The recently decided Federal Circuit case of CLS v. Alice Corp. was supposed to clarify what constitutes patentable subject matter in the world of software. Unfortunately, clarity here is in short supply. Alice went to court and was given a confusing trip through “Patent Wonderland.” Ten judges heard the case, and in 135 pages they gave seven separate opinions, concurring, dissenting and reflecting on one another’s statements. In some of the claims, there was even a 5-5 split as to whether they were patentable.
But there are some things we can learn from the decision. A majority of the court held that certain method claims of the patents at issue were directed to ineligible subject matter. So how should a software inventor best frame his or her patent claims? Let’s first look at the opinion that five of the 10 judges delivered.
Judges Lourie, Dyk, Prost, Reyna and Wallach began their analysis by characterizing the invention as a “computerized trading platform used for conducting financial transactions in which a third party settles obligations between a first and second party to eliminate settlement risk.” They proceeded to ask whether this subject matter was a “process, machine, manufacture, or composition of matter.” If not, it was unpatentable. If the invention fell into one of these four categories, they then asked whether the subject matter falls into the (patent ineligible) categories of “laws of nature, natural phenomena and abstract ideas.”
These five judges also emphasized that patents should not be allowed to “preempt the fundamental tools of discovery.” To do this, the claims must be analyzed to determine whether they “threaten to subsume the full scope of a fundamental concept,” and a determination must be made whether the claim limitations prevent the claim from “covering the concept’s every practical application.” Not surprisingly, they noted that there is no easy bright-line test to determine this, but felt that the device claims were not patent eligible because they merely recited an abstract idea.
Next, these same judges found that the method claims were also not patentable because they supposedly lacked any express language to define the computer’s participation. It appeared to these judges that the computer only “accelerated the calculations” that were previously done in any typical escrow arrangement (i.e., tracking obligations and performance of the parties). The judges also didn’t like claiming the software as a physical apparatus by describing it as a method carried out on a “computer readable storage medium”(i.e., the physical “thing” on which the software resides). The judges felt that this was merely an attempt to claim a method “in the guise of a device.” They also held that the system claims (which set forth a data storage unit and a computer – with description of how each operated) were little different from the (rejected) apparatus claims.
The judges said that any computer is a tangible machine made of wires, plastic and silicon, but “that is not the question.” They felt that the question to be asked instead was whether the claim represents something more than an abstract idea. The judges rejected the claims as being an abstract method carried out on a tangible machine.
This May 10 per curiam opinion represents the biggest threat to software patents to date. If one more judge would have concurred (thus making this a majority opinion), software patents would be in for a rough time. Also, by devising a two-part test in which you first look at the invention to see if it falls into a patentable category, and then see if it is an abstract idea (which should be rejected), this gives software patents two hoops to jump through.
The takeaway lessons for the software inventor are:
(1) Make your invention look as much like a machine as possible, and include as many functional details as possible as to how the individual software components actually operate.
(2) Never just claim the idea of doing something on a computer. Instead, if you claim a method, be sure to claim a way of doing something with a computer showing how the individual components of the computer perform their required functions.
(3) Ensure the claims add “significantly more” in terms of real limitations to an abstract idea. Therefore, write claims that would not pre-empt others from using tools for discovery. In doing this, don’t rely on hollow field-of-use limitations (for example, don’t claim the use of a method in only a particular industry).
Several more important separate opinions were delivered by the court, as follows.
Chief Judge Rader and Judges Linn, Moore and O’Malley stated that the claim must be considered as a whole to assess whether it is patentable subject matter – and to avoid dissecting the claims into old and new elements prior to analysis. They noted that it is not the breadth or narrowness of the abstract idea that mattered, but whether the claims cover every practical application of that abstract idea.
These judges emphasized that the key is to determine whether the claim ties an abstract idea to a specific way of doing something on a computer, or to a specific computer. If so, this could be patentable. However, they believed that claims that simply recite performing an abstract idea on a general purpose computer are not patentable. They asked: Is the computer itself part of the solution? Does the computer itself contain an important improvement in computer technology? In short – if the claim is tied to the computer in a meaningful way, such that the computer performs a meaningful role in the invention and such that the claim does not pre-empt the underlying abstract idea, then the claim is likely patentable.
In dissent, these four judges strongly protested the conclusions of the other five judges when they stated that failure to recognize the system claims as patentable would be “the death of hundreds of thousands of patents, including all business method, financial system, and software patents.” They felt that the system claims were directed to “tangible machine components with limited specialized functions programmed with detailed algorithms disclosed in the patent.” The computer is a “tangible item that you could put on your desk . . . it is not a method claim disguised as a machine.” They even stated that a computer does not become an abstract idea just because it is running software.
Importantly, these four judges thought that the system claims were patentable subject matter. This was the most pro-patent part in the collection of decisions. They felt that “the machine itself, with all its structural and functional limitations, would rarely, if ever, be an abstract idea.” They also noted that the use of a computer is not inherent in an escrow, so the claimed invention would not be obvious. However, on the method claims, these four judges split. Rader and Moore thought the method claims were not patentable. Linn and O’Malley believed the opposite.
The takeaway message here for software inventors is to always write device / apparatus claims, and never rely on method claims alone to protect their inventions.
Some of the other statements made in the various decisions should be mentioned, but are less relevant (in that they attracted little support from other judges). For example, Judge Newman stated that patent eligibility should not depend upon the form of the claim (i.e., that method and system claims should be treated the same way); and that experimental use of patented information should not be barred. He would hold all of the claims to cover patent-eligible subject matter, but remand to the district court to further consider patentability. Judges Linn and O’Malley stated that the method claims should be treated the same as the system claims. They also noted that Congress should consider shortening the life of software patents and requiring “functional claiming.” However, they gave no explanation as to how this would work in practice – e.g., how to determine whether something is in fact a software patent (since most inventions involve software nowadays).
In view of this strange decision split (with five judges holding at least some of the claims to be patentable and five judges stating that none of the claims are patentable), it is important for all parties to realize that software patents are presently sitting on a fence. Subject matter eligibility of software may depend as much on the composition of the court – or the identity of the patent examiner as it does on the claims in question. Therefore, until these issues are resolved, the objective for software inventors must be to write their patent claims to increase the odds they will be received favorably (i.e., making them look more like a physical machine and less like an abstract idea).
But what does the U.S. Patent and Trademark Office say about all this? Three days after the decision was released, the Patent Office officially told its examiners that there would be “no change” in examination procedures for software patent applications in view of Alice. The Patent Office stated that Alice affirmed the test for patent eligibility must be flexible and made on a case-by-case basis. The Patent Office also noted that the judges’ test for subject matter eligibility is different from other patentability concerns such as novelty and obviousness. Lastly, the Patent Office emphasized the judges’ conclusion that the claim must be looked at as a whole and that it must add “significantly more” to an abstract idea in terms of meaningful limitations to be eligible subject matter.
In other words, the debate as to when software is patentable continues.
To read the opinion, click here.