![136px-Spline_(PSF)[1]](https://cdn.patentlyo.com/media/2017/11/136px-Spline_PSF1.png)
Tough case for Livermore Software. The PTAB just rejected its claims as abstract ideas. Claim 1 is a bit long but is listed below as a method for joining mis-matched portions of an object’s surface (NURBS patches) so that a whole object model is prepped for finite element analysis (FEA). The problem for the patentee is that it presents a step-by-step algorithm with such precision that the Examiner (and PTAB) saw it as claiming a mathematical algorithm.
Ex parte Hartmann, Apn. 14/051,559 (Nov. 2, 2017) [PTAB Decision: Livermore101]
The PTAB writes:
We agree with the Examiner that claims 1, 7, and 13 are directed to the abstract idea of manipulating data through mathematical relationships, which is similar to the computing formula discussed in Parker v. Flook, 437 U.S. 584 (1978), and the Arrhenius formula in Diamond v. Diehr, 450 U.S. 175 (1981). . . . [All of the claimed steps] are abstract mathematical concepts and algorithms that could be performed in the human mind, or by a human using a pen and paper, without need of any computer or other machine.
Moving to Step-Two of Alice, the PTAB wrote:
We find that the additional limitations … do not add significantly more to the abstract idea. … Particularly, claims 1, 7, and 13 recite well-understood, routine, and conventional elements (i.e., computer system having an application module, memory, processor, and computer readable storage medium) that enable the creation of a computerized model using a “generic computer structure that serves to perform generic computer functions that are well-understood, routine, and conventional.”
Interesting aspect of the decision here is the notion that the spline adjustment could be done by a human with pen and paper. The historical approach to do this well required the use of a spline tool (shown above).
In its decision the PTAB suggests (but does not hold directly) that the claims may be eligible if they had expressly (rather than implicitly) recited “a CAD model, a computer aided engineering (CAE) model, or a mechanical engineering design analysis for a product (e.g., car, airplane).” One way to think about the claim below is as a step for enabling model conversion (an abstract idea) but not the actual full conversion process (a concrete application rather than an abstract idea).
Rejected Claim 1 below:
1. A method of joining imperfectly-matching Non-Uniform Rational Basis Splines (NURBS) patches to form a computerized model suitable for finite element analysis (FEA), said method comprising:
receiving, in a computer system having an application module installed thereon, definitions of a first NURBS patch and a second NURBS patch, the first and second NURBS patches to be joined together along a physical boundary defined in a first curve with a first set of control points, associated weights and a corresponding first plurality of knot-vector values in the first NURBS patch and defined in a second curve with a second set of control points, associated weights and a corresponding second plurality of knot-vector values in the second NURBS patch, wherein the first set of control points and said first plurality of knot-vector values are different from the second set of control points and said second plurality of knot-vector values;
normalizing said first plurality of knot-vector values such that the first curve’s parametric length equals the first curve’s physical length in the first NURBS patch;
normalizing said second plurality of knot-vector values such that the second curve’s parametric length equals the second curve’s physical length in the second NURBS patch;
determining a common curve as an overlapped section of the first and the second curves to represent the physical boundary;
adjusting the first and the second curves such that first and second projection points correspond to starting and end locations of the common curve, respectively;
designating one of the first and second curves having less number of control points along the common curve as a master curve, the other as a slave curve; and
determining a set of linear constraint equations for numerically connecting the first and second NURBS patches along the physical boundary by computing dependencies of the slave curve’s control points to the master curve’s control points,
whereby the first and second NURBS patches together with the set of linear constraint equations for the control points along the physical boundary enable a computerized model created therefrom suitable for finite element analysis.
Interesting aspect of the decision here is the notion that the spline adjustment could be done by a human with pen and paper. The historical approach to do this well required the use of a spline tool (shown above).
Spline tools, rulers, slide rulers, calculators and (oh yes) computers all fall into the category of “human with calculating tool”. That’s what “pen and paper” is about. It doesn’t literally mean “pen and paper”. I mean, “pencil and paper” work. “Chalk and chalkboard” work. “Digital pen and monitor” work. “Pen, paper and calculator” works. “Computer, chalk and chalkboard” work.
The point is you can’t claim a “new” method of calculating a number and make it eligible by reciting “wherein you use a pen and paper” or “wherein you punch the numbers and the multiplication signs in a computing device and hit the return key”. It also doesn’t matter if you recite “then use the number in a modeling application.”
People will struggle with the result in this case because it’s the absolute correct result. This stuff doesn’t belong in the patent system. Claiming “computerized models” is a j o ke, and a bad one. “Computerized models” are — at best — intellectual tools that might ultimately lead to a patentable (and patent eligible) thing that exists in the physical world, or a process that transforms the physical world. But the “computerized models” themselves are just expressions of math. Ineligible.
All this logic / math “on a computer” junk is ineligible. Start from the fundamentals and work outwards, not the other way around.
>Computerized models” are — at best — intellectual tools that might ultimately lead to a patentable (and patent eligible) thing that exists in the physical world, or a process that transforms the physical world.
