Today's Date: October 25, 2014 Search | Home | Contact | Services | Patent Attorney | Patent Search | Provisional Patent Application | Patent Application | Software Patent | Confidentiality Agreements

How to Draft Software Patent Claims After CLS Bank


Written by Dolly J. Krishnaswamy
Law Student, New York University School of Law
Project Manager/Law Clerk, Goldstein Patent Law
Posted: November 21, 2013 @ 7:55 am
Tell A Friend!

If you’re trying to draft a patent application for a computer-related invention, you’re basically left in a mess of opinions that the Federal Circuit refers to as CLS Bank Int’l v. Alice Corp. The decision, described by Judge Rader as a failure of his institution, left us wanting for just a tad –by a tad I actually mean a ton– more predictability. For now, we can only try to parse out why the opinions differed, take cues from our historical backdrop, and revise our prosecution strategy accordingly.

“A lot of times when we talk about patent eligibility we really need to know what the claims look like, so I thought we could spend four or five hours discussing this claim. Oh not enough time,” says Bruce Sunstein, founder of Sunstein, Kann, Murphy, and Timbers LLP at the 2013 AIPLA Annual Meeting. Fair point, Sunstein. So here’s the gist of CLS Bank. We have a computer readable medium claim in a format we’re accustomed to and that seems to be perfectly acceptable— program code for this program code for that. There are system claims that also would seem completely okay. Finally, there’s a method claim, which notably doesn’t include the word “computer.” Unfortunately for us, seeming completely okay was clearly not the test for patent eligibility. This was an en banc decision. The 10 available votes came out as follows:

  • Seven judges say method claims are not patent eligible.
  • Three judges say method claims are patent eligible.
  • Five judges say systems claims are not patent eligible.
  • Five judges say systems claims are patent eligible.

Judge Lourie writing with Dyk, Prost, Reyna, and Wallach says none of the claims are patent eligible. They reject the method claims for being too abstract. Then they say the storage and systems claims are just method claims in disguise, and so are also not patent eligible.

Judge Rader with Linn, Moore, and O’Malley says the systems claims are patent eligible, because the computer is performing a specific method. Rader with Moore again says that the method claims are patent ineligible, because they’re too abstract. On the other hand, Linn and O’Malley say all these claims are okay.

Newman says all the claims should be patent eligible, because they’re within statutory classes. “She as usual wrote a brilliant opinion well worth reading, cutting to the chase, and if everyone voted with Judge Newman our world would be simpler,” says Sunstein.

But since our world isn’t simple, Lourie’s opinion is the one to further analyze. He finds the systems claims patent ineligible, and believes the test for patent eligibility should be whether the claim in practical effect places an abstract idea at risk of preemption. If the claims do place preemption at risk, then we have to look at the inventive concept in the claim and decide if theres enough extra stuff in the claim to go beyond abstract idea. If there’s not enough extra stuff, the claims will not be patent eligible. “Now there’s a completely simple and easy to implement test,” jokes Sunstein.

The main authority for Lourie’s opinion is the Mayo v. Prometheus decision, which said laws of nature, natural phenomena, and abstract ideas are not patent eligible. Those exceptions came long before the Prometheus decision and even the Patent Act of 1952. To understand the Prometheus decision Lourie leans on, we need to go back and try to glean the still-standing principles that seem to come from these older cases.

Let’s start with the 1853 case O’Reilly v. Morse. “It was really astounding how many of the issues that were raised [in this case] are the same kind of issues that are going on today,” says Jerry Selinger, partner at Patterson & Sheridan LLP and current chair of AIPLA amicus committee. In this case, the claim at issue was basically for sending information at a distance using  “the motive power of the electric or galvanic current, which I call electro-magnetism…of which I claim to be the first inventor or discoverer.” (Oh, mighty Morse, could you have any more hubris?) There the court distinguished an earlier English decision, Neilson v. Harford. In the English case the purpose of the invention was to have a more efficient furnace by heating hot air in a receptacle before you blew it into a furnace. There we can see that the abstract principle was that hot air better transfers heat than cold air. The court pondered whether putting a receptacle between the fan and furnace was too abstract or whether that was in fact an application of the idea. It was upheld on the grounds that the invention was a machine embodying a principle and not just a principle alone. In contrast, the court in Morse said claiming the use of electromagnetism for communication was too abstract. The point to take away — other than the abstract idea exception itself– is that fact-specific assessment is not a recent invention of the Federal Circuit or the Supreme Court.  Moore brings attention to that lesson in her opinion in CLS Bank. She cites the Morse case but says the abstract idea exception is being given too much scope today. If you keep stretching the exception, we would abolish all business method, financial system, and software patents entirely.

