Bilski: Adding Obvious but Meaningful Limitations

Paul Gardner is PRG’s Academic Director. In an e-mail, I asked him whether claim drafting techniques and strategies can be effectively tailored to satisfy Bilski’s requirements without sacrificing valuable claim scope. Mr. Gardner says yes it can be done most of the time (and PRG is developing the CLE to tell you how). For Gardner, an important consideration in Bilski is between “meaningful limits” versus “nonobvious limits.”

While Bilski requires that process claims recite machine or transformation limitations that “impose meaningful limits on the claim’s scope,” such limitations need not themselves be new or nonobvious. In other words, “meaningful limits” is not to be equated with “nonobvious limits,” and the “meaningful limits” requirement may be satisfied – insofar as Section 101 patent-eligibility is concerned – by machine or transformation limitations which, standing alone, are old or obvious. Once Section 101 patent-eligibility is found to be present, novelty and nonobviousness of the claim as a whole may be satisfied by a novel and nonobvious algorithm in combination with the structural machine or transformation recitations.

This difference is seen in the Federal Circuit’s discussion of Abele.

As Chief Judge Michel points out in Bilski, in Abele the CCPA found a broad method claim reciting only data manipulation steps (calculating the difference between two values and displaying the value of the difference) to be patent-ineligible, but found a dependent claim adding only that the data is “X-ray attenuation data produced in a two dimensional field by a computed tomography scanner” to be patent-eligible, because the data represented physical and tangible objects.

My own caution comes from the CAFC’s nonobviousness analysis in the 2007 Comiskey decision. In describing that case, I led with the headline “35 USC 101 Finds its Teeth (Biting into Nonobviousness)” because Comiskey could be read to indicate that any portion of an invention that constitutes nonstatutory subject matter will be considered de facto obvious. Under this reading of Comiskey, obvious but meaningful limitations may overcome §101, but leave the claim extremely vulnerable under §103(a). The Supreme Court’s 1978 Parker v. Flook decision follows this same line of thinking – treating a non-statutory (but previously unknown) algorithm “as though it were a familiar part of the prior art.”

   

     

     

   

60 thoughts on “Bilski: Adding Obvious but Meaningful Limitations

  1. 60

    “That lays waste to a huge swath of **poorly drafted claims or obvious** inventions.”

    Fixed your typo.

  2. 59

    JBW-

    I see this exact point as one of inconsistency in the law. Benson’s claims were rejected in part because they could only be practiced by a computer. Single-marker diagnostic claims can generally be practiced EITHER by a computer OR by the doctor (in his mind). But the additional of convention modern technology either will not solve the 101 problem because it is insignificant extra-solution activity (a la Bilski) or will not solve the 103 problem because it is obvious (a la Comiskey).

    So it’s a Catch-22: if the claim CAN be practiced entirely in the mind, it’s patent-ineligible; if the claim MUST be practiced by a computer, it’s patent-ineligible; if you try to recite a computer for something that can be practiced in the mind, it’s either patent-ineligible or obvious. That lays waste to a huge swath of inventions.

  3. 56

    I’ll post this in a slightly different way that I did on another thread….

    The FedCir relied on the Gottschalk v. Benson test, which states that a method invention is eligible for patenting if “(1) it is tied to a particular machine or apparatus, or (2) it transforms a particular article into a different state or thing.”

    So, my question is, if the specification of a patent does in fact describe a particular “machine or apparatus”, but a respective claim is drafted in a way that the method steps could be performed either (1) with assistance of said particular machine, 0R (2) solely by a human w/o any machine (e.g., a clinician with a clipboard and pen, for example, performing claim “sensing” and “determining” steps), would it still be patentable-subject matter?

    Stated differently, could you FAIL the Gottschalk test if (1) a method as claimed is CAPABLE OF BEING performed without any machine, though a particular machine or apparatus that enables the method is disclosed in the specification (but is not itself recited in the particular claim), and (2) there is no transformation?

  4. 55

    Hi, this is Sally from Patent Resources Group. We are planning to incorporate best practices that result from this decision into our course “The Bilski Impact: Procuring & Enforcing Software, Business Methods & Bioinformatics Patents” to be held at our next Advanced Courses, April 23-25, in Orlando. By that point, the full impact of this decision will be more apparent, and we believe we will be able to provide the most balanced and tested acumen available in patent law. Please see our CLE course description here: link to patentresources.com

    We should be posting more information about all of our April CLE courses to our website in approximately 2 weeks.