Representing information takes space, time, and energy. A “computerized model” does exist in the real world and has structure. (I know, according to you things exist with functionality that have no structure.)
Representing information takes space, time, and energy
Yes, writing instructions takes space time and energy.
Did you have a point?
I knew you didn’t. I’m just mocking you because you’re silly and you deserve it.
MM, the problem is that the mathematical algorithm is not applied to do anything actually useful in terms of providing a new or improved manufacture. The method could be highly useful in a graphics engine for a display. If claimed like that, it would pass 101 per Alappat.
graphics engine
The answer to the question: when is an “engine” not an “engine”?
Out of the frying pan into the fire?
If the CAFC were to re-affirm the reasoning of Alappat, in a case not distinguishable from Alappat, then that might be sufficient to induce SCOTUS to grant certiorari on a fifth 21st century 101 case.
DS,
I do not recall the name of the case, but one of the very first after the Bilski decision did that very thing: affirmed Alappat as good law.
To my knowledge, there have been exactly zero that affirmatively (and properly) states otherwise.
The other point is that, if the novelty or non-obviousness of a mathematical algorithm is to be decided by the USPTO, who would be qualified to examine or practice before the USPTO.
According to this document
link to uspto.gov
not only are mathematics majors not qualified to practice before the USPTO, but “mathematics courses” are listed as “typical non-acceptable course work”
That is perfectly fine on the basis that, under the law as interpreted by SCOTUS, mathematical algorithms, principles etc. are ineligible subject matter.
But if that were not the case, it would be absurd if electrical engineers, chemists, computer engineers lacking basic knowledge of the state of the art in contemporary mathematics were to consider themselves qualified to determine the novelty or non-obviousness of a mathematical algorithm, or to practice as attorneys before the USPTO. Suppose, for example, that the alleged novelty or non-obviousness applied principles of martingale theory, stochastic calculus, Ito’s lemma etc. in an apparently novel way to yield an allegedly new and better method of pricing financial derivatives (more demanding that the absurd kindergarten “mathematics” of the formulae in Bilski and Flook, or the trivial subject matter of In re Bernhart). Or if someone claimed to have invented some other purported application of homology theory to some practical end. Which art unit would have the task of evaluating it properly on the basis of 102 and 103?
And does it really serve a useful purpose if basic mathematical algorithms are tied up in monopolies on the basis that a PHOSITA in some “technical” field can be presumed in law to be mathematically incompetent?
At least if cases got to the CAFC, they might end up on a panel including Judge Taranto, who, from his curriculum vitae, could be presumed qualified and capable of distinguishing the wheat from the chaff.
…or perhaps look to cryptography…
ALL of this is captured already here in the comments – see post 2.
A cultural note. When I was taking Math 55 at Harvard, the Four-Color Theorem was proved by grinding through all 1,936 cases.
Professor Andrew Gleason the next day came into class the next day and lectured to us students that use of computers was evil because it made people lazy and did not create any new knowledge at all (except perhaps that the theorem was true). He mentioned that at least ~300 of map subtypes relevant to the proof were known in the 1930s.
I really try to distinguish the issue of algorithms and of protocols, which use those algorithms, in Undermining the patent system and dismantling the healthcare system are two steps in an ongoing effort to harden class and racial divisions.
It is unfortunate that there are situations in which the common practice is to use algorithm were protocol is meant and vice versa.
Algorithm is ill-defined in general. I think it mostly means a small constructive theorem. Thus the Spanning Tree Algorithm is analytically derived from a “bigger” theorem in Graph Theory, which is itself an analytical result.
Radia Perlman used prior and analytic knowledge of Graph Theory to construct the Spanning Tree Protocol, which was new synthetic knowledge and which could transform an interconnected collection of communications computers, into a usable computer network. (The appropriate terminology of the 80s would have been communications subnet or comms subnet.) The Spanning Tree Protocol could easily been encompassed within the metes and bounds of an allowable patent claim.
Now not all algorithms are used within the context of communications protocols as the cases of Diamond v Diehr, Gottschalk v Benson, Parker v Flook, and In re Alappat show.
I would be hesitant to describe all of the claims of these preceding cases to be directed toward algorithms. I would describe some to be directed toward mathematical results derived analytically from prior knowledge (natural phenomena, natural laws, or mathematical postulates). It seems not sufficient to make allowable for the metes and bounds of a claim merely to encompass a combination of prior knowledge and analytic knowledge.
Now can pure mathematical reasoning provide new synthetic knowledge. I think the answer is no even though mathematicians describe some algorithms or results as constructive or as synthetic. Such algorithms or results are analytically derived from postulates or previously derived theorems. Postulates count as prior knowledge because they are not derived from experience. The parallel postulate is prior knowledge of Euclidean geometry.
Also when a physicist describes the space in which we live to be locally Euclidean but globally non-Euclidean, that statement constitutes prior knowledge because it is a natural phenomenon that may discovered by experimentation, but it is not analytically derived from other prior or posterior knowledge.
