CLE: How to Draft Software Claims under Bilski

In Bilski, the Federal Circuit laid down the law of subject matter eligibility under Section 101 of the Patent Act. To be patent eligible, a claimed process must either: (1) be tied to a particular machine or apparatus, or (2) transform a particular article into a different state or thing. In all likelihood, claim limitations focused at general machines such as a "computer" will not be considered tying to "particular machine." Bilski did not, however, decide that question – Bilski's claims were admittedly divorced from any particular machine. On the transformation side, the court provided the example from Abele where graphically displaying "X-ray attenuation data produced in a two dimensional field by a computed tomography scanner" was sufficient transformation. Bilski's claims were essentially knocked-out on the new mental steps doctrine – a process where all the claimed steps "may be performed entirely in the human mind is obviously not tied to any machine and does not transform any article into a different state or thing."

Two additional points: Although the Federal Circuits points to the machine-or-transformation test as the sole test for Section 101 patentability, the court approved two ‘corollaries.’  First, field-of-use limitations continue to be “generally insufficient to render an otherwise ineligible process claim patent-eligible.” Second, "insignificant postsolution activity will not transform an unpatentable principle into a patentable process." (quoting Diehr).  Thus, merely reciting a specific machine or particular transformation will not bring a claim into the realm of patentable subject matter unless the recitation is more than mere insignificant postolution or extra-solution activity.

Going forward, I do not believe that these limitations will have a significant impact on a skilled practitioner's ability to patent software innovations. However, I would like community input on how you might properly claim computer software in a way that avoids § 101 rejections?

Comments

[curious]

Malcolm, we are waiting for your proposed claims.

[UTEX]

In discussing the patent inellible claims of Abele, the Bilski court noted that, "The claim did not specify any particular type or nature of data; nor did it specify how or from where the data was obtained or what the data represented." The court also clarified the specific holding in Abele by saying that "the electronic transformation of the data itself into a visual depiction in Abele was sufficient; the claim was not required to involve any transformation of the underlying physical object that the data represented." I think these two statements from the court provide, at least, an interim roadmap for how real software inventions (not ones like a computer-implemented method for making money) need to be drafted. First, characterize your data or information including, where applicable, where it came from, what it represents, etc. Second, provide a physically perceived transformation of that data. For example, many software inventions display a resulting feature or state or whatever to the user, it could also cause transmission of some kind of information to another process or trigger some other physically perceivable result. It appears that in order to get over the examples of patent inelligible processes is to claim something other than abstract data and provide some physically perceivable "transformation" such as putting something on a display that represents the processed data.

[Kevin Grierson]

The CAFC may not have spoken on the general purpose computer issue, but the BPAI, in Ex Parte Langemyr, expressly held that a general purpose computer won't cut it. So, if you want your claims allowed now, folks, you'll have to find a more specific machine, or use the transformation arm of the analysis.

I must admit that the court's example of the x-ray data display leaves me puzzled. Are they saying that tying a program to a particular type of data makes the program transformational? Or was it the fact that the data was a representation of a real world object? If it's the latter, it would be difficult to see how most financial software could be claimed in a way that satisfies the Bilski analysis. On the other hand, if a particular type of data is all that's required, that seems kind of counterintuitive, unless it would not be obvious to apply the program to that kind of data.

[Dennis]

UTEX & KEVIN Good posts.

Regarding Transformation, it appears that the transformed article needs to at least be “representative of physical objects or substances.” Notably, transformation of “legal obligations or relationships, business risks, or other such abstractions” do not qualify as ‘transformations’ under the new test.

[Patent Attorney]

What does this do to Beuregard claims e.g.:

computer-readable medium claims (comprising instructions ...)

computer-readable means claims

and other types of software claims? (E.g. Claiming a method as a software system)

Especially considering the fact that many software applications are shifting from being on CD-Roms, floppies, disks, etc., to being distributed and downloadable on the Internet, on websites and even mobile phones or related devices?

[Kevin]

I have had several Beauregard claims rejected based on Bilski (prior to the CAFC's decision), as have other members of NAPP (National Association of Patent Practitioners). Whether the CAFC would agree, I can't say, although their language about tying processes to a physical step not being sufficient seems to me to indicate that just reciting that the instructions are in a computer readable medium will not suffice.

[JK]

can any one define what is "STATE" in the area of "trasformation"?

[cheesepep]

So stating "A computer-implemented method for...."

and then stating "receiving data and transforming data into...." is OK?

