Finisar v. DirecTV: Software Means-Plus-Function Claim must be Supported by Particular Structure

Finisar v. DirecTV (Fed. Cir. 2008) [This is Part II of the case discussion. Read Part I]

In a cross-appeal, Finisar asked the CAFC to overturn the lower court’s findings that its means-plus-function claims were invalid for failing to disclose any particular structure to perform the claimed function.  A patent applicant is allowed to claim that her invention covers the “means” for accomplish certain results so long as the application also discloses the “structure” for accomplishing the result.  That structural disclosure must be found somewhere in the specification. Means plus function claims are generally easier to write and are thought by many to be quite broad. Over the past decade, however, courts have increasingly been wary of particularly overbroad claims.

In this case, the claimed means and functions were software related. A number of CAFC cases have required that software means-plus-function claims provide more structural disclosure than just a general purpose computer. Rather, the disclosure must be “the special purpose computer programmed to perform the disclosed algorithm.” WMS Gaming; See also Aristocrat. The CAFC sees its structure requirement as quite minimal and easy to satisfy. Here, however, the complete lack of structure leaves the claims so vague that “one of skill simply cannot perceive the bounds of the invention.”

Notes:

  • One of the Finisar MPF clauses reads as follows:
    • database editing means, coupled to said one or more computer memory devices, for generating a hierarchically arranged set of indices for referencing data in said information database, including distinct indices for referencing distinct portions thereof, and for embedding said indices in said information database . . .”

83 thoughts on “Finisar v. DirecTV: Software Means-Plus-Function Claim must be Supported by Particular Structure

  1. I see that you are indeed correct it seems MM, I had taken the recitation of Aristo to mean that they were actually using that. But, upon closer review it seems you have it. It’s just plain ol’ indefinite without regard to having improperly invoked or any such thing. And yes your example shows the concept though really just rereading the decision from page 28 does it.

    I had taken the Aristo quote to mean they were actually caring about the “simply disclosing a computer as the structure designated to perform a specific function does not limit the scope of the claim … as required by 112 6th. But that was wrong, they did not care so much about that and just through it in for the sake of completing the quote I suppose. They could have just cut off the 2nd sentence of that quote if they didn’t care about it. Still with the 2nd sentence floating around I’d be more than a little careful with my structure in my spec I referred back to.

  2. “Idk about that MM, the ruling is imo (they are quite vague at particularly this point in the decision so it has to be an opinion) 112 6th was not properly invoked, not necessarily that it was nor it wasn’t, but that the improperness of the invocation (you could I suppose interpret this instead as the improperness of the structure) leaves the claim wanting for definiteness.”

    The only way anybody can find this portion of the ruling “vague” is if they are trying really hard not to understand what the court is saying.

    If the CAFC felt that 112P6 was improperly *invoked*, one has to assume that they would say so. Instead, they plainly approve of the district court’s 112P6 analysis and even reiterate the 112P6 test which they then proceed to apply to the claim: “the patent must disclose, at least to the satisfaction of one of ordinary skill in the art, enough of an algorithm to provide the necessary structure under § 112, ¶ 6.” Then, looking at the facts, they conclude that “The district court correctly determined that the structure recited in the ’505 **specification** does not even meet the minimal disclosure necessary to make the claims definite.”

    How can anyone possibly argue that the CAFC was not doing a straight-up 112P6 analysis? 112P6 was invoked as surely as the sky is blue. It was invoked by the magic words that no professional would ever use *unless* they wanted to invoke the 112P6.

    The problem illustrated by this case is that mere invocation is not enough. After you summon the demon, you must remember to feed it.

    “the improperness of the invocation (you could I suppose interpret this instead as the improperness of the structure)”

    Again: what you are granting as a plausible alternative is, in fact, the only reasonable way to read the district court’s decision and the CAFC holding. 112P6 was invoked with the magic words, but the alleged “structure” recited in the spec was, in the context of this claim, not sufficiently definite.

    I’ll try a third (forth?) example to illustrate the point. A claim recites “a means for carrying fluid” and the specification lists as the corresponding “structure” the following: “a solid that can hold a liquid.” Sure, it’s a “structure”. But it doesn’t limit the claim, any more than the “means+function” invocation recited in the claim. We can call this sort of practice “sloppy” or we can call it “cheating” but it fails regardless.

    Note also that the CAFC hints that, had 112P6 not been invoked, the decision regarding indefiniteness might have come out differently, although a tiny bit more than a mere microprocessor might have been needed in place of the M+F language: “This court does not impose a lofty standard in its indefiniteness cases…. But in this case, the claims are already quite vague. Without any corresponding structure, one of skill simply cannot perceive the bounds of the invention.”

  3. I’m saying that it is structure in either event, claim or spec, but that when in the spec without an algorithm it is not a sufficiently limiting structure for a “means” in a claim to be interpreted by a POSHITA as being a sufficient boundary for a claim because all you are trying to do is limit the generic “means” in terms of a generic means. In essence you are failing to present a “corresponding structure” to the means as required by 112 6th, you are merely presenting a “corresponding functional language”.

    slippery slope you have a different problem in that you think that because Mas-Ham makes any pure functional claim element that lacks sufficient structure are interpreted under 112 6th, but what does that matter if if the functional language of a mircroprocessor in a claim would be purely functional after having referred to the language in the spec? That does not, by itself, make the claim indefinite, I see functional language with no structure backing it up in the spec all the time, I don’t hold them all indefinite, because the claim makes up for that portion not helping to define the invention with other limitations. A claim may include a lot of pure bs functional language limitations not backed up in the spec by any structure before it’s hit with indefiniteness unless those limitations are the only subject matter making up the claim. In this case there simply was no other subject matter helping to define the invention.

  4. e6k: Sorry, I’m not sure I follow. Are you saying that a generic processor would be structure if recited in a claim, but not structure if recited in the spec? If so, I fail to see the basis for that kind of split interpretation.

  5. “This is not the same as saying that the term “microprocessor” in a claim would not tell POSHITA what the bounds of the term in the claim is.”

    Actually, it indirectly is. First, assume that a programmed microprocessor is essentially purely functional if there is no algorithm (that’s what Aristocrat is saying). If the same programmed processor recited in a claim, it must be interpreted under 112p6 because, under Mas-Hamilton Group Inc., pure functional claim elements that lack sufficient structure are interpreted under 112p6. Next, under Arisocrat, since there isn’t any algorithm corresponding to the claimed programmed processor, it’s indefinite.