  5. 54

    Product by process ?

    product: the computer processor configured, by execution of a software program, to X.

    process: directing the processor to read and execute the software program.

    Comments?

  6. 53

    As I see it, the situation in the US post-Bilsky is quite similar to that in Europe: US claims now require “machine implementation”, which seems awfully similar to the EPO Board of Appeals’ “technical character”. Now the question is, as in Europe, whether those “machine or transformation limitations” have to be nonobvious, or whether a claim will be allowable even if the only nonobvious limitations are not technical.

    Good luck, we Europeans have already been argueing about that for a couple of decades, and we are no closer to an answer…

  7. 52

    BCD encoding … that was Benson.

    Flook was the alarm setpoints application.

    Comiskey left unanswered the question whether 101-ineligible subject matter should even be considered in a 103 analysis.

  8. 51

    I am still confused by definition of a “general machine” as opposed to a “particular machine.”
    On another blog, someone wrote a ‘particular’ machine would be, for example, a dedicated “router” as opposed to my laptop which would be a “general machine.”

    HOWEVER, in many cases, I can easily load, for example, game software, onto the dedicated router hardware – i.e. also, many times, the router manufacturer may find it cheaper to configure the router with SOFTWARE and use a general purpose CPU rather than “router chips” – to the outside world, the blackbox router would look EXACTLY the same as a “dedicated hardware” router. The outside world could not tell the difference.

    SO is “particular machine” meant to imply “dedicated machine” – i.e. with only a single purpose??
    this would endanger many HARDWARE claims because most dedicated hardware can ALSO be configured for multiple purposes.

  9. 50

    David Stein:

    I guess the language that concerns me is “We remand to the PTO to determine whether the addition of a general purpose computer . . . to Comiskey’s otherwise unpatentable mental process would have been obvious.” And “The rountine addition of modern electronics to an otherwise unpatentable invention typically creates a prima facie case of obviousness.” It seems like they are saying that it’s the addition of the computers or electronics that needs to have been not obvious to support patentability. At least that seems like a plausible way of reading it, especially in light of Flook.

    I confess again that the remand puzzles me. The PTO and Board already determined that the claims were obvious–presumably including the computers and electronics.

  10. 49

    “In Benson (the binary to BCD conversion), the claim was “purported to cover any use of the claimed method in a general-purpose digital computer of any type,” according to the Court. How is a preemptive process (like the binary to BCD conversion) is less preemptive because it was performed by a computer?”

    I completely agree with you here, Sky. The sole advantage of adding “obvious but meaningful limitations” of hardware components (as per this thread) is to meet the new definition of “process.” If the software, without those elements, can’t pass every other test – novelty, non-obviousness, non-preemption, non-vagueness, enablement, written description, yadda yadda yadda – then adding those elements won’t help it.

    “By the way, if the CAFC and the Supreme Court, can define ‘process’ narrower, don’t you think that the terms ‘machine’ and ‘manufacture’ could be similarly narrowed?”

    Sure… but they haven’t done it yet. ;)

    I think there are several bases for hope that the “particularity” of the “machine” tied to the software won’t have to be hyper-specific. Even the majority opinion in Bilski suggests some breadth in this test by arguing that Bilski’s claims “don’t require a machine,” “can be performed mentally,” etc. That suggests that the CAFC just wants to see *something* limiting the claim to a method actually running on a computer. And that argument seems to resonate with the general consensus on the pro-Bilski side of the ball (notwithstanding the USPTO itself, which took an extreme viewpoint on par with Richard Stallman’s…)

    – David Stein

  11. 48

    MG:

    “But doesn’t Comiskey seem to say that if (A) is non-statutory subject matter it will not help you satisfy 102 and 103 because it is treated under Flook as though it is well known? So, if (B) and (C) are well known, and (A) is treated as though it is well known, what is left that is not obvious?”

    Well – you’re trying to have it both ways. Your analysis depends on knocking out A for non-statutoriness (notwithstanding the statutoriness-conferring presence of B and C), and knocking out B and C for obviousness (notwithstanding the non-obviousness of A.)

    That’s not how the analysis works. You have to read the elements of the claim together for each of these tests. While A isn’t statutory by itself, A+B+C is statutory. And while B and C aren’t non-obvious (even together), A+B+C is non-obvious.

    You’re construing Comiskey as stating that if a *single element* is non-statutory *standing by itself*, then it is deemed obvious. I don’t see that in Comiskey – can you cite to a section that supports that concept?

