« Amazin' Claimin' | Main | PCTs Down Under »

November 4, 2008

Divining Bilski

Bilski was nothing short of a jurisprudential obscenity, horrendous in its incoherence. To the degree it was coherent, it was ludicrous. That makes Bilski a definitive statement on the quality of the CAFC, most notably majority author Chief Judge Michel. Only Judges Newman and Rader had a clue. As Judge Newman put it:

This exclusion of process inventions is contrary to statute, contrary to precedent, and a negation of the constitutional mandate. Its impact on the future, as well as on the thousands of patents already granted, is unknown.

From one judge to another, that's harsh language. None more harsh.

Pundit readings of Bilski have been wildly divergent. Those who considered it "status quo," including this pundit, very likely got it wrong. Mea culpa. Those who considered Bilski obliterating software patents in toto also got it wrong. I hope. Upon more careful review and contemplation, the hideous truth of Bilski's corpse lies between the two camps.

The first Patent Prospector coverage of Bilski has a different set of quotes. Herein, extracts and alternate takes.

A claimed process is surely patent-eligible under § 101 if: (1) it is tied to a particular machine or apparatus, or (2) it transforms a particular article into a different state or thing.

What the CAFC referred to as "the machine-or-transformation test."

"Article" is a 13th century word, meaning, among other things, "an item of business; matter". Its dictionary context is so ambiguous as to make it a very poor choice of words without further definition. In other words, "article" was used by Judge Michel as a lexicographical straw man.

A claimed process is patent-eligible if it transforms an article into a different state or thing. This transformation must be central to the purpose of the claimed process. But the main aspect of the transformation test that requires clarification here is what sorts of things constitute "articles" such that their transformation is sufficient to impart patent-eligibility under § 101. It is virtually self-evident that a process for a chemical or physical transformation of physical objects or substances is patent-eligible subject matter.

Here's where the anti-software patent view comes home to roost.

The raw materials of many information-age processes, however, are electronic signals and electronically-manipulated data. And some so-called business methods, such as that claimed in the present case, involve the manipulation of even more abstract constructs such as legal obligations, organizational relationships, and business risks. Which, if any, of these processes qualify as a transformation or reduction of an article into a different state or thing constituting patent-eligible subject matter?

Our case law has taken a measured approach to this question, and we see no reason here to expand the boundaries of what constitutes patent-eligible transformations of articles.

For the majority, Benson was a touchstone in the majority, as to what was not patentable: something with "no utility other than operating on a digital computer," that is, software ipso facto.

[I]n Benson, the limitations tying the process to a computer were not actually limiting because the fundamental principle at issue, a particular algorithm, had no utility other than operating on a digital computer. Benson, 409 U.S. at 71-72. Thus, the claim's tie to a digital computer did not reduce the pre-emptive footprint of the claim since all uses of the algorithm were still covered by the claim.

Diehr was another touchstone, for what was patentable: computerized process control.

The process claimed in Diehr, for example, clearly met both criteria. The process operated on a computerized rubber curing apparatus and transformed raw, uncured rubber into molded, cured rubber products. Diehr, 450 U.S. at 184, 187.

Abele was the most illuminating example provided by the muddled majority, affording patentability, in the Bilski opinion, to software that represented a "physical object." Hence, software is patentable "so long as the claimed process is limited to a practical application... to transform specific data, ...to a visual depiction that represents specific physical objects or substances."

Our predecessor court's mixed result in Abele illustrates this point. There, we held unpatentable a broad independent claim reciting a process of graphically displaying variances of data from average values. Abele, 684 F.2d at 909. That claim did not specify any particular type or nature of data; nor did it specify how or from where the data was obtained or what the data represented. Id.; see also In re Meyer, 688 F.2d 789, 792-93 (CCPA 1982) (process claim involving undefined "complex system" and indeterminate "factors" drawn from unspecified "testing" not patent-eligible). In contrast, we held one of Abele's dependent claims to be drawn to patent-eligible subject matter where it specified that "said data is X-ray attenuation data produced in a two dimensional field by a computed tomography scanner." Abele, 684 F.2d at 908-09. This data clearly represented physical and tangible objects, namely the structure of bones, organs, and other body tissues. Thus, the transformation of that raw data into a particular visual depiction of a physical object on a display was sufficient to render that more narrowly-claimed process patent-eligible.