  6. Huh, I see what your issue is, that a microprocessor in the spec should be considered sufficient structure for “means” in the claim to refer back to. Well, I have news for you. This little issue was settled prior to this case. That was what Arito was about iirc with smatterings of WMS Gaming thrown in for good measure and concuring opinion.

    “So, if we can show that a corresponding processor claim _is definite_, then it logically follows that the M+F claim invokes 112 6th, and is definite. Proof by contradiction.”

    No. Just because putting a structure in a claim leads to there being a structure in the claim (but doesn’t necessarily makes it definite, though it might) does not mean putting a generic structure in the spec and referring to it as a “means” in as the actual limiting words (the claim) makes a claim definite as well. Why? Because saying “a generic processor performing x” in the spec (which is what they put) is naught by functional language. Or so the court has held in cases prior to the instant case, and they did so with the reasoning that the “structure” of a generically produced microprocessor is not sufficiently limiting to tell a POSHITA what the bounds of the term “means” in the claim are. This is not the same as saying that the term “microprocessor” in a claim would not tell POSHITA what the bounds of the term in the claim is. If you want to take issue with the cases where this was held, Aristo etc. feel free, but I’m not having that discussion right now. We’re taking that holding as a given in this case.

    “I would rather find broad claim terms that acheive the same purpose.”

    PDS says something worthwhile now once in … hmmm, I can’t remember how long. Keep your flames down eh?

    “That is not the finding. 112P6 was invoked. The problem, as noted by the district court judge and illustrated by me upthread two or three times (at least) is that the “structure” provided in the specification is effectively a reiteration of the vague (but plainly invocative) language in the claim.”

    Idk about that MM, the ruling is imo (they are quite vague at particularly this point in the decision so it has to be an opinion) 112 6th was not properly invoked, not necessarily that it was nor it wasn’t, but that the improperness of the invocation (you could I suppose interpret this instead as the improperness of the structure) leaves the claim wanting for definiteness.

  7. >X means for [performing function Y]
    >In the spec, which one is sufficient, and which one is not:

    >iii) An X microprocessor programmed to perform function Y. IOE, function Y can be
    >implemented using algorithm A which includes steps 1, 2, and 3. IAE, function Y
    >can be implemented using algorithm B which includes steps 4, 5, and 6.

    The software M+F cases I’ve read in the last year or so lead me to say that only option III suffices — regardless of how simple function Y is. I say this because arguments by patentees along the lines of “no additional structure or [no structure at all] is necessary because a PHOSITA would understand how to write code to perform the function” haven’t seemed to work at the Federal Circuit.

    But is there a “Dossel exception” for really simple functions?

    Suppose function Y=”computing a sum from two addends”. Would disclosure of a generic processor along with a statement in the spec that “known algorithms can be implemented on the processor to compute a sum from two addends” (ie, the statement in Dossel) be enough structure? How about a statement that “a PHOSITA would know how to program the processor to compute a sum from two addends”? [Second statement is similar in spirit, but doesn't have the possibly-magic-word algorithm.]

  8. “MM: your response indicates that you’ve missed the entire rest of the thread. ”

    Maybe you should read the entire thread to see how much of it I’ve “missed.”

    “As for my use of the term “invoking”, I think it’s appropriate here because the finding in this case is supposedly that the “means” claim fails to even _invoke_ 112 6th because there is no corresponding disclosure of structure.”

    That is not the finding. 112P6 was invoked. The problem, as noted by the district court judge and illustrated by me upthread two or three times (at least) is that the “structure” provided in the specification is effectively a reiteration of the vague (but plainly invocative) language in the claim.

  9. MM: your response indicates that you’ve missed the entire rest of the thread. A claim to “microprocessor for performing X” generally won’t get shot down by enablement or obviousness – hence the need for some other way to torpedo the claim, and hence the “result-oriented decision” in this case as per Step Back (whose post I generally agree with).

    As for my use of the term “invoking”, I think it’s appropriate here because the finding in this case is supposedly that the “means” claim fails to even _invoke_ 112 6th because there is no corresponding disclosure of structure. Thus, the “means” recitation in the claim is simply interpreted on its face, with the meaning of “means” in plain language, and is therefore indefinite.

    And there’s no “chicken little” going on, at least as far as I’m concerned; this is just a logic exercise.

  10. “What is worthwhile about never using M+F claims?”

    The problem with M+F is that they are a PITA to deal with, as the multitude of these cases bear out.

    I would rather find broad claim terms that acheive the same purpose.

  11. step back:
    >>The issue is one of sufficiency of disclosure. Is it sufficient to be enabling? Is it sufficient to disclose the best mode if there was one? If yes to both questions, then what is the court’s problem? Didn’t the inventor do everything he (they) was supposed to by law?>>This is a court that wanted to have a particular outcome irrespective of how illogical their legal and technological analysis are.<<<<

    Evidence to support this assertion?

  12. pds,

    What is worthwhile about never using M+F claims?

    If done properly, an M+F claim can provide the benefit of allowing equivalents apart from those provided by the DOE.

    One shouldn’t ONLY draft M+F claims.

    For software inventions the trick is to write several different types of claims (e.g., method, Beauregard, M+F).

  13. “And avoid M+F claims unless the client demands that you do so. And keep that demand in writing in the file.”

    Wow …. MM says something worthwhile now twice in about a month.

    I hope the aliens never return the original.

  14. “But if it’s not a M+F claim, who cares if it’s sufficiently “structure” or not? Answer: nobody.”

    I think huh’s statement should read that “if disclosing a microprocessor in the spec isn’t ‘structure’ for purposes of -invoking-_satisfying_the_corresponding_structure_requirement_ of 112 6th, then …”

  15. “the issue here is that if disclosing a microprocessor in the spec isn’t “structure” for the purposes of invoking 112 6th, then it wouldn’t be structure when actually _written into the claim_ either.”

    But if it’s not a M+F claim, who cares if it’s sufficiently “structure” or not? Answer: nobody. The term should simply be construed on its face and the claim should die, in that case, for lack of enablement or obviousness.

    It’s always tempting to go chicken little over stuff like this. Resist the temptation. And avoid M+F claims unless the client demands that you do so. And keep that demand in writing in the file.

  16. Thanks huh (Apr 25, 2008 at 08:05 AM) for articulating the issue.

    Assume that a processor configured to perform a function (in the spec) is not sufficient corresponding structure for a means plus function element.

