10 November 2014

Australia’s ‘Alice’: Appeals Court Denies Business Method Patent

No computerIn a unanimous decision, a Full Bench of the Federal Court of Australia has ruled that the ‘substance’ of an invention trumps the form in which it is claimed, dismissing Research Affiliates’ appeal against a decision of a single judge of the court, which found its invention, relating to the construction and use of passive portfolios and indexes for securities trading, ineligible for patent protection: Research Affiliates LLC v Commissioner of Patents [2014] FCAFC 150.

The decision canvasses many of the same issues tackled by the US Supreme Court in its recent Alice Corporation v CLS Bank ruling.

In essence, the Australian court addressed the question of whether an invention that would otherwise be ineligible for patent protection – in this case, a scheme for generating a financial index which can be used, for example, to measure the rise and fall in value of an investment portfolio – becomes patentable when implemented in software.  This has been an open question in Australia since the court’s 2006 decision in Grant v Commissioner of Patents [2006] FCAFC 120, in which it found that an asset protection method was ineligible for patenting.  In that case, however, the court was not required to consider whether the outcome may have been different had all or part of the method been performed using a computer.

In the Grant case the court stated that, for a claimed invention to be patent-eligible, a ‘physical effect in the sense of a concrete effect or phenomenon or manifestation or transformation is required’.  Where a computer is involved there is always, of necessity, some physical effect or change of state, at least within the electronics of the machine.  The question that the court has not previously addressed is whether or not this is sufficient to confer patent-eligibility? 

In Research Affiliates, the court has answered this question in the negative, finding that describing a scheme – no matter how ingenious it may be – that is not itself eligible for patenting, and merely giving directions to implement it on a standard computer is not sufficiently transformative.

However, while – subject to any appeal to the High Court – this decision would seem to spell an end to ‘do it on a computer’ patent claims in Australia, the Full Court has been cautious (certainly more so that the US Supreme Court in its recent Alice v CLS Bank decision) to ensure that most computer-implemented inventions remain patent-eligible in Australia.

Invention Not An ‘Artificially Created State of Affairs’

Patent-eligibility in Australia is assessed according to the venerable ‘manner of manufacture’ test, and the leading decision on this aspect of the law remains the Australian High Court ruling in National Research Development Corporation v Commissioner of Patents (‘NRDC’) [1959] HCA 67.  The NRDC decision was most recently followed by the Full Court in the Myriad Genetics case, in relation to which I described it as ‘without doubt one of the greatest and most prescient judgments to have been delivered anywhere in the English-speaking world.’  The fact that the same principles can be applied effectively to technologies as disparate as genetics and computer-implemented financial instruments reinforces this description.

The guiding principle established by the High Court in NRDC was essentially that the question of whether or not new subject-matter is eligible for patenting should not be constrained by the example of what was previously considered patentable, but rather by whether or not the invention ‘consists in an artificially created state of affairs’ (i.e. something which, but for human intervention, would not exist), having economic significance (i.e. that it ‘belongs to a useful art as distinct from a fine art’).

There was no dispute about the economic significance of Research Affiliates’ invention in this case.  As for the artificially created state of affairs, Research Affiliates argued that both the index produced by its method, and the software-implemented method itself, result in physical effects in a computer, which is an artificially created state of affairs in the sense required by NRDC (decision at [80]).

It was not disputed that there are physical changes within a computer when a program is executed, however, as the court asked (at [82]):

Is that sufficient? As the Commissioner puts it: is an “invention” which is in the nature of a business, commercial or financial scheme rendered patentable “merely” by reason of the fact that it is implemented by means of a computer?

In the case of the Research Affiliates invention, the court concluded that this was not sufficient, reasoning (at [106]) that:

The determination whether the claimed invention is truly “an artificially created state of affairs” in satisfaction of NRDC is made not by some mechanistic application of the criterion of artificiality or physical effect, but by an understanding of the claimed invention itself. The invention is to be understood as a matter of substance and not merely as a matter of form. (Emphasis added.)

The court went on to evaluate, on the basis of what is disclosed and claimed in the specification, the substantive contribution made by the inventors, concluding (at [115]) that:

Here, that subject matter [which needed to exhibit the required characteristics of a manner of manufacture] is truly the scheme, the idea, the index. As set out in the specification it may be, and in the claimed method it is, implemented in a computer, but the ingenuity of the inventors, the end result of which is the invention, is directed to the idea, which is not patentable.

Problems with the Decision?