[SF]

"What does this do to Beuregard claims"

nothing because they are apparatus claims and Bilski is limited to process claims

[non-senior counsel at major corp]

Proposed claim and questions

A method comprising:
(a) representing a specific original physical reality with specific original data
(b) electronically transforming the specific original data into specific transformed data (b’) using a novel and nonobvious process
(c) representing the specific transformed data (c’) as a specific transformed physical reality (c’’) in the form of a visual depiction

Apologies for the clumsy language.

My questions:
1)Do I need c’’ for 101?
2)Do I need c’ for 101?
3)Do I need b’ for 102/103 or can I derive novelty and nonobviousness from other elements?
4)Is a “computer simulation” (e.g. CAE, FEA, FSI, CFD) a transformation?

Thanks for any thoughts.

[Budge]

The issue of what is a "general purpose ccomputer" has been decided in the past by WMS Gaming. I belive under Bilski, a computer that can perform an algorithm would satisfy 101.

[breadcrumbs]

Lots of scroll-down-and-hit-Next (DC can you fix this?), but Kevin, this post on another Patently-O thread goes directly to your question:

Haven't courts already denied this line of logic from the PTO (general purpose computer vs. particular computer)? as cited in In re Lowry, In re Bernhart and Fetterer states:
There is one further rationale used by both the board and the examiner, namely, that the provision of new signals to be stored by the computer does not make it a new machine, i.e. it is structurally the same, no matter how new, useful and unobvious the result. . . . To this question we say that if a machine is programmed in a certain new and unobvious way, it is physically different from the machine without that program; its memory elements are differently arranged. The fact that these physical changes are invisible to the eye should not tempt us to conclude that the machine has not been changed. Bernhart, 417 F.2d at 1400.

Posted by: Patent Agent | Oct 30, 2008 at 12:45 PM
http://www.patentlyo.com/patent/2008/07/the-death-of-go/comments/page/11/#comments

I believe the courts outrank the BPAI. Of course, the BPAI may need to be reminded of what the law actually is.

[dblagent007]

Many people seem to think that software will fall under the machine prong of the court's new test. However, the court explained Benson as follows (from pages 12-13 of the opinion):

"Interestingly, Benson presents a difficult case under its own test in that the claimed process operated on a machine, a digital computer, but was still held to be ineligible subject matter. However, in Benson, the limitations tying the process to a computer were not actually limiting because the fundamental principle at issue, a particular algorithm, had no utility other than operating on a digital computer. Benson, 409 U.S. at 71-72. Thus, the claim's tie to a digital computer did not reduce the pre-emptive footprint of the claim since all uses of the algorithm were still covered by the claim. "

The court seems to be saying that any process that has no utility other than operating on a computer does not qualify under the first prong. Most software processes do not have any utility outside of operating on a computer, which means that software claims are likely to fall outside of the first prong.

Strangely, the court seems to say that software may fall under the second prong so long as it is limited to a specific type of data that represents physical and tangible objects.

[I haven't a clue]

What about in financials if you created some great hedging algorithm and based it on a real world representation of value say a commodity like gold or pork bellies and "transformed" those into another real world representation of value such as barrels of oil. I know commodity markets you are not actually buying a bushel of wheat to deliver to your house but maybe it could be enough. I guess this could create "straw-man" arguments and form-over-substance type arguments.

[I haven't a clue]

What about in financials if you created some great hedging algorithm and based it on a real world representation of value say a commodity like gold or pork bellies and "transformed" those into another real world representation of value such as barrels of oil. I know commodity markets you are not actually buying a bushel of wheat to deliver to your house but maybe it could be enough. I guess this could create "straw-man" arguments and form-over-substance type arguments.

[sr]

"Lots of scroll-down-and-hit-Next (DC can you fix this?)"

Amen to that.

[SF]

"Many people seem to think that software will fall under the machine prong of the court's new test."

It depends. Do you mean (1) software processes actually being performed or (2) software embodied in an apparatus, like a computer-readable medium? Bearegard claims are not computer-process claims, but rather apparatus claims (a medium). As long as you haven't defined the medium to be a signal, you're fine.

[UTEX]

dblagent007 - I don't agree with your conclusions regarding the Bilski court's discussion of Benson. The claims at issue in Benson were written purely to translating the BCD to binary numbers without any application or context. Thus, the claim was purely the steps of converting one piece of abstract data into a second piece of abstract data. I believe that is the definition of an algorithm. Regardless of whether or not that algorithm is performed on a computer doesn't change the fact that the claim is defining the scope of an algorithm, which is not patentable. Because the algorithm had no other utility than operating on a digital computer, the effect of simply tying it to a digital computer was an attempt to pre-empt all uses of that algorithm.