More than a century after Morse, and after the Patent Act of 1952, the Supreme Court revisited patentable subject matter again in Gottschalk v. Benson. The invention in this case was a method of converting binary-coded decimal into pure binary numbers on a general-purpose computer. Just a reminder: we’re working from the backdrop that the discovery of a phenomenon of nature is not patentable, but an application of a law of nature to a new useful end is patentable. Also, transformation or reduction of article to different state or thing is a clue to patentability of a process claim that does not include particular machines. Benson later went on to clarify that the process patent does not have to be tied to particular machine or apparatus or operate to change articles or materials. Given these principles, Douglas with 5 other justices made a fact-specific decision saying the claims were ineligible because granting a patent in this case would be like granting a patent on the algorithm itself since there was no practical application except in a digital computer. Lourie followed suit in CLS Bank; he used Benson to find the fact that the system claims occurred in a computer insufficient to make them more than abstract.

Six years after Benson, in 1978, the court in Parker v. Flook reiterated that the discovery of a natural phenomenon alone cannot be patented. The invention provided a way to figure out the limit numbers between which the catalytic conversion of hydrocarbons can occur; once outside the range an alarm would sound. The court said that limiting the formula to the catalytic conversion of hydrocarbons was not enough to make the claim patent eligible. The court went out of its way to say any “competent draftsman could attach some form of post-solution activity to almost any mathematical formula.” In other words, they’re saying you can’t just add some non-inventive stuff onto the end of a mathematical formula to make the claim patentable. “I think what that says is you may have achieved ‘transformation’ but it may not matter.. [the court] never said nor implied that the transformation test trumps the law of nature exclusion,” says Selinger. That concept is an important line of demarkation you see between the opinions of CLS Bank.

Just a few years after Flook, the court in Diamond v. Diehr held that a process for molding rubber products by using the Arrhenius equation was patent eligible on the grounds that claiming a physical process that used a mathematical formula was not the same as just claiming the abstract formula. (And then all the software developers got a glimmer of hope.) Rader relies on the logic in Diehr to support his opinion in CLS Bank, saying that the claim there is for when a computer performs a specific application of a principle.

So then we –finally– come to Prometheus. Here we have a claim for a method to figure out how to adjust dosage. The doctor administers a drug to a subject who has a gastrointestinal disorder by giving them a pill. Then, the doctor determines the level of a metabolite by drawing blood. These steps seem concrete enough. Then if the measurement of metabolite is below some threshold he indicates a need to increase drug dosage to increase efficiency, and if its above a threshold then he identifies a need to decrease the dosage to avoid unsafe side effects. A unanimous decision by Breyer said the correlation between a specific metabolite level and the dosing “sweet spot” was a law of nature, and therefore not patent eligible. The court was looking for something else in the claim to go beyond just claiming the law of nature– an inventive concept. However, according to the court the administering steps refer simply to the relevant audience: doctors, who treat patents; the wherein clause simply tells the doctor about natural laws,  and the determining step tells the doctor to engage in well understood routine, conventional activity. The court slips in one sentence about a new drug or a new way of using a drug being safe from this inventive concept analysis. This tells us two things: 1) if we’re in the law of nature realm, we’re going to have to look outside that law for an inventive concept; and 2) as put by Courtenay Brinckerhoff, partner at Foley & Lardner LLP, “In law school we learned bad cases can make bad law, and here we have perhaps an example of bad claims making bad law.”