    Moreover, that sort of claim analysis conflicts with the venerable, overarching principle that the claim must be read “as a whole,” etc. (Granted, the CAFC only trots out that rule when it’s convenient ;) )

    – David Stein

  12. 47

    6000 – thx for your comment.

    6000 wrote:

    Question is, does a temporary “modification”, that isn’t even specified in the claim, count as a “transformation”?

    So a claim which modifies flash or magnetic medium should be MORE patentable.

    Also, as for RAM, as long as electrical power is supplied to the machine, the change could be quite permanent – assuming that you have a non-Windoze machine, you could leave the machine on for years with the transferred RAM.

    As for your second comment:
    “change in form, appearance, nature, or character.”

    why would the court exclude nature or character? Also, with “appearance” is this limited to naked eye appearance or appearance using some sort of instrument?

  13. 46

    Readers, when above I typed “commerce” I wasn’t thinking of the sort of “commerce” that EdStirling has now zoomed in on. Didn’t mean to lower the tone. Just goes to show how many different meanings words in a Decision can have. Sorry about that, folks.

  14. 45

    “- Aim: To promote reutilization by enabling the patent examiners of each office to understand each other’s search strategy ”

    Mah search strat: Find the invention. On google hopefully, within 10 minutes. End search.

    Unfortunately, sometimes mah strat isn’t enough. It works good on amendments tho.

    “What’s wrong with reciting transforming the state of the MEMORY (i.e. the particular RAM memory) rather than the state of the DATA?
    any physicist could tell you that software code when executed modifies the state of the MEMORY.”

    Question is, does a temporary “modification”, that isn’t even specified in the claim, count as a “transformation”?

    Well, we have several definitions to choose from:

    “the act of changing in form or shape or appearance”

    that being one that doesn’t really support your position.

    “change in form, appearance, nature, or character.”

    That being one that does support your position, but could be easily modified by the court to exclude “nature” and “character”.

    On the whole, I doubt you’re going to be able to pull transformation of RAM etc off.

    Also, the preemption thing might be a problem, but then again, it might not.

  15. 44

    Bilski is not about software claims. It’s about whether investment strategies are patentable. Bilski’s method is not drawn to a method performed with a computer, it’s about engaging in certain social interactions, i.e. transactions, as a way to enhance one’s financial position. A more pertinent example than MRI software would be a method for picking up women at a bar. Obviously, an effective method in this endeavour is useful for several purposes, such as procreation. Suppose the claimed method steps are, 1) saying hi to a girl, 2) getting the girl’s phone number. This is a transaction, in step 1, you have given the girl recognition, in step 2 you have acquired a phone number, quid pro quo. But the transaction is an exchange of intangibles, similar to commodity shares, per Bilski. Transactions of intangibles alone apparently falls short of the 101 requirement. You should write the number on a piece of paper, maybe buy the girl a drink.

  16. 42

    I think Gardner’s analysis is tough to square with both Comiskey (as Dennis notes) AND Abele. From Abele:

    “Taking each claim as a whole, it is assumed, for analysis purposes only, that any mathematical calculation in the claim is part of the prior art. If what is left is new and unobvious, then the claim, taken as a whole, protects more than a mathematical calculation and it is deemed statutory. But if the remainder of the claim is not novel nor unobvious, then the claim, taken as a whole, merely seeks to protect the mathematical calculation and, as such, does not comprise statutory subject matter.”

    I would argue that Comiskey reversed this analysis by announcing that statutory elements turn the claim into statutory subject matter that is then vulnerable to s.103. The result is essentially the same and, regardless of this tension, Gardner seems to be wrong under either case.

  17. 41

    Disnae, business method claims are not my field but in my field I already see the EPO cranking out its patentability opinions with its search reports within less than 12 months of the filing of the app. Given the internationally acknowledged competence of the EPO search, and the “settled” EPO 101 and 103 caselaw in tune with the Bilski banc mindset (except Newman), I venture to suggest that those EPO opinions, addressing “technical character” and “solution to an objective technical problem” constitute the star for the world’s other Patent Office Examiners to navigate by. If the non-obvious innovation’s “technical” the EPO will give you a patent. Conversely, if you’ve improved commerce and done nothing more, no patent. With the room for manoeuvre given them by the CAFC in Bilski, USPTO Exrs can just go online, read the opinion, then build their own (pre-empt, abstract, post-solution) text around it. All the rest, Highways etc, is just hot air to amuse you. Wishful thinking me surmises that the CAFC understands EPO jurisprudence and (despite protestations to the contrary) is rowing towards it enthusiastically, even while avoiding the obstacles of stare decisis.