I think that the court has struggled with this case, which may explain why it has taken nearly a year to issue a decision that is not, by Australian Federal Court standards, especially lengthy or complex.  Even given the time taken to produce the judgment, it appears to contain some unresolved internal tension

When I attended the hearing last year, it seemed to me that the judges were already troubled by the idea that an otherwise ineligible method might be made patentable by adding little more than ‘do it on a computer’.  The difficulty is that any programmed computer is a man-made construct and, in this sense, an ‘artificially created state of affairs’.  So on what basis might this somehow avoid passing the test set out in NRDC?  And how might a distinction be drawn without, as it were, throwing the baby out with the bathwater and rendering all computer-implemented inventions ineligible for patenting?

At [110], the Full Court reiterated its earlier warning (in CCOM Pty Ltd v Jiejing Pty Ltd [1994] FCA 1168) ‘against bringing into the determination of “manner of manufacture” considerations of whether what was claimed involved anything new and unconventional in computer use’.  So, on this basis you would assume that there is no requirement for ingenuity to reside in the implementation of an invention using a computer.  This is, more properly, the role of the inventive step inquiry.

Yet the court went on to say (at [118]) that:

The claimed method in this case clearly involves what may well be an inventive idea, but it is an abstract idea. The specification makes it apparent that any inventive step arises in the creation of the index as information and as a scheme. There is no suggestion in the specification or the claims that any part of the inventive step lies in the computer implementation. Rather, it is apparent that the scheme is merely implemented in a computer and a standard computer at that. It is no part of the claimed method that there is an improvement in what might broadly be called “computer technology”. (Emphasis added.)

I am not entirely sure how these statements are to be reconciled, however it appears that the solution may lie in the nature of the ‘idea’ underlying the implementation.

In this case, the inventive concept is an ‘abstract idea’ (in the court’s view), and thus ineligible for patent protection, and the implementation – according to the inventors’ own evaluation – involves no additional inventive ingenuity.  However, if the underlying inventive concept had been something that itself fell within the scope of patent-eligible subject matter (such as, for example, a method of operating an internal combustion engine in order to achieve improved fuel economy) then the fact that the computer-implementation may then have involved only routine and non-inventive programming steps would not render the invention ineligible.

Yet, even this explanation of the court’s approach is not entirely satisfactory.  If correct, then it seems that the court has dissected the Research Affiliates claims to isolate the unpatentable content (i.e. the ‘abstract’ index-generation method) from the potentially-patent-eligible limitations (i.e. specific technological implementation via computer software), before asking whether there is some ‘inventive’ contribution in the latter that would overcome the ineligibility of the former.

This analysis – if it is indeed what the court has done here – has much in common with the US Supreme Court’s approach to the Alice case, and to the European ‘technical contribution’ approach.  However, this type of dismantling of patent claims, rather than considering the claimed subject matter as a whole, has never been a valid approach in Australia.

Conclusion – A Narrow Ruling?

I believe that the Full Court intends its decision in this case to be interpreted narrowly.  It perhaps stands for nothing more than the proposition that disclosing an unpatentable concept in a patent application, and then seeking to claim this as an invention by ‘merely’ directing that it be implemented via computer using standard techniques, is insufficient to confer patent-eligibility.

I fear, however, that it will be interpreted more expansively than this – most imminently by the Patent Office – to capture a wider range of allegedly patent-ineligible ‘schemes’ and ‘abstract ideas’ than the court may have intended.  A continuing difficulty in Australia – as is proving to be the case in the US – is the lack of any satisfactory definition of what constitutes an ‘abstract’ idea.  Almost any invention is, at some fundamental level, based on a ‘abstract’ concept conceived by the inventor.  And many computer-implemented inventions employ standard hardware, well-known programming techniques, and existing building blocks to bring such ideas to practical fruition.

On the other hand, the judgment does contain some useful discussion of considerations distilled from the various Australian authorities that may be applied to identify patent-eligible inventions.  I intend to discuss these in a separate article.

As for the Full Court’s apparent ‘dissection’ of Research Affiliates’ claims, the validity of this approach could be a question for the High Court, should Research Affiliates elect to apply for special leave to appeal, and should such an application be granted.  By all accounts, the invention is quite valuable, and so it is certainly on the cards that Research Affiliates may decide to take its case all the way to the top.


Unknown said...

A good result and not a "narrow" result at all.

Unknown said...

This case appeared to try and differentiate it from RPL. It will be interesting to see how RPL pans out

Unknown said...

Well, there is some positive treatment of the RPL Central decision, which I am quite pleased about! However, this is directed to the principles applied by Justice Middleton, and his discussion of the earlier authorities. It is not directed to whether he applied those principles correctly to the claims at hand (and rightly so, since the Full Court is well aware that the RPL decision is also on appeal, and could well be before them in due course).

It is possible (though I suspect unlikely) that the Commissioner of Patents could see the writing on the wall, and elect to drop the appeal in RPL Central. At the same time, I think that an application by Research Affiliates for special leave to appeal to the High Court is practically certain: on a simple cost/benefit analysis, it is almost a no-brainer. The bigger issue is whether they can put up a question that the High Court would be inclined to take on.