    If so, then a claim that recites the same processor likewise lacks sufficient structure. Under Mas-Hamilton Group Inc. and other cases, the processor element is interpreted under 112p6. And, now, if there isn’t an algorithm, BAM! Indefiniteness. It isn’t a stretch to apply this reasoning to computer-readable medium claims too. Hence the slippery slope …

    Anyone who doesn’t care about this opinion because it’s a mpf claim should care about it if they have ever wrote a computer-readable medium claim.

  17. Dear step back,

    Re:
    “This is a court that wanted to have a particular outcome irrespective of how illogical their legal and technological analysis are.”

    I don’t know how common this fitting description of your example is used, i.e., a “Result Oriented Decision,” or how frequent such abominations occur. Of course, even one is too many, and, like a plague, it seems to be a growing threat to our justice system – I just read about another horrendous Result Oriented Decision posted by none other than “Oliver Wendell Holmes, Jr. The Great Dissenter” on this thread:

    link to patentlyo.com

  18. e6k: As I’m sure you realize, the issue here is that if disclosing a microprocessor in the spec isn’t “structure” for the purposes of invoking 112 6th, then it wouldn’t be structure when actually _written into the claim_ either. The implication is that if a M+F software claim doesn’t invoke 112 6th with a processor disclosure, then a processor claim of the same form is, by the same reasoning, indefinite.

    So, if we can show that a corresponding processor claim _is definite_, then it logically follows that the M+F claim invokes 112 6th, and is definite. Proof by contradiction.

  19. Where to begin? Where to begin unraveling the nonsensical Gordian knot weaved both by the court and the upthread commenters?

    Suppose there were only 3 known ways of performing a given function F and the specification listed all 3 ways. Then under 35 USC 112 P6, the M+F element in a “combination” claim must “cover” all 3 ways. There is nothing in the statute that says M+F “must” be narrower than all the known ways (structures) for performing the stated function.

    Moreover, the claimed invention is a “combination” of elements and not the one M+F element taken by itself. There was no pre-existing law to limit the scope of every element to only that which is disclosed.

    Moreover, a micro”processor” is not a micro”computer” and neither “micro” is a general purpose of any size computer. There is no industry definition of “general” purpose when it comes to computers unless you mean to say that Intel and Microsoft have a court-granted monopoly over the whole field of computational machines.

    This is a court that wanted to have a particular outcome irrespective of how illogical their legal and technological analysis are. A microprocessor is a structure. A programmed microprocessor (didn’t they mean microcomputer?) is a structure. The issue is one of sufficiency of disclosure. Is it sufficient to be enabling? Is it sufficient to disclose the best mode if there was one? If yes to both questions, then what is the court’s problem? Didn’t the inventor do everything he (they) was supposed to by law? Where in the statute is their a requirement for disclosing a specific computer program to the exclusion of all other readily available ways of achieving the function? And if the claim drafter had called it a “database editor” instead of a database editing “means”, would that “form over substance” swap have changed the whole of what the inventor disclosed to the world? And what if one found in the prior art a database editing means of different “structure”? Would that anticipate the M+F element? Isn’t the court trying to have it both ways as long as the inventor gets the mine shaft?

  20. “”microprocessor for performing X” is not indefinite regardless of whether 112 6th is invoked”

    That would be correct, but it certainly is not definitely limiting enough to invoke 112 6th. Notice that “definitely” is being used differently than “definiteness” is in a 112 2nd context. Huh, you have presumed that a limitation stating a “microprocessor for performing X” was in the claims in the case under discussion. There is no such limitation. There is simply a “means” limitation. That is why the case was rendered indefinite. There is no structure in the claim because 112 6th never was invoked properly and the claim does not have structure without invoking 112 6th.

    As a side issue you might be interested in, there are arguments above stating that 112 6th should be necessarily invoked because the claim would be indefinite without invoking it (and that’s what current policy is to do in such situations) but, according to this court the failure to supply an algorithm preempts that consideration and causes the invocation to fail. After the invocation fails, then the claim is indefinite because it has not structure.

    “A finding of indefiniteness indicates that an artisan would not know the metes and bounds of the claim – plainly false in this case.” No, without the “means” limitations properly invoking 112 6th an artisan would not know the metes and bounds and such is the courts correct ruling.

    If you still cannot understand, try reading the page 10? Aristo quote I think it was again in the morning.

  21. “”microprocessor for performing X” is not indefinite regardless of whether 112 6th is invoked”

    That would be correct, but it certainly is not definitely limiting enough to invoke 112 6th. Notice that “definitely” is being used differently than “definiteness” is in a 112 2nd context. Huh, you have presumed that a limitation stating a “microprocessor for performing X” was in the claims in the case under discussion. There is no such limitation. There is simply a “means” limitation. That is why the case was rendered indefinite. There is no structure in the claim because 112 6th never was invoked properly and the claim does not have structure without invoking 112 6th.

    As a side issue you might be interested in, there are arguments above stating that 112 6th should be necessarily invoked because the claim would be indefinite without invoking it (and that’s what current policy is to do in such situations) but, according to this court the failure to supply an algorithm preempts that consideration and causes the invocation to fail. After the invocation fails, then the claim is indefinite because it has not structure.

    “A finding of indefiniteness indicates that an artisan would not know the metes and bounds of the claim – plainly false in this case.” No, without the “means” limitations properly invoking 112 6th an artisan would not know the metes and bounds and such is the courts correct ruling.

    If you still cannot understand, try reading the page 10? Aristo quote I think it was again in the morning.

  22. Wasn’t me wrote:

    “Your desire for this to be a bright line test reveals a fundamental policy disagreement you have with the creators of paragraph 6. You misunderstand what policy they were going after (you think it’s a gift to lazy prosecutors) and then you are frustrated when the implementation of it does not support what you thought was the policy (you ask why don’t they just give them a “check box” to invoke the gift). Rather than just expressing frustration with the state of things, maybe you should check your premises and figure out what they were really trying to do with paragraph 6.”

    It has happened, and I am sure it will continue. I have responded to more than one Office Action in which the Examiner “required” that I identify which statements were to be interpreted under 35 U.S.C. 112, 6, and “required” that I identify the corresponding structure.

