23 November 2014

Research Affiliates: A ‘Hard’ Case Making Bad Law

Chaos and OrderIt is sometimes said that ‘hard cases make bad law’.  This is not actually generally true, and good and bad law emerges from a wide variety of different cases.  Perhaps, in fact, the quality of judges, or the quality of argument presented at trial, are more important factors than the difficulty of the case.

That being said, ‘hard’ cases do present judges with special challenges.  In this context, a hard case is one in which the most literal or straightforward application of the law – i.e. the binding precedents and relevant statutory provisions – is liable to produce a result that may strike the judge as illogical, unfair, absurd, unjust, or just plain wrong.

The temptation in such cases must be very great to try to find a ‘creative’ approach to applying the law in order to make the legal outcome match the subjectively ‘right’ decision.

Patent law is not immune from such ‘hard’ decisions.  Sometimes, for example, an alleged infringer’s efforts to ‘work around’ a patent, and escape infringement on a technicality, are so flagrant that judges are inclined to find ways to broaden the interpretation of patent claims in order to capture the infringer’s variation.

So what would happen if a judge (or, indeed, three judges) were to look at an invention and think ‘surely that cannot be eligible for a patent?’  And what if they then looked at the current state of law and found that, in fact, a literal and straightforward application of the law could well lead to the conclusion that the suspect invention is patent-eligible?  The answer might be that you end up with a decision like that of the Full Court of the Federal Court of Australia in Research Affiliates LLC v Commissioner of Patents [2014] FCAFC 150.

Research Affiliates – A Shambolic Decision?

In initially reporting the Research Affiliates decision, I suggested that the court intended it to be interpreted narrowly, but that I feared it would be applied more expansively.  I have already seen evidence that this concern is well-founded, in the form of a further examination report on an application that had not previously received any objection on grounds of subject matter, now citing the Full Court’s decision in raising such an objection.

With this in mind, I have decided that a more critical look at the Research Affiliates decision is in order.  This should have been a good judgment.  The court spent nearly a year coming to a decision.  In the meantime five judges of the Full Court issued a unanimous decision on patent-eligibility in D'Arcy v Myriad Genetics Inc [2014] FCAFC 115 which has been praised for its scientific and legal rigour not only on this blog, but also by such highly-regarded jurists as former Chief Judge of the US Court of Appeals for the Federal Circuit Randall Rader.

Yet where the Myriad decision is consistent and coherent, the Research Affiliates judgment is something of a shambles.  It lurches through a review of UK, European and US law for no apparent purpose, since this seems to have no bearing on the analysis under Australian law.  Worse yet, however, the Full Court in Research Affiliates appears to have developed three new theories about patent-eligibility that are not only without clear precedent in the leading High Court case on the subject, National Research Development Corporation v Commissioner of Patents (‘NRDC’) [1959] HCA 67, but which cannot even be reconciled with the court’s own recent Myriad decision.  This is a problem, not least because Myriad was a decision of five judges of the court, whereas Research Affiliates was decided by the more usual panel of three judges.  That makes Myriad binding on the Research Affiliates court, though you would not know it from reading the decision.

The three arguably new theories, which I discuss in greater detail below, are:
  1. that the NRDC test, requiring an ‘artificially created state of affairs’ which has economic utility, is not a ‘sufficient or exhaustive’ test of patent-eligibility;
  2. that for a practical application of a ‘abstract idea’ to be patent-eligible, some element of ingenuity must lie in the implementation itself; and
  3. that the test of patent-eligibility requires looking to the ‘substance’ of the invention, as determined from reading the patent specification as a whole.

The NRDC Test as Necessary, But Not Sufficient?

The Full Court in Research Affiliates contended, at [101], that:

The characterisation of patentability by reference only to the description in NRDC of a product which consists of an artificially created state of affairs of economic significance was part of the High Court’s reasoning but did not represent a sufficient or exhaustive statement of the circumstances in which a claimed invention is patentable.  (Emphasis added.)