Unknown said...

I would be interested to know in what way you consider the result to be 'good', and why you think it is not 'narrow'. When it comes to judicial decisions at this level, there is a difference between 'good' and 'correct outcome'. The very fact that we can disagree as to whether or not this is a 'narrow' ruling indicates that it is perhaps not that 'good'.

Unknown said...

The "substance of the invention" appears to me to be a new test. The decision at first instance seemed to allow for a claim with no details of implementation to be cured through the addition of these details (although the claim would still have to be supported by the specification). However, this decision implies that the claims are largely irrelevant - it is more important what the specification is describing in substance (whatever that means). This seems like a difficult position to reconcile with the fact that the invention is defined by the claims.

Unknown said...

I agree. One could argue, I suppose, that the 'substance' should emerge on a 'proper construction' of the claims, in light of the specification as a whole, but if so then that does not really come through clearly in the decision.

Unknown said...

The problem with trying to use inventive step for this purpose is that if the invention resides wholly within ineligible subject matter, but is nonetheless non-obvious, the computer-implementation would get it over the 'manner of manufacture' hurdle, while the non-obviousness of the underlying idea would provide the inventive step. There is no scope under Australian law for disregarding 'ineligible' components of a claim when considering inventive step: the claim must be considered as a whole.

The EPO, on the other hand, does do this. The requirement found in the EPC for an invention to make a 'technical contribution' to the art is understood to authorise the disregard of any non-technical elements. The UK courts have reached essentially same result by a slightly different path.

The new exclusion in the New Zealand law is expected to work in a similar manner.

The US Supreme Court has a line of authority for treating ineligible subject matter as being, in effect, part of the prior art, and conducting a type of informal inventive step analysis, justified by the words 'whoever invents... in 35 USC 101.

I am not sure that any of this is justifiable under Australian law, unless authority can be gleaned from the Microcell/Philips v Mirabella line of decisions, i.e. that there is no 'invention' to be found on the face of the specification. I note, however, that the Full Court has not even cited Microcell or Philips in this case.

I think Tom Lehrer best summed up the problem with Justice Stewart's formulation when he wrote/sang that:

All books can be indecent books,
Though recent books are bolder.
But filth (I'm glad to say) is in the eye of the beholder!
When correctly viewed,
Everything is lewd:
I can tell you things about Peter Pan,
And the Wizard of Oz - there's a dirty old man!

Unknown said...

I would have to agree. Take these paragraphs from RPL,

"I further find that each of the computer-effected steps of the invention constitutes or gives rise to a change in state or information in a part of a machine, and therefore produces a physical effect in the sense of a concrete effect or phenomenon or manifestation or transformation (Grant at 70 [32])." at [143].

"I do not accept that the Australian decisions that are binding upon me in respect of this issue – NRDC, CCOM and Grant – expressly import any such requirement of substantiality or centrality of physical effect" at [147].

These passages in particular seem to stand out as inconsistent with the "substance of the invention" test outlined in the Research Affiliates decision. Regardless of whether the court finds this to actually be inconsistent or not, I think the outcome of the appeal in RPL will have far greater ramifications (at least for software patents) than this decision does.

Unknown said...

I think the court reached the right result, but I'm not at all convinced by the path it took to get there.

The passages where the court announces its "substance, not form" test just don't fit with the rest of the judgement, which seems to point inescapably at the conclusion that RA's claims would have been A-okay if instead of being in the form " + do it on a computer" they'd been expressed as " to produce physical changes in that correspond to ". Or even " configured to perform ", maybe with a few non-limiting processing units and storage media thrown in. Each of those forms of claim would protect, in substance, the same thing.

The problem is that the courts have painted themselves into a corner; they've stretched the NRDC test to the point where they have no principled way of rejecting things like the RA claims that cannot in any sensible way be described as a "manner of manufacture" and that are so bad that even the US system has found a way to control them. So they produce judgements like this one. What a mess.

Unknown said...

The US system has not (yet) come up with an entirely satisfactory solution. The difference is that the US cases have gone all the way to the Supreme Court, whose word is law. Whether or not the reasoning in Alice or Bilski is logical or coherent, it is law. It certainly "controls" the most extreme cases, but it is also (on at least some analyses) resulting in considerable collateral damage.

I am going to write more on this. But the more I think about it, the more I hope this case goes to the High Court. That is the only venue in which NRDC is not immutable binding precedent.

We need a new NRDC for a new era. This Full Court decision is not it!

Unknown said...

The Full Court trumps a single judge... even if Justice Middleton's decision in RPL Central is more logical and coherent (though, of course, I am biased :).

Post a Comment

Copyright © 2014
Creative Commons License
The Patentology Blog by Dr Mark A Summerfield is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Australia License.