    TMP

  23. Overall, my argument (tying in policy, law, etc.) is: legally speaking, these kinds of claims aren’t indefinite, but these claims aren’t good from a policy standpoint either, so you’ve got to find some way to torpedo them. I’ve outlined my arguments above, but a bullet recap:
    - Enablement should work only rarely, because programmers are good at their jobs.
    - Novelty/obviousness depends on what’s been done before (it might be easy but never actually done, or no record of it being done), as well as the novelty of the function. It can be novel and enabled without disclosure, provided that the function is novel, but programming the function is easy.
    - Subject matter is easy to satisfy with magic words.

  24. huh- I see that argument, and I don’t disagree with it. I gamed the system, as you may note, by assuming we were already in paragraph 6 land. I was just shorthanding a few arguments that had been made against e6k together, but hadn’t read your argument closely to see the difference you were making.

    I don’t see how your situation is important in very many instances, because again it can’t really be the point of novelty. If it’s structure, it’s not paragraph 6, but that’s only true because it’s known in the art, so it can’t be the point of novelty. If it’s not structure, it’s in paragraph 6, and is indefinite unless it appears in the spec – regardless of the state of “general knowledge” on the point.

    I think this all works together OK, I guess I don’t see a policy problem.

  25. Wasn’t me: I think we’re crossing lines somewhere. The court in Mas-Hamilton uses the concept of “generally understood” to determine if a “lever moving element” is a structure, and thus, whether that language invokves 112 6th. Their finding was that since there was no general understanding in the art of the structure of a “lever moving element”, the claim was in M+F form even absent actual “means/step for” language. My comment to ‘me’ is that, if “algorithm A” has a generally understood structural meaning, then recitation of the name alone, absent specific steps, should be good enough under the same reasoning. This doesn’t really have anything to do with e6k’s comment.

    On the other hand, my comment to e6k is that “microprocessor for performing X” is not indefinite regardless of whether 112 6th is invoked. A finding of indefiniteness indicates that an artisan would not know the metes and bounds of the claim – plainly false in this case. “A microprocessor for performing X” is any microprocessor that performs X, by any algorithm. Broad, not indefinite.

    Now, as I argue above, I don’t think that claims like this are at all desirable from a policy standpoint, for reasons I’ve outlined. But, as you say, this is law, not “real world” or policy.

  26. George, since you appear to have all the answers, if a microprocessor, computer, or computer readable medium elements without an underlying algorithm is not sufficient structure for definiteness purposes, how could it be sufficient structure to avoid being pushed into a 112p6 analysis under the older caselaw?

  27. huh, me, etc:
    I’m in the uncomfortable position of agreeing with e6k here. He is describing the situation correctly.

    Your “generally understood” meaning fails also. First, once you’re in paragraph 6, the specification is invoked. If there are three ways to do it, and you list two in the specification, the third is excluded from the claim, even if the third is “generally understood”. As e6k said, this isn’t the “real world”. Secondly, as you said, if you can really manage to leverage something into the claim that you don’t have to describe (which you shouldn’t succeed at), clearly that cannot be your point of novelty and you’re not going to be fighting with the Examiner over it anyway.

  28. me: in that case, I think Mas-Hamilton may apply – specifically, does the algorithm have a “generally understood” meaning in the art? If “algorithm A” was a particular type of binary search algorithm, for instance, it seems to me that (ii) might just pass muster. Unfortunately, that runs you right into 102/103 issues.

  29. “Thus, while we’re technically in the MPF context, it really is just an ordinary indefiniteness holding. The claim is indefinite, end of story.”

    I agree.

  30. “In the real world … bl ah bl ah”

    And in the real world 99.9% of all of the stuff you’re patenting is obvious. So what? This is legal world, you simply must remember that. But anyway, your point is not what we’re trying to say. What we are trying to say is that for purposes of being sufficient structure for a wanna be 112 6th invoking “means” in a claim to refer back to in the spec “a microprocessor programmed to perform function A” is not a definitely limiting structure as is required by 112 6th in order for it to be properly invoked. Thus, “a microprocessor programed to perform function A” in the spec is not limiting enough for it 112 6th to have been properly invoked.

    De ar go d “me” just say in the claim “a means for performing function Y”, then in the spec say an X microprocessor programmed to perform algorithm A then state the algorithm including function Y being the result. Why do you guys make this so hard? btw, if you put “a microprocessor for performing function Y” in the claim then the claim has structure in it already and likely doesn’t even invoke 112 6th. Then again, any generic microprocessor likely 102′s it.

  31. “As has been said, 112 6th requires merely structure in order to be definite.”

    Yes folks keep saying that but the CAFC inteprets the statute. It should come as no surprise that the CAFC has held that the “structure” in the specification has to be meaningfully more structural than the means recited in the claim. i.e., you can’t recite a “fastener means” and then refer in the specification to “an attacher.” Please try to understand the logic behind this requirement. It’s straightforward.

    Likewise, you can’t recite a “data editing means” and then refer to “a microprocessor” without anything more. You might as well refer to “a person” as the structure. It is not at all *clear* what the microprocessor is doing. And my gosh it must be doing something truly fabulous here to warrant the grant of a patent.

    Right?

  32. Except really MM this case boils down to that MPF language exists in the claim but it doesn’t “properly” exist. Thus, while we’re technically in the MPF context, it really is just an ordinary indefiniteness holding. The claim is indefinite, end of story.

  33. Huge Ego: “Then in what terms should a software M+F claim be described?”

    Disclose an algorithm for embeding the indices. Mention that the microprocessor is to be programmed with suitable code for executing the algorithm.

    Don’t just say: “Fig. XX shows a microprocessor for embeding indices.”

  34. “As per usual trying to explain complex things in terms of the simple for the resolution of problems that apparently have their basis deep within the complexities has escaped my ability.”

    Yup. That is the usual.

  35. me: I don’t see a difference between (ii) and (iii), unless “algorithm A” is a name of an algorithm, and not an actual description of one.

    e6k: I think the problem here is that indefiniteness doesn’t work regardless of whether it invokes 112 6th. A claim to a “microprocessor for doing X” clearly delineates the boundries of the invention – any microprocessor that does X, using any algorithm. It’s a broad mother of a claim, but broad != indefinite.

  36. George says: “The holding appears to be that the corresponding structure for a software M+F claim cannot be described merely in functional terms.”

    Then in what terms should a software M+F claim be described?

  37. George and e6k,

    You’re saying that “a microprocessor programmed to perform function A” is not a structure because function A is a function? In the real world, “a microprocessor programmed to perform function A” is a structure. Your logic would mean that the microprocessor’s task must be disclosed at the source code level, without any generalization, else it would be disclosed as a “function.”