Algorithm in Computer Science means a set of determinate steps.
That’s it.
A protocol if determinate is an algorithm.
Plus, am I sitting on a chair or the static equations that describe the chair?
Many networking protocols may use a graph theoretic algorithm.
While a networking protocol may provide a means for organizing the activities of many devices, a graph theoretic algorithm does not.
I think this is the distinction that Radia Perlman was making in her famous 1985 paper entitled “An Algorithm for Distributed Computation of a Spanning Tree in an Extended LAN.”
In simplified Kantian epistemological formalism a chair represents posterior knowledge because it was constructed directly or indirectly by humans. At some time in the past, the knowledge in the construction of the chair represented NEW synthetic knowledge. Now it is posterior knowledge and therefore prior art.
The static equations that describe the chair represent analytic knowledge derived from natural law (prior knowledge) that are part of an area of physics called statics.
Imprecision in terminology is one of the problems that make 101-eligibility analysis to be difficult.
>While a networking protocol may provide a means for organizing the activities of many devices, a graph theoretic algorithm does not.
That doesn’t mean that a “networking protocol” isn’t an algorithm only that it is not a “graph theoretic algorithm.”
Plus, I don’t think you can say that an algorithm cannot “provide a means for organizing the activities of many devices.”
I’ll have to take a look at the paper.
>Imprecision in terminology is one of the problems that make >101-eligibility analysis to be difficult.
This is definitely true.
Plus, the software code is not the equations any more than the chair is the static equations. The code was constructed directly or indirectly by humans.
“I claim the static equations of the chair” makes about as much sense as the “software is a logic patent” mantra that seems to run rampant here.
There is a far easier way of making the distinction, Joachim – see post 2.
“ but it is not analytically derived from other prior or posterior knowledge.”
Not so sure about that – you might check into the work being done by the VR people who DO – on a daily basis – perform such analytical derivations.
Malcolm – see post 2 (as you are one clearly lost on the differences)
So the PTAB also does not “get” the differences between:
math
applied math (as is found in most all engineering), and
mathS (the philosophical view that overlaps math and applied math)
The thing about 101 that is so appalling is that it puts the examiners in a position of being all knowing. I have a case with a 101 now–that’s it–and the examiner said that he knew he couldn’t find it in the prior art, but if he let it out (as in allowed the claims) then he didn’t know what we would do with it and his career could be lost if it was a big litigation.
So, he hides behind 101 and fabricates nonsense.
I probably have at least five cases that only have 101 rejections. This is just what they do now.
Out of how many?
Relevance? The murderer killed five people. Really, out of how many?
“then he didn’t know what we would do with it and his career could be lost if it was a big litigation.”
lol wut?
I know 6. I get that a lot now where part of the interview is the examiner assessing why I want the patent.
If they feel you might try to license it, then it is much harder to get an allowance. I actually have to yell at some of them. I tell them listen. We are X and we spend billions in research and development. Stop acting like we are a troll.
DIRECTLY compliments of the SAWS mentality.
Never mind the fact that from the very beginning, the US system was built so that the property obtained was fully alienable – on purpose.
” I actually have to yell at some of them. ”
I’m going to have to disagree with you there skipper.
“I tell them listen. We are X and we spend billions in research and development. Stop acting like we are a troll.”
I think you could just use your for srs tone instead of yelling.
I think it might be easier to just tell em “who even knows what all they’ll do with it, maybe try to get some business partners, maybe let some outside folks use it for a fee, or whatever etc. etc. it isn’t really any of OUR business I don’t suppose”. Emphasis on the our, make sure to draw them into your own third party perspective (as an office of the court and an agent of the federal gubmit conversing) the convo. Even if you do work in house.
B-b-but 6, this particualr personal property (if granted) happens to also be a “public right,” so the government wanting to know what you plan to do with it is entirely fair game.
Right?
This is all just meaningless goobly goop. There is simply no meaning to any of this and no way to determine what the outcome will be of any given claim.
(And, the judges/examiners become these ultimate finders of fact that seem to just know somehow everything about every invention. No evidence needed. No worries about hindsight reasoning.)
Factual findings by an executive branch body.
Does the PTAB being the “judicial function” within that executive body excuse it from the otherwise applicable Administrative Procedures Act and/or fact-based “taking-Official-Notice-even-while-denying-that-taking” limitations that apply to examiners?
Do we have an Admin Law White Knight high atop a horse somewhere that might deign to point the way through this morass?
These are great points. I would like to see more written about this.
Yes, including what auto-plural-photos-stitching art was of record in this mere ex parte PTAB decision, and what arguments against this dubious 101 rejection were presented. Doesn’t this claim fit into non-101 decisions on specific improvements in computer operations?
The Supreme Court spawned “I-know-it-when-I-see-it” combined with its own institutional disregard (“We are the SUPREME Court, so we do as we please”) have been the fire hose training (apparently) not only the rogue panel-to-panel differences at the CAFC, but also the Emperor-like (clothes optional) Article I “courts.”
Comments are closed.