We further note for clarity that the electronic transformation of the data itself into a visual depiction in Abele was sufficient; the claim was not required to involve any transformation of the underlying physical object that the data represented... So long as the claimed process is limited to a practical application of a fundamental principle to transform specific data, and the claim is limited to a visual depiction that represents specific physical objects or substances, there is no danger that the scope of the claim would wholly pre-empt all uses of the principle.

Majority manipulation of Benson sucks oxygen from the idea that software can find succor in being able to particularize a machine, prong one of the process test.

The Court explicitly stated in Benson that "[t]ransformation and reduction of an article 'to a different state or thing' is the clue to the patentability of a process claim that does not include particular machines."

The court found State Street's "practical application" test "inadequate." And the physicality ("physical steps") of AT&T is bashed, in view of State Steet.

In AT&T, we rejected a "physical limitations" test and noted that "the mere fact that a claimed invention involves inputting numbers, calculating numbers, outputting numbers, and storing numbers, in and of itself, would not render it nonstatutory subject matter." 172 F.3d at 1359 (quoting State St., 149 F.3d at 1374).

The ruling leaves me wondering how software could ever meet the test of being tied to a "particular" machine, other than transforming something physical, as in computerized process control altering physical objects.

But the main aspect of the transformation test that requires clarification here is what sorts of things constitute "articles" such that their transformation is sufficient to impart patent-eligibility under § 101. It is virtually self-evident that a process for a chemical or physical transformation of physical objects or substances is patent-eligible subject matter.

Physicality itself is not patentable, unless tied to a machine.

The correct analysis is whether the claim meets the machine-or-transformation test, not whether it recites "physical steps." Even if it is true that Applicant's claim "can only be practiced by a series of physical acts" as they argue, see id. at 9, its clear failure to satisfy the machine-or-transformation test is fatal.

In my view now, nothing but uncertainty abounds. Even though a software process is necessarily tied to a machine, such as "a computer-implemented process," is it not patentable if it only transforms data which is not representative of a physical object? Does a computer-implemented process become patentable by transforming data representative of a physical object? Is software inherently patentable because its processes necessarily involve physical transformation of the disk drives that store data processed by software? Is a novel way of deleting files patentable, because graphical user interfaces depict file deletion as a trash can, which is a physical object?  Does any relation to a physical object make a software method patentable? Aren't electronic documents representative of physical documents, and, in fact can be used to create physical documents? Hence, is electronic document processing patentable subject matter? Are all database processing patents invalid, because their data does not necessarily represent physical objects? But what if the database is populated by data for physical goods, such as for an online vendor like Amazon?

The majestic Judge Newman, for every precedent cited, show Michel's analysis as facile and ill-considered. Judge Newman then played what should have been a trump card, but wasn't:

In Tilghman v. Proctor the Court considered a patent on a process for separating fats and oils, and held that the process was not restricted to any particular apparatus. The Court held that a process is an independent category of invention, and stated:

That a patent can be granted for a process, there can be no doubt. The patent law is not confined to new machines and new compositions of matter, but extends to any new and useful art or manufacture.

102 U.S. at 722; see also Corning v. Burden, 56 U.S. (15 How.) 252, 268 (1853) ("It is for the discovery or invention of some practical method or means of producing a beneficial result or effect, that a patent is granted, and not for the result or effect itself.")

Perhaps the worst is to be feared from Bilski. Certainly Judge Newman feared the worst from her colleagues' insanity in Bilski.

The court thus excludes many of the kinds of inventions that apply today's electronic and photonic technologies, as well as other processes that handle data and information in novel ways. Such processes have long been patent eligible, and contribute to the vigor and variety of today's Information Age.