It is not at all clear how the court arrived at this reading of NRDC.  As stated by the Full Court in Myriad at [168], ‘[t]he principle to be taken from the High Court’s explanation in NRDC is in terms of the intervention of man to produce from, or by means of, a naturally occurring product or the laws of nature, something artificial or of an artificial effect that can bring benefit to mankind. That benefit generally has economic utility.’  The reference here to ‘a naturally occurring product or the laws of nature’ can equally be extended to encompass so-called ‘abstract ideas’, and other excluded categories of subject matter.  The point is that while a ‘mere’ idea, discovery or natural law may not be patentable, a practical application of the idea having an economic benefit will be patentable.

The Myriad court further stated, at [208], that the High Court in NRDC ‘held that it is sufficient for a product to result in “an artificially created state of affairs”, leading to “an economically useful result”.’  This bears emphasising: a panel of five judges of the court (which is binding on the three-judge panel in Research Affiliates) held that it is sufficient, for an invention to be a manner of manufacture, that it consists of an artificially created state of affairs of economic significance. 

Requiring Ingenuity in Application?

The Research Affiliates court refers to the Myriad decision only in passing, and in listing ‘principles’ established by NRDC (at [8]) is silent as to the following relevant consideration, identified in Myriad at [115]:

One can distinguish between discovery of a piece of abstract information without suggestion of a practical application to a useful end, and a useful result produced by doing something which has not been done by that procedure before. It is no answer to ingenuity in the discovery that the materials would produce a useful result to say that there was no ingenuity in showing how the discovery, once made, might be applied. (Emphasis added.)

It is hard to reconcile this with certain of the reasoning in Research Affiliates, such as the statement (at [108]) that ‘[t]he computer that may be utilised is described in general terms, without an indication that any unusual technical effect is utilised’, or in the court’s conclusion, 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.)

So, the court concedes that the ‘idea’ behind the claimed method may be inventive but, in stark contrast to what was said just weeks earlier in the Myriad decision, concludes that the claim is nonetheless ineligible due to the lack of ingenuity in applying the idea using a computer.

A New Test of ‘Substance’?

The court in Research Affiliates also appears to have developed a new test of ‘substance’ of the invention.  As stated by the court at [106]:

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.)

So far this is not necessarily inconsistent with authority.  As the Myriad court stated (at [170]), ‘[i]t is necessary to look at the substance of what is claimed to see to what extent it departs from unpatentable subject matter, such as an abstract idea.’  The key to consistency is to recognise that ‘the invention’, in patent law, must always be referred to the claims (e.g. subsection 18(1) of the Patents Act 1990: ‘an invention is a patentable invention for the purposes of a standard patent if the invention, so far as claimed in any claim…’).

The Research Affiliates court went on to argue, at [113], that:

The effect of Research Affiliates’ submissions is that the mere implementation of any abstract idea or scheme in a well-known machine is sufficient to render that unpatentable subject matter patentable because it gives rise to an “artificial effect”. This approach is inconsistent with NRDC and is one of form not substance.

As will be apparent from the discussion of ‘ingenuity’ in implementation above, I do not agree that Research Affiliates’ approach is necessarily inconsistent with NRDC.  If the idea is inventive, the fact that its practical application may be straightforward (and thus readily dismissed as ‘mere implementation’) is no barrier to eligibility.  The claims here require the use of a computer – a meaningful limitation that takes all non-computer-implemented applications outside of its scope – and NRDC (and Myriad) do not authorise summary rejection on the basis that the computer implementation may be relatively simple.