  38. me:

    The court didn’t give much guidance as to how to write software M+F claims that pass muster. The holding appears to be that the corresponding structure for a software M+F claim cannot be described merely in functional terms.

  39. Malcolm: “How tough is it to write ‘means for’?
    Seriously.”

    “This” is tough – meaning this area of the law is tough. Why are you attacking me with cheap out of context snippets, when all you have is clearly just your desire for the way things should be?

    Using magic words is not required by the law, and neither does using the magic words automatically invoke the law. That’s the well settled state of things, despite your clear desire to go to a bright line test. Even the PTO acknowledges this, for example:

    link to uspto.gov

    Your desire for this to be a bright line test reveals a fundamental policy disagreement you have with the creators of paragraph 6. You misunderstand what policy they were going after (you think it’s a gift to lazy prosecutors) and then you are frustrated when the implementation of it does not support what you thought was the policy (you ask why don’t they just give them a “check box” to invoke the gift). Rather than just expressing frustration with the state of things, maybe you should check your premises and figure out what they were really trying to do with paragraph 6.

  40. Okay, I need to get this clear. Let’s assume this is a software M+F claim element:

    X means for [performing function Y]

    In the spec, which one is sufficient, and which one is not:

    i) An X microprocessor programmed to perform function Y.

    ii) An X microprocessor programmed to perform function Y. IOE, function Y can be implemented using algorithm A. IAE, function Y can be implemented using algorithm B.

    iii) An X microprocessor programmed to perform function Y. IOE, function Y can be implemented using algorithm A which includes steps 1, 2, and 3. IAE, function Y can be implemented using algorithm B which includes steps 4, 5, and 6.

  41. e6k,

    Yes… you explained it well.

    What it comes down to is that the asserted corresponding structure, i.e., the microprocessor programmed to embed indices, was described in functional terms and, to invoke 112, para 6, structure must be disclosed.

  42. “Would there be a similar understanding of “a microprocessor performing function A”?”

    apparently, it cannot have any structural meaning as a matter of law

  43. Sure, and I bet there’s tons of art published right before the critical date of this patent saying “It’s impossible to send information for a hierarchical database through the airwaves. Don’t even try it!”
    <<<<

    Sure, you can transmit a hierarchical database. But how does transforming a data set into a hierarchical database improve the transmission time of the data set. It doesn’t; you are adding overhead (both in metadata and in frequently updating the metadata). The claim then goes on to disclose other elements, which *gasp* depend on the database element.

    Don’t get so hung up on the fact that you know what a hierarchical database is.

  44. As per usual trying to explain complex things in terms of the simple for the resolution of problems that apparently have their basis deep within the complexities has escaped my ability. So, let’s delve deeper shall we? Look at page 29 of the decision, bottom of page. Read the entire quote from Aristo, then understanding that the entire “structure” of the means language in the claim amounts to nothing but functional language because it refers to a microprocessor that does not have an associated algorithm and thus the words “means” simply mean “means” instead of a 112 6th invoking “means” that refers to the structure of the application. Because disclosing a computer as designated to perform a particular function does not limit the scope of the claim to the corresponding structure, material, or acts that perform the function as required by 112 6th the computer in the spec cannot be considered a complete structure for “means” to refer back to. Thus, this claim cannot have it’s “means” terms construed as being 112 6th invoking. Now, after having realized this, they state that the court does not impose a lofty standard for indefiniteness and since these claims are already quite vague one of ordinary skill simply cannot perceive the bounds of the invention.

    So to recap, the “means” limitation provides no structure IN THE CLAIM because 112 6th is not properly invoked because there was no algorithm (as is required to properly invoke 112 6th in these microprocessor cases). And without that structure the claim is indefinite in this specific case. This is not saying that claims would always be indefinite if this happened, but just that in that case the claims were. Though I suspect this could happen often because people rely on the “means” to provide needed structure to provide definiteness in many instances. Ok? Understand? Are we all clear?

    Just curious, Do you have several claims ready to be invalidated because of this?

  45. Er, should have read George’s comment first, as he covers the same issue more succinctly than I did.

    As for holding “microprocessor performing function A” as being the function, we then turn to the question I quote from Mas-Hamilton: whether the element has “been shown to have a generally understood structural meaning in the art”. We can say something like “telephone” because there’s a “generally understood structural meaning in the art” of what a telephone is. Would there be a similar understanding of “a microprocessor performing function A”?

  46. Anon: don’t get me wrong, I don’t necessarily think that using indefinite in this case is really a legally tenable position. As has been said, 112 6th requires merely structure in order to be definite.

    As for the case where the purpose is not obvious, note that the Applicant doesn’t actually have to reduce to practice to get a patent issued. Since the purpose is novel and unobvious, it will pass 102/103. Enablement is a given, since given the disclosure of the purpose, any programmer would be able to write the corresponding software. It will pass 101 as statutory subject matter provided it’s claimed with the right “magic words” – but it can still be claimed broadly enough (processor executing instructions for performing the method …) to capture pretty much every possible practical application without technically capturing every possible application (it can still be done by hand …).

    The problem here, I think, really revolves around the term “structure” as applied to software. Unlike a conventional machine, a computer system running a program isn’t physically different from a computer system not running the program (unless we get into the gritty details of electron positions and whatnot). As such, treating, e.g., “microprocessor”, as structure is not useful in a software context. Stated another way, all of the “structure” in a software claim is actually virtual (i.e., the algorithm), and really has nothing to do with the physical hardware that the software is running on. As such, claiming a “processor for doing X” is similar to saying “a machine for doing X” – is there really any structure there?

    While the “indefinite” grounds might be shaky, I think the above problem is what this court (and the court in Aristocrat, and in WMS gaming) was considering. Remember, the way software claims work now is that a computer system running the software is considered to be a distinct system from a computer system that is not running the software – as stated in WMS Gaming, “a general purpose computer programmed to carry out a particular algorithm creates a “new machine” because a general purpose computer ‘in effect becomes a special purpose computer once it is programmed to perform particular functions pursuant to instructions from program software.’”

  47. George,

    “So we cannot correspond the ‘means-plus-function’ claims with any corresponding structure…”

    Sure you can. If the disclosed structure is “a microprocessor performing function A,” then you can make a determination that the “microprocessor performing function A” is the corresponding structure and that a microprocessor performing function B is different.