Have a software innovation to patent?
Experience with software patents and related technologies.
CLICK HERE for MORE INFO or e-mail Gene Quinn


As we know, five judges in CLS Bank would use Prometheus to find the claims were patent ineligible. Alice Corp’s claims are for methods of reducing fraud and non-payment by effecting trades through a third party intermediary who makes sure both parties can pay before allowing the trade to happen. The court found that the method was an abstract idea, because it is a “disembodied concept . . . a basic building block of human ingenuity, untethered from any real-world application.” As for tying the methods to an application in a computer, Lourie says the systems claims also don’t provide any significant inventive concept in addition to that abstract idea. (Here’s where Rader chooses to rely on Diehr to approve the systems claims.)

Wait but isn’t this idea articulated by Lourie about looking for an inventive concept old hat? “That’s the stuff that was before the Patent Act of 1952 right? I agree that all this is blarney,” says Sunstein. The Patent Act promoted the idea that there is no gist of invention in patent claiming. Even §103 was revised to steer away from requiring this “flash of genius” inventiveness.

We’re likely to hear more from the Supreme Court on the subject. “But they’re the ones who brought us Mayo v. Prometheus, and the problem with Mayo v. Prometheus in my view is there no articulated standard for evaluating anything,” says Sunstein.

We’ve got a couple cases following CLS Bank that give us clues as to what a computer-related claim should look like post-CLS Bank. In the Ultramercial v. Hulu case, Rader and Lourie are surprisingly on the same side. The patent covers a method relating to a user seeing an advertisement before getting exposure to his desired media content. A point that I’ll circle back to in a minute is that there were 11 steps recited in the method claims in Ultramercial. As expected, Rader writes an opinion saying this stuff goes on in a computer so we find it’s patent eligible– think Diehr/CLS Bank logic. What’s interesting is Lourie writes a concurring opinion, using as precedent his oh-so-decisive plurality opinion in CLS Bank. He found that unlike in CLS Bank where intermediation was too abstract a concept and the claims added nothing inventive, in this case the limitations represent significantly more than the underlying abstract idea of using advertising.

The Accenture case is the most recent in the string. This patent was for systems and method claims relating a computer program for handling insurance-related tasks. The interesting thing here is the method claim ineligibility determination by the federal court was not appealed. So Judge Lourie writing for the majority says the system claims on appeal are just like the method claims, and since the method claims were bad, the system claims are ineligible too. He goes on to say the system claims would be bad independently too, but that doesn’t need to be analyzed.

This means that including different types of claims may not save you in terms of eligibility, because Lourie will just say they’re just like the method claims. Rader in the dissent says you can look at the abstract idea behind something to invalidate any invention, and you can’t just go hunting for abstractions by ignoring the specific stuff that is in the claim. Indirectly, this is a criticism not just of Judge Lourie in CLS Bank, but also of the Supreme Court in Mayo v. Prometheus. “So if you want to know how your claim is going to be evaluated, just find out which judges are on your panel, and you have your answer. See how they voted in CLS Bank,” says Sunstein.

It seems like if you get past Judge Lourie you optimize your chances, so the real question is how do you get past Judge Lourie? Well Accenture had an extremely long disclosure for the claims, so length is clearly not enough on its own. In the claims in Accenture all the limitations were in the last element compared to Ultramercial where there were 11 steps. This suggests that the rule is to atomize your structures. In other words, if you want to get past Judge Lourie instead of putting all the good stuff in the last claim like in Accenture, spread the limitations throughout the claims because then it’ll be harder for him to abstract. “It’s a terrible lesson, because it goes against the court telling us it shouldn’t be up to a clever claims drafter, but thats what it’s all about isn’t it? That’s how we earn our living,” says Sunstein.

There’s a petition for writ of cert in the CLS Bank and Ultramercial cases so hopefully we’ll have more clarification on the matter soon.


About the Author

Dolly J. Krishnaswamy is currently a law student at NYU and holds a position as the vice chair of the communications committee in the ABA-IPL. She has previously worked as a science journalist in for Science magazine, had jobs in the tech community in New York City, and spent her undergrad studying genetics at Emory University.


8 comments
Leave a comment »

  1. The Morse decision was fine: essentially it was an undue breadth type of claim, or you can say the claim was so broad it read on lots of stuff that was not enabled or for which there was no written description. Notice that these claim-killers can all be invoked without entering the murky waters of section 101.