Furthermore, because the Research Affiliates court cannot, therefore, find a basis for ineligibility in ‘the substance of what is claimed’ (as the Myriad court put it), it proceeds to look elsewhere, i.e. the body of the specification (at [114]):

The invention set out in the specification is directed to the index itself. The method of the invention is not one that has any artificial or patentable effect other than the implementation of a scheme, which happens to use a computer to effect that implementation. There is no technical contribution to the invention or artificial effect of the invention by reason of the intervention of the inventors. To take the words of NRDC at 268, the process does not produce “either immediately or ultimately, a useful physical result in relation to a material or tangible entity.” The claimed method, the result of the ingenuity of the inventors, does not produce such a result; the ingenuity is in the scheme. Again, drawing from NRDC at 270, there is a useful result of the claimed process but there is no physical thing “brought into existence or so affected as the better to serve man’s purposes”. There is no “physical phenomenon in which the effect, be it creation or merely alteration, may be observed” (NRDC at 276).  (Emphasis in original.)

Here the court is looking to the specification in order to justify disregarding express limitations recited in the claims.  It is also relying on statements in NRDC that are specific to the facts of that case, i.e. where the invention was a new and advantageous method of treating crops with a known chemical substance.  In doing so, it is ignoring the general principles that it is bound to apply, and in particular the binding precedent of the five-judge panel in Myriad, already quoted above, that ‘[i]t is no answer to ingenuity in the discovery … to say that there was no ingenuity in showing how the discovery, once made, might be applied.’

Conclusion – A Case of Bad Law

What is particularly disturbing about these apparent contradictions between Myriad and Research Affiliates is that two of the judges in the latter case (Justices Bennett and Kenny) were members of the Myriad panel, while the third (Justice Nicholas) was the first instance judge in Myriad whose decision was affirmed by the five judge panel.

Ultimately, the only reasonable inference that can be drawn from the inconsistences is that the court considers isolated DNA to be worthy of patent protection, in principle, whereas computer-implemented financial methods are not.  Faced with the ‘hard’ choice between applying the law consistently with its own precedent, and that of the High Court – which may have had the result of finding the Research Affiliates invention patent-eligible – or getting ‘creative’ in order to produce a subjectively desirable outcome, the court appears to have taken the second course, and in the process made a mess of the law.

Should Research Affiliates seek leave to appeal (as I expect it will), there is plenty here for the High Court to look at.  It may be that the NRDC principles need to be extended to account for recent developments in technology and business, and that something more may be required than to bring about an artificially created state of affairs having economic value.  Maybe there should be a requirement that ingenuity reside at least partly in a technical implementation of an idea.  Or perhaps it is sometimes necessary to look beyond what is claimed in order to identify the true ‘substance’ of the invention.

The High Court has the power to explain, extend or overrule its own previous judgments.  The Federal Court does not – whether one, three or five judges are sitting on the bench.  As matters stand, what are applicants, patent-holders, accused infringers, or the Patent Office supposed to do when faced with differing outcomes depending upon whether NRDC/Myriad or Research Affiliates is applied?  Technically, the answer should be to apply the higher authorities.  In practice (and I am seeing this already in examination reports coming out of IP Australia) it is likely that Research Affiliates will be applied in cases of computer-implemented inventions, despite its evident flaws.

Sometimes, hard cases really do make bad law!  My hope is that the High Court will be offered – and will take up – the opportunity to fix it.

7 comments:

Mark Summerfield said...

What is missing in the judgment is a clear sentence saying "it is not patentable because ". The judgment only goes through precedent, overseas cases and facts and then states that the invention is not patentable. But there is little nexus between all that. How is it possible to know when this judgment is applicable to the facts at hand?

Mark Summerfield said...

I agree. The lack of a coherent line of argument running through the judgment towards a logical conclusion is a further issue. I think we have a right to expect better from the Full Court, particularly when the decision took nearly a year to produce. It is almost as though the Federal Court wants this case to be taken up by the High Court... which is not a completely implausible suggestion!

Mark Summerfield said...

I agree that the decision is all over the place. Although I believe that the Court was trying to indicate that it thinks that the RPL decision is patentable subject matter (which is good for you Mark!!) I am having great difficulty in determining where to draw the line. In particular, both the RA and RPL methods seem to have the necessary transformation of data (which sets them apart from invention pathways which was a mere checklist). However both methods could be performed manually.