  48. “There is nothing in the statement of the statute that indicates you need to use magic words to invoke it. Because this is tough,”

    How tough is it to write “means for”?

    Seriously.

  49. Anon,

    The “structure” would be the microprocessor along with a stored set of instructions which would execute the algorithm.

    If the microprocessor were to be programmed using a algorithm A it would be “structurally” different from one being programmed using algorithm B.

    The trouble with the ’505 patent is that it doesn’t disclose ANY algorithm for embedding indices. So we cannot correspond the “means-plus-function” claims with any corresponding structure, as required under 112, para 6.

  50. Anon,

    The “structure” would be the microprocessor along with a stored set of instructions which would execute the algorithm.

    If the microprocessor were to be programmed using a algorithm A it would be “structurally” different from one being programmed using algorithm B.

    The trouble with the ’505 patent is that it doesn’t disclose ANY algorithm for embedding indices. So we cannot correspond the “means-plus-function” claims with any corresponding structure, as required under 112, para 6.

  51. David, means claims are very narrow. Nevertheless, there are specific situations where they are useful.

  52. “Is there really anyone that cannot tell whether an accused device is using a microprocessor? That is what a holding of indefiniteness means.”

    News flash: determining indefiniteness in the M+F context is *different* than determining indefiniteness in the “regular” context. The reasons for that difference have been discussed upthread. This is not new.

  53. Malcolm: “Remember that 112P6 is a “bonus” for prosecutors. It’s a gift. It’s a short cut for people who, for whatever reason, don’t like to write or read long claims. In all other respects, means + fxn claims STINK at performing the job that claims are intended to do: put the public on notice as to what the claims actually cover.
    The CAFC is right to hold prosecutors feet to the fire. The worst cases in this area are those where the CAFC has allowed applicants to invoke 112P6 without using “means for” language. That case law is a travesty and embarassment, right up there with the preamble garbage. The CAFC should fix it.”

    112P6 is not a bonus or a gift, it is a statute. There is nothing in the statement of the statute that indicates you need to use magic words to invoke it. Because this is tough, Examiners prefer to treat it like a secondhand, almost-law, that really shouldn’t be there. That’s not the job of the Examiner – the job of the Examiner is to apply the law, not advocate for positions of the law except as a private citizen outside of the scope of their employment.

    My opinion is that 112P6 covers an important gap in claiming procedure, and apparently Congress agrees. Opinions can differ and everyone has one, but an Examiner should not substantively attack the law through actions during examination just because he/she thinks it’s not a good law s/he’s being asked to apply. Such an Examiner should stop being a Federal employee and go work at a law school and write papers advocating their position, not sabotage the examination of applications at the patent office.

  54. David, as I recall, the CAFC has held that equivalents of structures in M+F claims must perform the IDENTICAL function and achieve the IDENTICAL result, leaving the “substantially similar way” prong remaining as a litigation playground.

  55. David French writes:

    Here’s a sample of a claim that relies heavily on the use of “means”:

    (WO/2004/074681) METHOD OF CONTROLLING AERODYNAMIC LOAD OF A WIND TURBINE BASED ON LOCAL BLADE FLOW MEASUREMENT

    16. A wind turbine according to claim 15, comprising one or more of the following features: storing means for storing data, transferring means for transferring data to analysing means, analysing means, such as processor means, for analysing the data stored in the storing means; pitching means being adapted to set the pitch of each blade and being in data communication with the analysing means, means for transmitting data to other wind turbines down stream, means comprising a membrane or a pitot tube for recording pressure on the front and/or of each blade and/or means comprising piezoelectric sensors for recording the mechanical load such as the torque measured at the connection between the blade and the hub.

    I have only just recently been alerted to the danger of using the “means” claim structure in United States patent specifications.
    When section 35 USC 112(6) is applied to claim interpretation, the claim element so framed does not cover all ways of performing the function. It only covers ways as described in the specification and their equivalents.

    Many people mistakenly think that a “means-for” claim element is necessarily very broad, but this is not correct. Such a claim can be narrow according to what was disclosed in the specification. If the specification discloses only a single way to perform a function, then a means-for claim element will cover only that disclosed structure and its equivalents.

    The words “and its equivalents” seems to give the claim broad scope. However, under US patent law the “triple identity” test (Graver Tank & Manufacturing Co. v. Linde Air Products Co., (1950)) provides that something is deemed equivalent if:

    1. It performs substantially the same function
    2. in substantially the same way
    3. to yield substantially the same result.

    Under the second test (Warner-Jenkinson Co. v. Hilton Davis Chem. Co. (1997)) something is deemed equivalent if there is only an “insubstantial change” between each of the elements of the accused device or process and each of the elements of the patent claim.”

    Both of these tests are very limiting. In particular, element 2 of the triple identity test, namely that the equivalent must perform in substantially the same way, is very restrictive.

    If I am right in my interpretation, then I’m going to avoid using the “means” language in future applications.

    I would appreciate it if anyone would tell me whether I am on the right or wrong track.

  56. George,

    And why SHOULD that be a problem? The statute does not require that a means-plus-function element be “construed to cover the corresponding ALGORITHM…disclosed in the specification,” only “the corresponding structure…disclosed in the specification.” Is there really anyone that cannot tell whether an accused device is using a microprocessor? That is what a holding of indefiniteness means.

  57. So assuming that “a microprocessor programmed to sort the array using a bubble sort” was disclosed, a “means for sorting an array” would be definite but a “means for bubble sorting an array” wouldn’t?

  58. John: “Unless your spec discloses algorithms at the source code level, then any description of an algorithm almost certainly includes functions and functional descriptions of what the source code does. An algorithm is simply any set of functions a microprocessor is performing.”
    >>>>>>>>

    Well, all I think that is necessary to meet 112, para 6, is to specify the algorithm that you use. So, if one wanted to sort an array, it might be acceptable to merely disclose “a microprocessor programmed to sort the array using a bubble sort”. If the algorithm is not well known, I believe that all that would be necessary would be to describe it in a fairly high-level fashion. But in the ’505 patent there is simply no disclosure at all as to the specific algorithm to be used. And therein lies the problem.

  59. “No. It is a reasonable test for determining whether a policy is reasonable. And you flunked. Now, beat it, troll.”

    Sounds like you drank a dud cup of coffee this morning, George. You seem to be confusing a test for determining “the best policy” with a test for determining whether a policy is “reasonable.” It’s an important distinction, but also an incredibly boring and irrelevant one in this context.