Most real software related inventions don't just deal with the processing of abstract data for the purpose of performing a cool algorithm. Most real software related inventions process real data that means something and performs some kind of signaling, transmission, or display of processed results to the user. The court implied in its discussion of the Abele case that those types of claims would be patent eligible. It seems like many of the "death to all software patent" folks posting here are really trying to stretch Bilski into calling into doubt all software patents, when, in fact, Bilski not only doesn't do that, but even seriously calls into question the PTO's and BPAI's overreaching conclusions regarding the general purpose computer. I believe that the Bilski court's statement that, "the claimed process operated on a machine, a digital computer, but was still held to be ineligible subject matter" shows that tying software-related inventions to a gp computer will be sufficient in many cases, unless the claims simply address processing of abstract information for performing some kind of algorithm without any context or purpose.

Breadcrumbs is right on point with this that, according to prior caselaw, a general purpose computer ceases to be a general purpose computer when programmed to perform the tasks of a particular program. So, even though a software related application is tied to a gp computer, when the software is executed by that gp computer it becomes a special purpose computer for performing the claimed steps.

[I did it]

Thanks, UTEX, for a refreshing dose of intelligent and articulate commentary.

However, Malcolm is still a poopyhead.

[3L mudder]

UTEX, I'm reading this as a line between Library code and Application code. Library code meaning general-purpose routines, ("compress this file"), vs. Application code, ("read these glucose levels from a file, compress them using LZW, calculate a checksum of the result, then if it is an even number, send a signal to an alarm.") [Purposefully silly example.]
The library code lacks the limitation to a general principle, compression in the abstract.
As an aside, to everyone blaming the courts, I think the problem is a huge gap in legislative activity. The judges are just struggling to deal with a fundamental problem: computer software is ultimately just a mathematical equation, and no amount of sophistry can solve that.

[Jim]

Two points from the opinion:

(1) For a machine implementation, that issue was not decided by the CAFC; and

(2) For a transformation, it must be central to the purpose of the claimed process and the electronic transformation of the data itself into a visual depiction was sufficient in Abele.

Fundamentally, one can maybe look to the EP's handling of software claims related to "technical effect" that goes beyond the normal physical interactions between the program and the computer, as a guide. See, http://www.epo.org/patents/law/legal-texts/html/guiex/e/c_iv_2_3_6.htm

Maybe?

[Jim]

"computer software is ultimately just a mathematical equation."

@3L Mudder: and so are physics, chemistry, and biology. Heck, even music and architecture too. Nature is mathematics in its purist form. It's not that simple.

But if you think it is, then you are much smarter than I, and I need to find a new profession.

[David Stein]

3L Mudder: The argument has often been raised that software = mathematics, because it's just a "conceptual arrangement" of instructions. Such arguers sidestep the fact that software is physically represented as data in (physical) memory connected to (physical) computing components that (electronically, i.e., physically) perform the operation.

But by the same token, chemistry also = mathematics. The invention is really a description of a composition that is expected to react according to the laws of chemistry... which are just applied laws of physics... which are applied models described via mathematics. Any patented chemistry invention is a "conceptual arrangement" of atoms.

Similar arguments can be applied to the mechanical and electrical arts. The common theme - in ALL of these cases - is that the patented invention is still an applied, useful concept, which MAY be implemented in the physical world. That's a basic tenet of patent law, and the software arts do not differ from the others in this regard.

- David Stein

[David Stein]

To everyone:

Even today, while the shock waves of Bilski are still reverberating, we need to start reaching consensus on three issues:

(1) What the CAFC's majority opinion (and SCOTUS precedent) actually mean;

(2) How the USPTO examining corps and BPAI are likely to react; and

(3) How the patent bar should respond - both in terms of drafting cases today, and in terms of options for addressing (1) and (2).

I am expecting PLI, AIPLA, NAPP, etc. to serve as focal points.

Just - PLEASE - let's try to coalesce as a body of practitioners ASAP. I have clients to advise, and while I can definitely provide counseling based on my own answers to these three questions, I'd feel much more comfortable with some backing by trusted sources.

- David Stein

[3L mudder]

And here I thought a theory of everything remained to be found, but it's right here, on this blog. Seriously though, we don't know the equations that transform a physical object into pure math, but we suspect it is possible and people are working on models. Software, however, is a different beast -- in my opinion.
But David is right that this isn't practically helpful. Unless this connection to physicality is the heart of what the court was struggling with as its limiting principle.
But I respectfully accept that others see object code in RAM [which isn't really a process claim, is it?] as distinct from a description of the process of computer software.
Apologies from a student. Carry on.

[JJL]

I just second everything UTEX said. Amen brother.