  18. 40

    anonymousAgent: The answer to your question is that transforming memory likely does not solve the preemption issue. Thinking that the mere mention of generic components somehow will affect preemption is not a strong strategy. Focus on avoiding preemption by defining the data and tying-in to the physical world (apart from generic components that don’t meaningfully affect preemption).

  19. 39

    David Stein:

    Thanks for your replies. I realize I’m late responding so no hard feelings if you don’t reply to this one. You say:

    “MG: No, I’m not relying on the obvious subject matter of 101 to get around 103.

    In my example, the software (A) has to stand on its own to satisfy 102 and 103. If so, then to that novel and non-obvious software (A), you can add obvious limitations (B) and (C) to satisfy 101.”

    But doesn’t Comiskey seem to say that if (A) is non-statutory subject matter it will not help you satisfy 102 and 103 because it is treated under Flook as though it is well known? So, if (B) and (C) are well known, and (A) is treated as though it is well known, what is left that is not obvious?

    I will admit I’m not clear on that point because the Comiskey decision remanded back to the Board for a determination of whether there was evidence of non-obviousness that would overcome the prima facie case. However, the implication seemed to be that (B) and (C) needed to be combined with (A) in a nonobvious fashion to satisfy 103. The non-obviousness of (A) would not suffice.

  20. 38

    “No, but for a different reason: preemption. That requirement is expressly referenced (and unchanged) by Bilski.”

    Again, why do you suggest that adding a general purpose computer will save you? If the process already isn’t preemptive, you don’t need to add a particular machine. But if the process is preemptive, then the general computer won’t save you. Either way, suggesting the addition of a general purpose computer is pointless.

  21. 37

    Sky,
    What’s wrong with reciting transforming the state of the MEMORY (i.e. the particular RAM memory) rather than the state of the DATA?
    any physicist could tell you that software code when executed modifies the state of the MEMORY.

    Also – what is a GENERAL PURPOSE machine? A “multi-purpose machine”? Even when you burn code on a chip (or build something ‘full-custom’ in hardware), often the chip has multiple purposes? Then there could be no hardware patents either.

  22. 36

    Given the apparent concensus that Bilski leaves nought but chaos in it’s wake, this seems like as good a place as any to laugh at the following, just received from a US associate:

    “Work Sharing Among Five Major IP Offices

    The offices established a cooperative framework of 10 Foundation Projects. These projects were devised to harmonize the search and examination environment of each office and to standardize the information-sharing process. The projects are expected to facilitate the work-sharing initiative by enhancing the quality of patent searches and examinations and building mutual trust in each other’s work.

    Common Rules for Examination Practice and Quality Control (lead: SIPO)
    – Aim: To execute patent examinations at a similar standard and quality through common rules of examination practice and quality control

    Common Approach to Sharing and Documenting Search Strategies (lead: USPTO)
    – Aim: To promote reutilization by enabling the patent examiners of each office to understand each other’s search strategy

    Common Search and Examination Support Tools (lead: USPTO)
    – Aim: To establish a system of common search and examination tools to facilitate work-sharing”

    Not that I’m cynical nor nothin’

  23. 35

    David Stein wrote, “In another section, the CAFC finds Bilski’s invention ineligible under 101 simply because “no machine is required.” If that is all that’s required, then including *any* bits of *any* machine should suffice.”

    Yes, if you assume, as you do, that any machine is sufficient, you’re done. But the CAFC is unwilling to go there, probably because those facts (a computer-implemented process) weren’t before them and, more importantly, the Supreme Court reached the opposite result in Benson.

    The CAFC said the “particular machine” issue wasn’t in play because there was no machine. The CAFC expressly refused to answer “whether or when recitation of a computer suffices to tie a process claim to a particular machine.”

    In Benson (the binary to BCD conversion), the claim was “purported to cover any use of the claimed method in a general-purpose digital computer of any type,” according to the Court. How is a preemptive process (like the binary to BCD conversion) is less preemptive because it was performed by a computer? The answer is: it isn’t any less preemptive. What would argue? “You can still convert it on paper!”?

    If your process is preemptive, reciting a general purpose computer won’t save you. If your process actually recites a sufficient transformation, you don’t need to add a general purpose computer or any other machine, particular or not.

