is part of the statutory definition of a process while remaining consistent with the five Justice majority opinion in Bilski v. Kappos. Nonetheless, we know the test is "a useful and important clue" and "an investigatory tool" for analyzing whether processes subject matter eligible under Section 101.
Since the machine-or-transformation test can no longer be considered part of the definition of a process, I believe that the test must now be considered a test for determining whether processes are preempting an abstract idea. Therefore, as the law now stands, processes that fail the machine-or-transformation test are likely to be unpatentable for preempting an abstract idea, while processes that pass the test are likely to be patentable for they do not preempt an abstract idea. The results of this test constitute strong evidence either way, but arguments can likely be used to prove that the test fails in some particular case.
This interpretation of current law is consistent with the current stance of the U.S. Patent Office. In a Memorandum issued on the date of the Bilski v. Kappos decision, the PTO stated that the machine-or-transformation test should still be used to analyze whether processes pass the Section 101 eligibility requirements. Processes that fail the test should be rejected by the examiner "unless there is a clear indication that the method is not drawn to an abstract idea." The applicant would then have the opportunity to explain why the claimed method is not drawn to an abstract idea. Similarly, processes that pass the test would be considered patent-eligible under section 101 "unless there is a clear indication that the method is directed to an abstract idea." In effect, the machine-or-transformation test is now the default test for determining whether processes preempt an abstract idea at the U.S. Patent Office. The test effectively establishes a preliminary conclusion, which can be overcome with arguments to the contrary. In other words, the test establishes who has the burden of proof as to whether a process preempts an abstract idea or not.
Prior to the Supreme Court's decision, patent attorneys writing software related patent applications learned that the machine-or-transformation test of In re Bilski could be avoided by drafting claims as machines or manufactures (i.e., by claiming the software as being stored in a memory device--i.e., a Beauregard claim, or by claiming the software as operating on a computer system). Since the machine-or-transformation test was so closely tied to the statutory definition of a process, the results of that test on process claims were deemed irrelevant to apparatus claims. Unfortunately, now that the Supreme Court has re-emphasized the ability to reject method claims as preempting an abstract idea, it is feared that the results of the machine-or-transformation test will spill over into apparatus claims. The argument for using the results of the machine-or-transformation test on non-process claim would run as follows.
Since the process claims within a patent or patent application have failed the machine-or-transformation test, that process is presumably preempting an abstract idea. The Supreme Court has made it clear that machines and manufacture claims must also not preempt an abstract idea. Since your apparatus claims have nearly identical scope to the process claims, the decision as to whether the apparatus claims preempt an abstract idea must be the same as whether the process claims preempt that same idea. Therefore, since your process claims fail the machine-or-transformation test, then all of your claims of similar scope must presumably be unpatentable under Section 101 for preempting an abstract idea.