The Court made clear that it was not barring patents on computer programs, and rejected the "argu[ment] that a process patent must either be tied to a particular machine or apparatus or must operate to change articles or materials to a 'different state or thing'" in order to satisfy Section 101. Id. Although my colleagues now describe these statements as "equivocal," maj. op. at 14, there is nothing equivocal about "We do not so hold." Benson, 409 U.S. at 71. Nonetheless, this court now so holds.

Until the shift represented by today's decision, statute and precedent have provided stability in the rapidly moving and commercially vibrant fields of the Information Age. Despite the economic importance of these interests, the consequences of our decision have not been considered. I don't know how much human creativity and commercial activity will be devalued by today's change in law; but neither do my colleagues.

We aren't told when, or if, software instructions implemented on a general purpose computer are deemed "tied" to a "particular machine," for if Alappat's guidance that software converts a general purpose computer into a special purpose machine remains applicable, there is no need for the present ruling. For the thousands of inventors who obtained patents under the court's now-discarded criteria, their property rights are now vulnerable.

Although this uncertainty may invite some to try their luck in court, the wider effect will be a disincentive to innovation-based commerce. For inventors, investors, competitors, and the public, the most grievous consequence is the effect on inventions not made or not developed because of uncertainty as to patent protection.

What does Bilski portend, now that the CAFC has told stare decisis to kiss its ass? From here, right now, it looks like pure chaos to me.

Posted by Patent Hawk at November 4, 2008 1:41 AM | § 101

Comments

Hawkie - I'll take a stab.
"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).
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?

Posted by: anonymousAgent at November 4, 2008 3:24 AM

No dice, a. Agent.

Transforming the actual memory won't suffice. Read Ex Parte Wasynczuk and Ex Parte Langmyr in light of Bilski majority and Comiskey...and Benson and Flook.

not patentable: method for a general purpose computer
patentable: same method for two general purpose computers
not patentable: method of operating a reentrant shift register to perform BCD (because the register is a general purpose computer!)
patentable: method of displaying data, where the data represents images of organs
not patentable: method of displaying data that works on all types of data (could even be the same method above w/o the arbitrary restriction on the type of data)

the list goes on...We're down the rabbit hole into Wonderland, where little girls are like serpents because they both eat eggs.

Posted by: Anon E. Mouse at November 4, 2008 6:28 AM

I agree with the Hawk's analysis.

The incoherence of the majority opinion is also apparent by their statement that field-of-use limitations are not enough to save a claim, while then referring to the field of use limitation in the dependent claim of Abele as an example of what suffices as a transformation. What the ??!?


Posted by: pikkumatti at November 4, 2008 8:08 AM

My reading of the case and Hawk's comments is that software instructions residing on a special purpose computer are not greatly endangered. If I am correct, router software residing on router hardware can still be protected.

What is strange, however, is that software performing exactly the same functions and resulting in exactly the same end is questionable if it is designed to operate on a general purpose computer and protected if it is designed to operate on a special purpose computer.

Posted by: dvan at November 4, 2008 9:32 AM

The problem is that "special purpose" and "general purpose" machines are not so well-defined.
In many cases, I can easily load, for example, game software, onto the router hardware - i.e. the router manufacturer may find it cheaper to configure the router with SOFTWARE and use a general purpose CPU - to the outside world, the blackbox router would look EXACTLY the same as a "dedicated hardware" router.
SO is "particular machine" meant to imply "dedicated machine" - i.e. with only a single purpose??
this would endanger many HARDWARE claims.

Posted by: anonymousAgent at November 4, 2008 9:52 AM

anonymousAgent said:

"In many cases, I can easily load, for example, game software, onto the router hardware - i.e. the router manufacturer may find it cheaper to configure the router with SOFTWARE and use a general purpose CPU - to the outside world, the blackbox router would look EXACTLY the same as a "dedicated hardware" router.
SO is "particular machine" meant to imply "dedicated machine" - i.e. with only a single purpose??
this would endanger many HARDWARE claims."