    By the way, if the CAFC and the Supreme Court, can define “process” narrower, don’t you think that the terms “machine” and “manufacture” could be similarly narrowed?

    For software-related inventions, the sky is, indeed, falling. You better start looking for transformative uses and particular machines, not merely general computers or general computer-readable mediums.

  24. 34

    Great insight Visiting. Could it be that the CAFC dislikes: 1) patents on business methods and 2) not only the European touchstone “technical” but also the US “technological arts” as being too vague, 3) tasks the majority Michel opinion with getting to a “patents in all fields of technology” endpoint, in tune with TRIPS, while 4) using the dissent opinions to rehearse all those “public policy” issues that should occupy the minds of the SCOTUS panel, (and all commenting law professors) so that 5) everybody appreciates how boxed in the CAFC is by stare decisis, and so 6) tempers their criticisms and frustrations when reading the 132 page epic. Surely there are enough interesting borderline cases soon to be decided by the CAFC to enable it to emulate the EPO (unencumbered by stare) and use that very caseflow, in aggregate, to crystallise the edge of patentability, and progrssively to move it with the onward rush of “technology”, as the EPO does. After all, the Boards of Appeal at the EPO decide thousands of cases, and its caselaw gets progressively sharper.

  25. 33

    “I’m trying to envision an example of a business method patent that would pass muster after in re Bilski and Comisky, and I’m having some difficulty.”

    Yeah, I think business methods are toast.

    Of course, this depends on one’s definition of “business method.” My definition of a business method is any method that doesn’t involve a computer, software, the making or using of an apparatus/composition/article, etc. And it’s difficult to imagine how most such methods might satisfy Bilski.

    It’s amusing that the CAFC went out of its way to reject a categorical ban on business method patents, while imposing requirements for the term “method” that do exactly that. (Of course, the CAFC also expressly rejected a “physicality” requirement for software patents, while simultaneously imposing a requirement that such patents are either tied to a physical machine or transform a physical article.)

    – David Stein

  26. 32

    I’m trying to envision an example of a business method patent that would pass muster after in re Bilski and Comiski, and I’m having some difficulty.

    In the case of claims written in method format, if you rely on the “machine” portion of the machine-transformation test, how do you avoid a rejections where the naked non-statutory method is given no patentable weight, and the machine is considered just modern electronics.

    If the method performs a statutory transformation of a non abstract thing, then the statutory/non-statutory unbundling should not take place. But such methods probably would not be considered business methods.

  27. 31

    “A claimed process is patent-eligible if it transforms an article into a different STATE or thing.”
    When I transform data, on a physical level, I transform THE ACTUAL MEMORY IN WHICH THE DATA RESIDES into a different PHYSICAL state – for example, I transform the physical state of RAM, flash, etc. (thus, we focus on transforming the memory rather than the data).
    [this is no different than boiling water – no chemical change but a change of state]
    In some memories, such as flash memory, this transformation of physical state is permanent.
    Electronic computers with RAM are not the only examples of ‘digital’ computers – thus, when we recite changing the state of the RAM (or flash or magnetic medium) we are avoid this:

    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.

    i.e. maybe the term “digital computer” is too general – there are certainly digital computers that do not include RAM or flash.
    Thoughts?

  28. 30

    “The CAFC expressly refused to say that such was sufficient to be a particular machine.”

    Yes, and it indicated that the degree of “particularity” would be refined in future cases.

    However, other parts of its opinion hint at the degree of “sufficiency” required. There are at least two hints about this:

    1) Although the CAFC withheld ruling on the “particularity,” it adamantly rejected a categorical ban on pure software patents. By definition, a pure software patent includes no elements of a machine; thus, at least one software invention is (theoretically) claimable in the absence of any type of hardware.

    2) In one section, the CAFC criticizes Bilski’s claimed invention simply because it can be performed solely in the mind. In another section, the CAFC finds Bilski’s invention ineligible under 101 simply because “no machine is required.” If that is all that’s required, then including *any* bits of *any* machine should suffice.

    Yes, it’s true that the magnitude of “particularity” and the requisite types of “tying” will be fleshed out by future cases. For now, though, it looks hopeful that only a modest degree of “tying” and “particuarity” are required.

    “If ‘calculating the difference between two values and displaying the value of the difference’ was not 101 subject matter, you really believe that ‘calculating the difference between two values BY A PROCESSOR and displaying the value of the difference’ would be?”

    No, but for a different reason: preemption. That requirement is expressly referenced (and unchanged) by Bilski. Ditto your Gottschalk example.