  60. Remember that 112P6 is a “bonus” for prosecutors. It’s a gift. It’s a short cut for people who, for whatever reason, don’t like to write or read long claims. In all other respects, means + fxn claims STINK at performing the job that claims are intended to do: put the public on notice as to what the claims actually cover.

    The CAFC is right to hold prosecutors feet to the fire. The worst cases in this area are those where the CAFC has allowed applicants to invoke 112P6 without using “means for” language. That case law is a travesty and embarassment, right up there with the preamble garbage. The CAFC should fix it.

    In this case, the CAFC is basically saying that if you put a “fastening means for blah blah blah” in your claim, you better recite more specific structure than “an attacher” in your specification.

    In the context of the invention of the patent at issue here, “a database editing means” and “a microprocessor” are indistinguishable. The M+F paradigm should never have been invoked. Nothing was gained by doing so.

  61. “Your solution is not reasonably tailored to solve the alleged problem.”

    Malcolm: “…That’s an absurdly strict test for determining whether a policy is reasonable.”

    <<<<

    No. It is a reasonable test for determining whether a policy is reasonable. And you flunked.

    Now, beat it, troll.

  62. huh,

    You forgot about obviousness. Notice the assumption in your analysis: “IF given the purpose of the software, any programmer would be able to do the same…”

    What if the purpose was not obvious? The claim should be held indefinite?

  63. “Databases add overhead. So in this case, structuring the data set as a hierarchial database would not have been obvious because it would just make that “something big” even bigger.”

    Sure, and I bet there’s tons of art published right before the critical date of this patent saying “It’s impossible to send information for a hierarchical database through the airwaves. Don’t even try it!”

    Seriously, let’s assume your premise is true for arguments sake. Am I to believe, then, that if I look at the claims of this patent I will find a detailed description of, e.g., the algorithm used to encode the database to make it “fit” in the available bandwidth?

  64. “Your solution is not reasonably tailored to solve the alleged problem.”

    Why? Because other solutions also address the problem? That’s an absurdly strict test for determining whether a policy is reasonable. The fact remains: it works. Computer type prosecutors appear to have relied heavily on means plus fxn claiming. One down. Hundreds more to go.

  65. If we’re talking about policy, treating these sorts of claims under enablement is unusable. For the vast majority of programs, if you gave a programmer a description of what the program was supposed to do, they would be able to bang out a working program pretty easily – and if not ‘easily’, then certainly within the lines of “undue experimentation”. We must remember, of course, that the bar for “undue” is based on what is regularly undertaken in the art; given that software is essentially a content-creation business, and software regularly involves test trials and debugging, the level of “undue” here is quite high.

    Novelty can be difficult to rely on also – consider the case of a program that’s well within the ability of one of ordinary skill to write, but nobody has actually done it yet. Legally speaking, this deserves a patent, but as a policy matter do you really want patents issuing on these sorts of “inventions”? The point of patents is to promote innovation by granting limited monopolies in exchange for disclosure – as a policy matter, patents are only successful instruments when the value of the disclosure equals or exceed the value of the monopoly. How often is that true in software patents, when we consider that, if given the purpose of the software, any programmer would be able to do the same in many cases?

  66. e#6k says: >>algorithms are lists of instructions, prodecures or formula, not functional language, and not functions.

    George says: I don’t see where any algorithm was disclosed in the ’505 patent.<<<

    Unless your spec discloses algorithms at the source code level, then any description of an algorithm almost certainly includes functions and functional descriptions of what the source code does. An algorithm is simply any set of functions a microprocessor is performing.

    The decision here and in Aristocrat treats computer-implemented “means” different than other types of “means.” Apparently, we can disclose a “fastener” as a “means for fastening,” but we can’t disclose “a microprocessor programmed to generate indices” as a “means for generating indices.”

  67. Malcom says: >>”What … policy reason is there to hold that a microprocessor is not a “structure” under 112, 6 regardless of whether an algorithm is disclosed?”

    That’s easy. There are too many bogus computer-implemented devices and methods out there. This is one way to weed some of the more ridiculous ones out.<<

    Your solution is not reasonably tailored to solve the alleged problem. Holding means-plus-function claims indefinite where the disclosed structure is a microprocessor does not solve the problem of “bogus computer-implemented devices and methods.” If they are “bogus” because they aren’t novel, then section 102 and 103 can take care of that problem. If they are “bogus” because they aren’t enabled, then section 112,1 can take care of that problem.

  68. e#6k says: >>They didn’t hold it wasn’t a structure, they held that the spec disclosed little about the structure and that it wasn’t a special purpose structure (computer) because there was no definite algorithm provided.<<

    The court held the claims “indefinite.” In re Donaldson says that a means-plus-function element is indefinite if there is NO structure disclosed for performing the claimed function, because in that situation, the claim cannot be “construed to cover the corresponding structure…described in the specification,” as the statute requires. However, there is no reason to hold a means-plus-function claim INDEFINITE if there IS a structure disclosed for performing the claimed function, such as a microprocessor performing the claimed functions. There is no requirement in 112,6 or anywhere else for a (more) SPECIFIC structure. There may be other grounds for holding the claim invalid because of its breadth–such as lack of enablement, anticipation, or obviousness, but not indefiniteness.

  69. e6k, wrote “No, learn to tell the difference between the different paras of 112. Dear god I hope this isn’t a lawyer.”

    I *am* a lawyer and apparently one needs to know the law, not just the MPEP, to understand what I was talking about. “huh” clearly understood: courts sometimes construe claim elements that lack the phrase “means for” under 112p6 because the claim element lacked sufficient structure.

    Thus, if a microprocessor, computer, or computer readable medium elements without an underlying algorithm is not sufficient structure for definiteness purposes, how could it be sufficient structure to avoid being pushed into a 112p6 analysis under the older caselaw?

    e6k, please use a different tone, especially when you might not be aware of the relevant caselaw.

  70. e6k, “means for” language isn’t necessary to invoke 112 6th. See Mas-Hamilton Group Inc. v. LaGard Inc. 48 USPQ2d 1010, 1016. The court found that the recited “lever moving element” discloses no structure because it “had not been shown to have a generally understood structural meaning in the art”. The pertinent question here is whether a “microprocessor” or “computer-readable medium” would escape the same analysis. “Computer-readable medium” is particularly troubling, because applicants can’t seem to agree on what exactly the medium encompasses – does it or does it not include transmission mediums? Signals? While the office policy on this may be clear, the opinion of the outside world is less so. The term “processor” is even worse – absent the “micro” limitation, a “processor” is identical in every way to a “means for processing”; anything that can process is, by definition, a “processor”.