Agreed.

One of the great economic benefits of the personal computer revolution is that as the power of the fundamental building blocks of PC's has increased, the hardware components can be utilized for more and more highly-specialized tasks that formerly required purpose-built hardware. As just one example, computational tasks that formerly necessitated massively-expensive super computers can be accomplished by linked low-end PC's.

The key, the flash of genius, that allows groups of PC's to perform work that would otherwise require multi-million dollar hardware, is innovative software. Bilski appears to penalize that class of genius.

For me, branding an innovation unpatentable simply because is takes advantage of a general purpose digital computer instead of a special-purpose machine runs counter to the clear benefits of the explosion of economic wealth and value that the efficiencies of Moore's Law and Metcalfe's law have brought into existence. As an inventor, I am financially incented to develop an innovative technology that can only be operated on expensive specialized hardware instead of a general purpose computer (and it may require even greater genius to develop for a general purpose computer). Thus, the resulting product is so expensive that its benefits are available to only a small fraction of those who want/need those benefits. Where's the societal payoff in that?

Sorry for the rant, but I think the CAFC has gotten some public policy issues very wrong in the Bilski decision.

Posted by: dvan at November 4, 2008 10:36 AM

For me, branding an innovation unpatentable simply because is takes advantage of a general purpose digital computer instead of a special-purpose machine runs counter to the clear benefits of the explosion of economic wealth and value that the efficiencies of Moore's Law and Metcalfe's law have brought into existence.

Actually, the rule isn't that its unpatentable if you recite a general purpose computer. The rule is that you need to recite either (1) a particular machine or (2) a transformative function. You're free to claim a transformative function in a general purpose computer, regardless of whether the general purpose computer is a particular machine.

What is left in the air is whether, if you don't have a transformative function, will a general purpose computer with the non-transformative function satisfy 101.

Posted by: anonymous at November 4, 2008 4:57 PM

Why not sidestep Bilski completely and just redraft your general purpose method as a Beauregard claim. It seems to me that then Bilski doesn't apply. And after all, you have to store that computer program somewhere to execute it.

Posted by: anonymous at November 4, 2008 11:49 PM

"not patentable: method for a general purpose computer
patentable: same method for two general purpose computers"

Oh Gosh

If the above sentence is true then I pity the US Patent system
What kind of ignorant idiot would propose such a criterion for patentability, in the age of multi-core CPUs and distributed computing ?

Sometimes I think that any punk from the street would do a better job writing opinions for higher courts of this country

Posted by: angry dude at November 5, 2008 8:15 AM

"As an inventor, I am financially incented to develop an innovative technology that can only be operated on expensive specialized hardware instead of a general purpose computer"

Wrong
As an inventor you are NOT financially incented anymore to produce a new technology which has anything to do with computing (unless you can keep it a trade secret, e.g. server-side web search technology, or server-side voice recognition software etc.)
Whatever specialized novel computing hardware you can come up with will be ultimately ripped off and implemented on a low cost commodity PC, essentially moving your invention into unpatentable domain and denying you even a small chance of making profit in the market

Posted by: angry dude at November 5, 2008 8:30 AM

Why not sidestep Bilski completely and just redraft your general purpose method as a Beauregard claim. It seems to me that then Bilski doesn't apply. And after all, you have to store that computer program somewhere to execute it.

That's fine right now. However, if "process" doesn't really mean "process" than "article of manufacture" may not really mean "article of manufacture" and you're no better off. The difference, of course, is the Supreme Court caselaw hasn't attempted to redefine "article of manufacture" so you might as well try the beauregard claims for now.

Posted by: anon at November 5, 2008 9:46 AM

"In my view now, nothing but uncertainty abounds."

Now you're catching on.

"this would endanger many HARDWARE claims."

Unfortunately, no it would not, because they meet the overarching "prong" of being a physical object themselves and thus meet 101. Remember, these two prongs are only applicable to PROCESS claims.