    – David Stein

  29. 29

    More questions …

    if “calculating the difference between two values and displaying the value of the difference” was not 101 subject matter, you really believe that “calculating the difference between two values BY A PROCESSOR and displaying the value of the difference” would be?

    Would converting binary into binary coded decimal BY A PROCESSOR be 101 subject matter?

  30. 28

    “To the ([hopefully] novel and non-obvious) steps enacted by the software, we’re adding the (admittedly obvious) steps of running that software on a processor. That is the key element that the CAFC stated was missing in Bilski”

    But the CAFC expressly refused to say that such was sufficient to be a particular machine. So there’s nothing “key” about it. The key was that both parties agreed that there wasn’t any particular machine and thus that prong of the analysis wasn’t addressed in Bilski.

    I don’t see anything in the Supreme Court caselaw that would suggest that reciting a processor is significant enough, whether in the preamble, in the body, here or there or anywhere, not in a fox, not in a box …

  31. 27

    Does this uncertainty then in the least leave us with the need to draw no fewer than 3-4x+ the number of claims we would normally; consisting of the 2-4+ various major general claiming approaches enumerated since Bilski; in order to insure that at least one set of claims will survive the now Byzantine examination/ BPAI/ Fed Cir/ litigation gauntlet?

    Is this really what our forefathers had in mind?

  32. 26

    “There’s no way that “computer-implemented” will satisfy the “particular machine” requirement.” I agree

  33. 25

    @Sky:

    “provided those obvious limitations are not ‘insignificant postsolution activity,’ which ‘will not transform an unpatentable principle into a patentable process’ according to the Court.”

    Ah, but the “activity” added is hardly “insignificant.” To the ([hopefully] novel and non-obvious) steps enacted by the software, we’re adding the (admittedly obvious) steps of running that software on a processor. That is the key element that the CAFC stated was missing in Bilski – note all of the rhetoric about “Bilski’s invention can be performed simply by a thinking human” and “Bilski’s invention does not require a computer.”

    Admittedly, this extra “activity” is extremely well-known – if you have some software, you can run it on a processor. ;) But the patentability hurdle overcome by this extra “activity” is not 102/103, but 101. So adding this “obvious” limitation to the “non-obvious but unpatentable” software may definitely confer patentability.

    “that seems like a difficult balancing act: adding limitations that insignificant enough to get you the breadth that you want, but significant enough to transform an unpatentable principle into a patentable process.”

    I don’t think so. If I have a novel software invention (x), how much conceptual breadth is cut away by adding the limitation of running (x) on a computer?

    There’s definitely some ENFORCEMENT breadth limited by this limitation, since the patent may only be infringed when an end-user runs the software. But I think that the uncovered infringement activity has been problematic under any other claim style… throughout the entire history of software patenting.

    “I can’t see how reciting ‘a computer implemented method’ is going to save a claim that otherwise would have claimed an unpatentable principle.”

    It won’t, because that phrase doesn’t really mean much. It wouldn’t even have survived State Street, because it’s arguably not even limiting – it doesn’t “breathe life and meaning” into the claim, right?

    There’s no way that “computer-implemented” will satisfy the “particular machine” requirement. It’s clear that the “particularity” requirement is pretty empty – the CAFC doesn’t even stick with it throughout its majority holding! – but this is one instance where it would be utilized with effect.

    – David Stein

  34. 24

    6000 — isn’t Bilski saying that if you amend the Bilski claim so that it meets 101, then the entire claim is examined to determine whether it meets 103?

  35. 23

    “DDC’s point is that you probably cannot rely on non-statutory subject matter to distinguish over the prior art. Further, even when arguing secondary considerations, you’ll be forced to establish that the nexus of the secondary condition is something other than non-statutory subject matter.”

    Keyword here kiddies: “probably”. “probably” is the prosecutors dream come true so far as I can tell. It’s when that word is left out that they’re upset. As to the part about arguing secondaries, meh, how likely is that to come up in prosecution/enforcement? Probably not very likely. If your combination is truely non-obvious, and you’re not just trying to push bs through the office then you should be in the clear.

    But, hey, maybe it is all a big waste of time. Don’t take me for an advocate of pulling sly tricks to get around good faith efforts by the courts to limit patentable subject matter to things that it was intended to cover.

  36. 22

    “I can’t see how reciting “a computer implemented method” is going to save a claim that otherwise would have claimed an unpatentable principle.”