[dblagent007]

UTEX, I hope you are right. However, the Bilski court, when discussing Benson, didn't make any distinction between the Benson algorithm processing "abstract data" vs "real data." Nor did it refer to the Benson algorithm as lacking application and context.

Instead the court said that it was ineligible patent subject matter because its only utility was to operate on a computer. Thus, limiting the claim to a computer was not really a limitation at all according to the court.

Keep in mind that some of the discussion preceding the court's description of Benson refered to processes that were patentable because certain steps were real world steps while other steps were performed by a computer. Adding all that together gave me the impression that a process that operates solely on a computer does not satisfy the machine prong of the test.

[Kevin R]

Kevin Grierson said: "I must admit that the court's example of the x-ray data display leaves me puzzled. Are they saying that tying a program to a particular type of data makes the program transformational? Or was it the fact that the data was a representation of a real world object?"

Kev - (can I call you Kev?), I have been scratching my head about this as well. Here is what the opinion 'clarifies' -

"This [x-ray] data clearly represented physical and tangible objects, namely the structure of bones, organs, and other body tissues. Thus, the transformation of that raw data into a particular visual depiction of a physical object on a display was sufficient to render that more narrowly-claimed process patent-eligible. . . . [T]he claim was not required to involve any transformation of the underlying physical object that the data represented. . . . So long as the claimed process is limited to a practical application of a fundamental principle to transform specific data, and the claim is limited to a visual depiction that represents specific physical objects or substances, there is no danger that the scope of the claim would wholly pre-empt all uses of the principle."

It seems like the majority is saying:

1. If a claim recites a method for displaying data wherein the data represents a “physical and tangible” object, then displaying that data on a computer screen will satisfy the transformation test. [What about claims for simulation software, e.g., MATLAB that displays a representation of a hypothetical (i.e., imaginary) “physical and tangible” object? Good question]

2. Transformation of data alone may satisfy the transformation test, provided that the pre-transformed data represents something “physical and tangible.”

3. Here is a tip for improving the odds that we will not shoot down your method claims: please endeavor to “more narrowly-claim[ the] process” and see that “the claimed process is limited to a practical application of a fundamental principle.” [In other words, if your client is considering a patent for a entirely novel method for transforming specific data, e.g., an improved method for data compression, you might want to consider the trade secret alternative, especially if it is NOT “limited to a visual depiction that represents specific physical objects or substances”]

AND

4. Hell hath no fury like a court considering a “claim [that] would wholly pre-empt all uses of the [fundamental] principle.”

[hinks]

Ex parte Langemyr provides insight into the PTO's current position on Bearegard claims, in asserting that "[s]imply placing instructions on a computer readable medium, wherein the instructions are designed to perform mere manipulations of abstract ideas, should not convert an otherwise non-statutory method into patentable subject matter." The BPAI appears to be taking the position that computer-readable medium claims that recite software instructions in the body, without more, are non-statutory.

Further, the BPAI appears to return to an argument against the patentability of software which was repeated to me years ago, and goes "[a]lthough the instructions, when executed in a computer, may cause a transformation of the computer, the step of executing the instructions in a computer is not claimed here. In other words, the claim is not directed to a computer or machine loaded with and/or executing the software."

Thus, perhaps practitioners now need to recite the loading of the software on a computer or the executing of the software on a computer to make the claim statutory, essentially turning your 'apparatus' claim back into a process claim. Although this approach may not pass muster under the machine test (general purpose computer), it might under the transformation test. However, how you would successfully assert such a claim on a direct competitor (e.g., one who develops and sells software), beyond arguing that they surely must have infringed in development test situations, is not yet clear to me.

[sb79]

Before Bilski, the 101 standard as applied to computer inventions was vague, susceptible to abuse by the patent office, and filled with wishy-washy language and concepts. It is so great to see that the CAFC, in it's en bancable wisdom, addressed all these issues with that old standard.

[Kevin R]

Worth reading:

http://www.patenthawk.com/blog/2008/10/software_patents.html#more

[Leopold Bloom]

"Worth reading:
http://www.patenthawk.com/blog/2008/10/software_patents.html#more ."

If you take advice on patent law from someone who is neither an agent nor an attorney, and who appears to be treading perilously close to the boundaries of the unauthorized practice of law, then you get what you deserve.

In my opinion, that is.

[Leopold Bloom]

"Ex parte Langemyr provides insight into the PTO's current position on Bearegard claims, in asserting that '[s]imply placing instructions on a computer readable medium, wherein the instructions are designed to perform mere manipulations of abstract ideas, should not convert an otherwise non-statutory method into patentable subject matter.' The BPAI appears to be taking the position that computer-readable medium claims that recite software instructions in the body, without more, are non-statutory."