"Actually, the rule isn't that its unpatentable if you recite a general purpose computer. The rule is that you need to recite either (1) a particular machine or (2) a transformative function. You're free to claim a transformative function in a general purpose computer, regardless of whether the general purpose computer is a particular machine.

What is left in the air is whether, if you don't have a transformative function, will a general purpose computer with the non-transformative function satisfy 101.
"

This man knows what he is talking about.

"Whatever specialized novel computing hardware you can come up with will be ultimately ripped off and implemented on a low cost commodity PC, essentially moving your invention into unpatentable domain and denying you even a small chance of making profit in the market"

Hardly. There is an entire semiconductor/hardware industry that isn't even affected by this ruling.

"Why not sidestep Bilski completely and just redraft your general purpose method as a Beauregard claim. It seems to me that then Bilski doesn't apply. And after all, you have to store that computer program somewhere to execute it."

Because they're next on the chopping block. Wait and see.

Posted by: 6000 at November 5, 2008 11:57 AM

"Whatever specialized novel computing hardware you can come up with will be ultimately ripped off and implemented on a low cost commodity PC, essentially moving your invention into unpatentable domain and denying you even a small chance of making profit in the market"

Hardly. There is an entire semiconductor/hardware industry that isn't even affected by this ruling."

???????????????????????

Watch for things to come
Ask those guys from fabless semiconductor companies like Amberwave about their feelings
I hope you don't consider them patent trolls ?

It is a well known fact that as CPU power becomes cheaper many functions tend to be at least partially implemented in software running on a general purpose CPU as opposed to add-on specialized DSP device, for example, DSP-based computer modem.

Ever heard of soft and hard modems ?

I guess you haven't...
BTW, is that you, little fella, a patent-hating patent examiner ?
No further questions...

Posted by: angry dude at November 5, 2008 1:16 PM

You guys are misreading the case completely. There is no suggestion that a computer program for a general purpose computer cannot be patented. Read the Benson case. It was a unique situation where the invention related to converting decimal-encoding numbers to binary-encoding numbers. Benson was a poorly reasoned opinion, but the court here distinguished its ruling from Benson, saying that the method in Benson was mathematical principle which cannot be patented, and therefore, the method claim at issue, which would foreclose all applications of the principle, would also not be patentable.

Posted by: lenehey at November 5, 2008 1:35 PM

I agree that Michel absolutely blew this one and made everybody's life not just harder but impossible; i.e., what the frick does "particular article" mean. Does it mean a separate patent or claim for every single computer in the world? Does it mean a "particular TYPE of article?"
This is the worse, most damaging opinion by any court I have ever read in patent law. It is the Dred Scot of patent law.

Newman wasn't all that great either, although I agree more with her general position.

My two favorite CAFC judges pretty well botched it.

But the question I have is this: I believe the position Hawk is selling is that a machine (hardware) is inherent in a software claim. The s/ware claim makes no sense and has no utility unless tied to the machine. That's inherency.

Inherency will be used by the examiners and courts at the drop of a hat to trash claims, and I don't have a particular problem with that when done properly. But does it swing the other way? Is there any case law where a claim is saved by reliance on inherency?

In this situation, for instance, can an applicant claim that the necessary tie to the machine is inherent in the s/ware? If not, why not?

Posted by: Babel Boy at November 5, 2008 2:54 PM

"But the question I have is this: I believe the position Hawk is selling is that a machine (hardware) is inherent in a software claim. The s/ware claim makes no sense and has no utility unless tied to the machine. That's inherency.

Inherency will be used by the examiners and courts at the drop of a hat to trash claims, and I don't have a particular problem with that when done properly. But does it swing the other way? Is there any case law where a claim is saved by reliance on inherency? "

Yes, many arguments on here have been relying on that of late, and that seems unfortunate, but nothing in a claim is "inherent". Things in references are "inherent".

Besides, can someone give us an example of where a computer is inherent but not explicit in a claim?

Posted by: 6000 at November 5, 2008 5:18 PM