    Because computers are so fast and awesome and important for everything and if we don’t allow those claims to be patented all the programmers will go back to playing video games and basically say FU America and we’ll lose to China which has the most awesome patent system ever.

    /patent crack addict off

  37. 21

    David Stein wrote, “If so, then to that novel and non-obvious software (A), you can add obvious limitations (B) and (C) to satisfy 101.”

    provided those obvious limitations are not “insignificant postsolution activity,” which “will not transform an unpatentable principle into a patentable process” according to the Court

    that seems like a difficult balancing act: adding limitations that insignificant enough to get you the breadth that you want, but significant enough to transform an unpatentable principle into a patentable process

    I can’t see how reciting “a computer implemented method” is going to save a claim that otherwise would have claimed an unpatentable principle.

  38. 20

    I think that tieing rejections of patents to an 1850 court decision is archayic a paticular machine? the spectrum of inventions in that era was so narrow that the judges couldent envision anything else.Also tests such as novel and non-obvious are also out of date. Ideas like valuable and in need of protection by the granting of a patent is the way to make decisions today.Things like marketable and capable of contributing to free market commerce are more the terms of modern judgements I would think.Business concepts should have been patentable long ago,as the financial reward for creating that segment of the economy that they created.Talk about a way to repress jobs growth and stagnate prosperity.!

  39. 19

    Once it has been decided that a claim meets the requirements of 101, then all elements/features/limitations of the claim are evaluated to determine whether the claim survives 103, right?

  40. 18

    MG: No, I’m not relying on the obvious subject matter of 101 to get around 103.

    In my example, the software (A) has to stand on its own to satisfy 102 and 103. If so, then to that novel and non-obvious software (A), you can add obvious limitations (B) and (C) to satisfy 101.

    Adding even well-known elements to a novel and non-obvious element cannot make the element known. It still includes that nugget of novelty and non-obviousness, PLUS additional limitations for other purposes.

    – David Stein

  41. 17

    “6000, that is sort of what I was thinking.”

    I don’t understand what you two are getting at.

    DDC’s point is that you probably cannot rely on non-statutory subject matter to distinguish over the prior art. Further, even when arguing secondary considerations, you’ll be forced to establish that the nexus of the secondary condition is something other than non-statutory subject matter.

  42. 16

    6000:

    How do you rely on the nonstatutory subject matter to pass 103 if the nonstatutory subject matter is treated “as though it were a familiar part of the prior art”?

  43. 14

    “Comiskey could be read to indicate that any portion of an invention that constitutes nonstatutory subject matter will be considered de facto obvious.”

    So what? The point of adding in that obvious subject matter is to get around 101 not to get around 103. You rely on your nonstatutory subject matter to pass 103.

  44. 13

    Curious:

    I think that is true. In Comiskey they found a few of the dependent claims to meet the statutory requirements because they were tied to machines (computers), but found them to be prima facie obvious because it appeared just adding computers (and other machines) to the process would have been obvious.

    David: The “so what” is that the non-obviousness of the mental process, A in your example, doesn’t help. It has to have been non-obvious to add a computer to the mental process. At least that is how I read Comiskey.

  45. 12

    Seems hard for someone to come up with guidance for claim drafting after reading Bilski. After all, the actual holding in the case relates to claims that had no physical limitations; no computing devices or steps that must be performed by a physical device. Technically, most of the discussion in Bilski is dicta.

    Narrowly, the test in Bilski applies to cases where claims are not limited to computer implementation. The “particular machine” or “physical transformation” are so poorly defined any guidance on how to satisfy the Bilski test should be viewed skeptically.

    How about this simple guidance? If your invention can only be implemented on a computer, then include limitations to that effect. Have steps that are explicitly performed by (or only performable by) the computer. If the data represents something in the real world, even better, and make sure it is clear in the claims.

    Per Supreme Court precedence, the governing issue is still whether the claim is too broad … whether all uses of the algorithm (assuming that is what the invention is) are pre-empted. Ask if the claims would prohibit someone from doing academic research… would their pencin and paper and brain infringe the claim?

  46. 11

    Yeah, that’s probably true – but so what?

    Let’s say you have software invention (A), and you add machine elements (B) and (C) to the claim. The point is to tie (A) to a “particular machine” (B)+(C) to meet the requirements of a patentable process under Diehr. So who cares if (B) and (C) are well-known components? The point of adding (B) and (C) is not to add non-obviousness to (A)… which had better be novel and non-obvious standing on its own.

    – David Stein

  47. 10

    curious: Yeah, that’s probably true – but so what?