    Software patents in general are a quagmire resulting from the fact that software per se is not patentable. Thus, we are stuck using the awkward mechanism of claiming software as a “component” of an otherwise well-known device.

  71. Malcolm Mooney,

    I don’t know how obvious the hierarchial database would have been.

    The patent seems to disclose a method for transmitting some data set over some limited and much smaller bandwidth. In other word, it is trying to fit something really big into something really small. Databases add overhead. So in this case, structuring the data set as a hierarchial database would not have been obvious because it would just make that “something big” even bigger. The obvious thing would have been to make that “something big” smaller. Granted, I try to give issued patents the benefit of the doubt instead of being such a smart guy. In any event, it’s just one element of the claim.

  72. I agree with the earlier comment – it’s very important for people to understand that, since the late ’90′s “means plus function” claims have been construed narrowly – they are typically limited to the corresponding structure disclosed in the specification with essentially no equivalents.

  73. I confess, I haven’t read the case.

    Judging from Denis’s extract, it seems like this is a case of inventing a method and then trying to claim the apparatus which implements the method. What’s wrong with a method claim?

    Also, not being an American I don’t know the history of 112p6, but I’d suggest that any legislation which says that “a means for connecting” shouldn’t be construed in the same way as “a connector” is simply bad law.

  74. Again, it’s not that it’s not “enough” structure it’s that the structure isn’t disclosed well enough. These are two separate concepts. Please learn to use your law school trained noggin.

    “”a microprocessor programmed to …”? Wouldn’t that claim be under 112p6 since it would lack any structure and thus would be interpreted under 112p6, just like “a mechanism configured to”?”

    “What if a claim recites “a computer readable medium including instructions for performing …”? Wouldn’t that claim likewise be subject to 112p6?”

    Um, no, you didn’t invoke 112p6 with “means for” etc. in either case. A microprocessor is a structure people. It’s only when the mircoprocessor is called a “means for” or similar in the claim and is only defined in the spec that a problem arises with this stuff we are discussing.

    “If so, would the microprocessor / medium claims require an algorithm to be definite?”

    No, learn to tell the difference between the different paras of 112.

    Dear god I hope this isn’t a lawyer.

  75. If a microprocessor is not sufficient structure …

    Then what if a claim recites “a microprocessor programmed to …”? Wouldn’t that claim be under 112p6 since it would lack any structure and thus would be interpreted under 112p6, just like “a mechanism configured to”?

    What if a claim recites “a computer readable medium including instructions for performing …”? Wouldn’t that claim likewise be subject to 112p6?

    If so, would the microprocessor / medium claims require an algorithm to be definite?

  76. “What … policy reason is there to hold that a microprocessor is not a “structure” under 112, 6 regardless of whether an algorithm is disclosed?”

    That’s easy. There are too many bogus computer-implemented devices and methods out there. This is one way to weed some of the more ridiculous ones out.

    Again, I ask: why should that lengthy clause even be in the claim? It surely adds nothing to “the invention.” It’s like a claim to a “new” food processor with a 25 word limitation going on and on about a “power source engagement means for connecting the processor to a power source.”

    Hierarchical database. Wow. When was the patent filed? 1969?

  77. n this case, structure WAS disclosed and an algorithm WAS disclosed<<<<<<<<<

    Anonymous,

    I don’t see where any algorithm was disclosed in the ’505 patent.

    And just disclosing a microprocessor that is programmed with some unknown algorithm for embedding indices doesn’t satisfy 112, para 6, IMHO.

  78. Anon, I’m not completly sure but I believe in the case they mean the microprocessor isn’t a specific structure, as in, it’s a general structure with specific software running on it. They didn’t hold it wasn’t a structure, they held that the spec disclosed little about the structure and that it wasn’t a special purpose structure (computer) because there was no definite algorithm provided.

    Further, I’m not sure what cotton-candy world you’re living in, but an algorithm isn’t a function. From your friendly wiki: an algorithm is a type of effective method in which a definite list of well-defined instructions for completing a task … is used. From a dictionary it is: An algorithm (pronounced AL-go-rith-um) is a procedure or formula for solving a problem. The word derives from the name of the mathematician, Mohammed ibn-Musa al-Khwarizmi, who was part of the royal court in Baghdad and who lived from about 780 to 850. Al-Khwarizmi’s work is the likely source for the word algebra as well. So let’s be clear, algorithms are lists of instructions, prodecures or formula, not functional language, and not functions. So I guess the policy reason is that the courts are finally learning to look up definitions?

    There’s a lesson to be learned here. Don’t use functional language in your spec in leiu of an algorithm and expect your 112 6th programed general purpose computer claim to hold up. Seems self-evident to most people …

  79. Dennis’s description of this case omits significant facts. In this case, structure WAS disclosed and an algorithm WAS disclosed. It was a “microprocessor” programmed with software that generates and embeds indices, precisely as the claimed functions required. The court held that the claims were “indefinite” because a microprocessor was not a “structure.” It further held that the disclosed PROGRAMMED microprocessor was also not a “structure” because the “algorithm” was no more specific than the functions recited in the claim. What legal or policy reason is there to hold that a microprocessor is not a “structure” under 112, 6 regardless of whether an algorithm is disclosed? What legal or policy reason is there to further hold that an “algorithm” must be more specific than the claimed functions? The breadth of the resulting interpretation? Breadth is not controlled by the requirement for “definiteness,” but rather by the prior art and by enablement. There is no vagueness here, only breadth.

  80. ““database editing means, coupled to said one or more computer memory devices, for generating a hierarchically arranged set of indices for referencing data in said information database, including distinct indices for referencing distinct portions thereof, and for embedding said indices in said information database . . .”

    As of the filing date, there is absolutely nothing inventive about a hierarchical database in any context. That leaves the question: why even put this in the claim? i.e., why not claim the *invention*?

    It’s a rhetorical question, of course.

  81. Dennis,

    I don’t believe any competent practitioners have considered MPF claims to be broad since the late 90s. The CAFC limited their scope pretty quickly.

    Post-Festo, they received a bit of a resurgence as another source of equivalents.

    However, I have always believed that claims can be broadened by use of sufficiently functional language (fastener, barrier, heater, etc.).

Comments are closed.