    The real mischief was by Justice Douglas in the Gottschalk case. Essentially, he said that although the claim is limited to use of the algorithm on a computer, since nobody would use the algorithm except by using a computer, the claim preempts the entire scope of the algorithm. Of course, Douglas was the precursor to Lemley and that bunch in thinking NOTHING is patentable, but still, what a bunch of goof-ball sophistry!

  2. “The real mischief was by Justice Douglas in the Gottschalk case. Essentially, he said that although the claim is limited to use of the algorithm on a computer, since nobody would use the algorithm except by using a computer, the claim preempts the entire scope of the algorithm.”

    What do you propose is wrong with that statement or the reasoning underlying it?

  3. ” If you keep stretching the exception, we would abolish all business method, financial system, and software patents entirely.”

    I think that’s the reason why some people press for the application of it in just such a way.

    “Just a few years after Flook, the court in Diamond v. Diehr held that a process for molding rubber products by using the Arrhenius equation was patent eligible on the grounds that claiming a physical process that used a mathematical formula was not the same as just claiming the abstract formula. (And then all the software developers got a glimmer of hope.)”

    Very true.

    I don’t know if I agree with Courtnay though, the claims are just fine for what they specifically wished to cover with them. Had they have wished to cover something else, then sure, then they’re bad claims.

    “In other words, if you want to get past Judge Lourie instead of putting all the good stuff in the last claim like in Accenture, spread the limitations throughout the claims because then it’ll be harder for him to abstract”

    Exactly, just hide the abstraction from Judge Lourie so he can’t find it by using confusing language and spreading things out. Never mind any federal regulations against such hiding. It’s a good strategy for now that will work on him until he becomes sly to it.

  4. “If you keep stretching the exception, we would abolish all business method, financial system, and software patents entirely.”

    Yes, that was exactly what I was trying to do when I submitted an amicus curiae to the Federal Appeals Court for the Federal Circuit in CLS v Alice.

    “Douglas with 5 other justices made a fact-specific decision saying the claims were ineligible because granting a patent in this case would be like granting a patent on the algorithm itself”

    Yes, I tried to generalize this idea in my amicus curiae because all software consists of algorithms and nothing but algorithms. Therefore no software is patentable.

    ———————————
    Steve Stites

  5. “The real mischief was by Justice Douglas in the Gottschalk case. Essentially, he said that although the claim is limited to use of the algorithm on a computer, since nobody would use the algorithm except by using a computer, the claim preempts the entire scope of the algorithm.”

    What do you propose is wrong with that statement or the reasoning underlying it?
    The algorithm could be carried out using pencil and paper or chalk on a chalkboard or the like. In fact, as I recall even Justice Douglas’s opinion recited the algorithm. The practical benefit of doing it on paper may not be apparent, but it can still be done, so the entire scope of the algorithm is NOT preempted.

  6. Contrast with the (superseding) case of Diehr (also a use of an algorithm case) and one can plainly see that Benson, by itself, is simply not the last (or totla) word on the law.

  7. Glad to see this piece sparked some discussion between @AmericanCowboy and @JohnSmith. @SteveStites, glad to hear your contributions as well. I definitely think this is an area where people find themselves wanting clarity and predictability– whether it be for or against the patent eligibility of software. @Anon, yes Benson was definitely only one part of a very messy case history. Thanks for commenting!

  8. American – “The algorithm could be carried out using pencil and paper or chalk on a chalkboard or the like. In fact, as I recall even Justice Douglas’s opinion recited the algorithm. The practical benefit of doing it on paper may not be apparent, but it can still be done, so the entire scope of the algorithm is NOT preempted”

    I’m familiar with the argument you put forward, specifically that non-substantial uses, such as carrying out the algorithm on paper, are uses which are not preempt. Therefore, you conclude that all uses are not preempt. If it comforts you then I’m sure the courts will make themselves clear that they’re talking about all substantial uses/applications of the algorithm.

    Ms. Krishnaswamy – I’ve seen very little in the way of cases that aren’t very predictable if a judicial exception is indeed at issue. There are admittedly a few cases that are pretty tough and might go either way, but for the most part I haven’t seen many.