    Let’s say you have software invention (A), and you add machine elements (B) and (C) to the claim. The point is to tie (A) to a “particular machine” (B)+(C) to meet the requirements of a patentable process under Diehr. So who cares if (B) and (C) are well-known components? The point of adding (B) and (C) is not to add non-obviousness to (A)… which had better be novel and non-obvious standing on its own.

    – David Stein

  48. 9

    Is this true following Bilski?

    “Comiskey could be read to indicate that any portion of an invention that constitutes nonstatutory subject matter will be considered de facto obvious.”

    Anyone have any comments?

  49. 7

    “Does inducement get around this problem of enforcement?”

    To a limited extent. But purveyors of infringing software haven’t really been the targets of direct infringement since… well, since web downloading became feasible, so that such software could be disseminated without a computer-readable medium.

    However… as we weigh down our software claims with more and more hardware elements, the portion of the claim infringed by the distributor shrinks. And that critically bears on the damages calculation, since the distributor contributed to an end user’s infringement of only a small portion of the claim…

    Back me up here, litigators – am I remembering that correctly?

    – David Stein

  50. 6

    “Does inducement get around this problem of enforcement?”

    Of course. Chem and biotech people have been dealing with this for years. And most of the Examiners have Ph.Ds.

    Welcome to our world.

  51. 5

    “It’s pretty tough to square that sentence with the Flook case, which pretty clearly held that a novel, not-obvioius alogorithm is not made patentable by combining it with well-known physical processes…”

    Good point. But Flook was more about preemption – an algorithm that preempts every use of a mathematical operation (like BCD encoding) can’t be made non-preempting by adding generic elements.

    I don’t think Bilski redefines preemption. An algorithm that’s so broad as to preempt every use is not going to be patentable no matter how tightly it’s tied to any particular machine. (But I suppose, at some point, that the machine *becomes* the subject of the claim, and the preempting software is only tangential…)

    Rather, the purpose of importing machine components into the claim here is to meet the CAFC’s restricted definition of “process” under a revitalized Diehr test.

    On a side-note… is it getting really difficult for any other practitioners to explain this area of law to clients, colleagues, or… well, anyone, really?

    – David Stein

  52. 4

    “I agree that that tactic will more likely result in a patent after Bilski. But it renders enforcement even more difficult, since an infringer must not only be in possession of the infringing software, but must also run the software on a computer with that hardware.”

    Does inducement get around this problem of enforcement?

  53. 3

    “Once Section 101 patent-eligibility is found to be present, novelty and nonobviousness of the claim as a whole may be satisfied by a novel and nonobvious algorithm in combination with the structural machine or transformation recitations.”

    It’s pretty tough to square that sentence with the Flook case, which pretty clearly held that a novel, not-obvioius alogorithm is not made patentable by combining it with well-known physical processes: “The notion that post-solution activity, no matter how conventional or obvious in itself, can transform an unpatentable principle into a patentable process exalts form over substance. . . . Even though a phenomenon of nature or mathematical formula may be well known, an inventive application of the principle may be patented. Conversely, the discovery of such a phenomenon cannot support a patent unless there is some other inventive concept in its application.”

  54. 2

    So the message is: continue claiming the invention as a novel and non-obvious method, but to add enough non-novel-but-somehow-‘meaningful’ machine elements as claim limitations to make it pass the first prong of Diehr?

    I agree that that tactic will more likely result in a patent after Bilski. But it renders enforcement even more difficult, since an infringer must not only be in possession of the infringing software, but must also run the software on a computer with that hardware.

    My problem with this tactic (from an enforcement perspective) is that as hardware elements are added to the claim, the patent becomes more tightly focused on the act of running the software. Thus, the cause of action is more tightly focused on the end user, while the *actual* infringer – the company that produced the software and gave it to a customer – becomes increasingly tangential to the cause of action.

    I’m wondering if we can walk a middle ground by dragging such hardware elements into the preamble – not as elements of a claim, but as the context in which the infringing method operates. I think it might be possible to “tie” the claimed invention to an apparatus sufficiently to pass Bilski, while not necessarily importing the hardware into the body of the claim.

    I’m still cogitating on this style – but feedback would be appreciated. (Can’t wait for PLI’s and PRG’s CLEs on Bilski.)

    – David Stein

  55. 1

    “obvious but meaningful limitations may overcome §101, but leave the claim extremely vulnerable under §103(a)”

    Feel the python.

Comments are closed.