In addition, the Court seems to indicate that a (European style) technical effect is required (which would also mean that RPL would also be unpatentable subject matter), but at the same time stating that you need to look at the "vibe" of the method described in the specification. This seems to indicate that if the RA specification was written differently that it may be patentable subject matter? It also seems to indicate that all "pure" methods implemented on a computer are not patentable, but also indicating that it does not want to rule out all computer implemented methods.


As is common with the patent office at the moment, the Court also seems to be confused between inventive step and patentable subject matter.


This decision has made me more confused than ever. Goodness only knows how patent attorneys are meant to advise their clients.

Mark Summerfield said...

How do we advise our clients? Well, as you say, I think we just have to go with the 'vibe'! :)

Mark Summerfield said...

I do believe the court arrived at the right decision, I think it had trouble distinguishing between cases where an abstract but ingenious idea (which is obviously not patentable) is not implemented in software as much as it is translated wholesale into code and can actually be performed with pen and paper without a computer at all, and cases where the implementation in software is some essential part of the procedure itself (e.g. network traffic management, or specialised processors).


I would actually disagree with the RPL decision in that any abstract idea which can be accomplished with pen and paper and not requiring a computer, should not be patentable merely by moving the same calculations onto a computer. In this sense, RA is correct - there needs to be some implementation in software more than merely "[abstract idea] done on a computer."

Mark Summerfield said...

Of course you are entitled to your view, Dylan. However, I am wondering if you have actually read the RPL Central decision, or if you are relying on some oversimplfied summary (such as the one in the Research Affiliates judgment).
The whole point of the invention in RPL Central is to retrieve data available from an existing server, reprocess it into a more palatable form for an end user, present it to the end user at a client computer (e.g. via a web interface), and upload and store electronic documents and other information provided by the user in response.
While it could be argued that this might all be done on a much more limited scale using a range of non-computer-implemented technologies (e.g. telephones, human operators, monkeys with typewriters, the postal service), the invention itself actually resides in automating and integrating these existing disparate, uncoordinated and inefficient elements into a single (i.e. 'Central') internet-based system.
One thing that can certainly not be said of the invention is that it could be accomplished with a pen and paper, or without a computer.
Of course, I am biased in this particular case. But if such things are to be ineligible for patent protection, we need to be provided with a far clearer, more consistent and logical rationale than we have seen to date.
What we need is certainty. However, the only time we get it is when something is found to be patent-eligible (e.g. NRDC, IBM, CCOM, Sanofi-Aventis, Myriad). Judges are good at telling us why something is patent-eligible, which at least sets a high-water mark each time it happens (at the higher appelate levels, anyway). They seem less adept at articulating their reasons when they consider something to be ineligible. This ailment does not appear to be confined to Australia.

Mark Summerfield said...

It's been said many times by many judgements that abstract ideas/discoveries cannot be patentable, and all inventions are derivative of ideas/discoveries. Patentability requires some physical manifestation. The difficulty with the US tests in State Street and Bilski, and even Alice is devising a just test that can be applied generally.
The tests pronounced in Research Affiliates are simultaneously described at such a high level as to be arguable one way or another in almost every situation (i.e. "substance over form") or so narrow as to be easily distinguishable on the facts (i.e. ingenuity in applying a method on a computer). This makes them close to useless.
This decision is (another) instance of the Court deciding on the outcome and then figuring out how to make the precedent fit. Partly, I think this is because many people (including members of the judiciary) feel that business methods in computer implemented inventions don't deserve the same protection as, say, the method of isolating human DNA (or the isolated DNA itself). To some, a new use of stock price indices just don't feel like an invention worthy of protection, and this influences how we choose to interpret and apply the law. It's therefore easier to apply NRDC strictly to Myriad than in the present case.
If that is the case, then we need to address it, but society's sentiments are best left to Parliament to construct new law around, not the High Court. I will be interested to see what happens.

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.