No, if your quotation is correct then the BPAI is taking the (legally correct) position that a software-readable medium claim is not patentable if the program instructions are directed only to manipulation of abstract ideas. This is completely consistent with Benson.

"Further, the BPAI appears to return to an argument against the patentability of software which was repeated to me years ago, and goes '[a]though the instructions, when executed in a computer, may cause a transformation of the computer, the step of executing the instructions in a computer is not claimed here. In other words, the claim is not directed to a computer or machine loaded with and/or executing the software.'

That statement may have been true then, and might be true now, but is completely irrelevant to the patentability of software-readable medium claims. True, a software-readable medium claim is not directed to a computer. Nor is it directed to a method. So? It is (if properly drafted) directed to a functional apparatus that can interact with a computer. The MPEP has this right. I recommend reading it carefully.

"Thus, perhaps practitioners now need to recite the loading of the software on a computer or the executing of the software on a computer to make the claim statutory, essentially turning your 'apparatus' claim back into a process claim."

Bad advice. If your computer-readable medium claim is truly directed to an unpatentable abstract idea, then converting it to a method isn't going to help. Similarly, if a method is unpatentable because it is directed to "mere manipulation of abstract concepts," then the corresponding computer claim and computer-readable medium are each likely to also be unpatentable. The converse is also true - if an apparatus claim is directed to more than the mere manipulation of abstract concepts, then it is likely directed to patentable subject matter. And, the corresponding method claim, if tied to the appropriate "particular machine," should also be patentable subject matter. This last point is the one that I expect Examiners will simply not get (or will be instructed not to get).

[GatorAtty]

A difficult question to answer. At this point, I want to wait for the USPTO guidelines concerning software claims in order to develop a reasonable strategy.

[Phil Marcus]

I think David Stein is getting close to the core of the problem. Both the Constitutional language and every version of the patent act since 1793 are based on a misunderstanding, albeit one justified by then current science: that matter is indivisible. Thus, mathematical and algorithmic constructs were different from matter.

Now we know that what we call matter is an interaction with our senses and instruments of clouds of probability of location of sub-subatomic particles, as described in a space of maybe 20 dimensions. Sounds like sci-fi, but it is not.

In addition, it is surely an abstract mathematical construct. So mathematical v. physical is a chimerical distinction. So using section 101 is not how to eliminate applications that attempt to patent, say, Newton's laws or the laws of thermo or corollaries to them.

Section 103 on non-obviousness is the key. So if someone works out something that is useful and non-obvious to a POSIT, then a reading of either the Constitution or the Patent Act in connection with a current understanding of the physical universe should not find any distinction between math describing finance and math describing physics or chemistry or biology.

[amused]

For your amusement, here's an Internet post that says: "With the Bilski court decision, method patents are eliminated."
http://www.fastcompany.com/news/2008/10/31-apple-patent-bilski-case.html

Gee, I read it on the Internet, so it must be true!

[Y A Anon]

The answer is it depends. Standards of review are different in prosecution, licensing, and litigation.

[Anon.]

"non-senior counsel at major corp" has pointed out exactly what you have to do to patent software.

Don't try to patent general-purpose data-transformation algorithms. They're unpatentable, just like they always should have been. If you could theoretically do the entire thing in your head, with the assistance of pencil and paper, given an unlimited amount of time, it's unpatentable.

If you're patenting a computer-assisted rubber-curing process, or something along those lines, you're fine, but you have to *describe* it that way.

If you can't tell the difference, you need to learn more about computers.

[Anonymous Four]

What a bunch of the more avid everything-should-be-patentable people have missed is the Progress Clause:

" To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries; "

The correct ruling would be a Supreme Court ruling declaring that patents are only allowed, *Constitutionally*, in fields in which they promote the progress of science or useful arts. Specifically, monopolies can promote this progress only by providing an incentive to publish.

Pure business method patents, and pure data-processing software patents, have undeniably retarded the progress of science *and* the useful arts. There is ample evidence for this. In fact, there is not one pure software patent (as opposed to patents on manipulating other physical processes using a computer) which has given anyone any incentive to publish; they have served soley to suppress new technology. In fact, in violation of the patent bargain, many patentholders fail to publish adequate descriptions (i.e. working software source code) even after acquiring a patent. In software, patents are simply collected after doing research which would have been done anyway, without the existence of patents.

This is the true Constitutional reason software patents need to be tossed, now.

[AllSeeingEye]

I recommend hiring a good patent attorney...