The patentability of computer-implemented inventions in Europe and how they should be examined have long been controversial issues. Legislation appears at first sight to exclude computer programs from patentability, but in practice many patents are granted in this area because the inventions covered are deemed not to reside solely in a program for a computer as such.
Recent developments in case law, particularly at the European Patent Office (EPO), suggest that a degree of stability and consistency is developing, which means that the extent to which a given invention might be vulnerable to exclusion can be predicted with increasing reliably.
It is clear that the patentability of a genuinely technical invention, e.g. a method of manufacture, will not be harmed by the use of a computer program to implement some parts of the method. On the other hand, a non-technical invention, e.g. a business method or rule for a game, will not be rendered patentable by specifying that some part of it is implemented by a computer. When such “mixed” inventions are examined in the EPO, all non-technical features are disregarded and the inventive step of the remaining subject matter is examined. Often, all that is left is a general purpose computer and the application will be rejected.
We discuss the current situation below and offer advice for drafting and prosecution of inventions that might be implemented with computers.
A “Trio” of EPO decisions issued by the EPO Technical Board of Appeal 3.5.01 between 2000 and 2004 - Pension Benefits Partnership, Comvik and Hitachi/Auction Method - developed the doctrine that mention of any technical feature in a claim is enough to satisfy Art 52(2) EPC (which defines the exclusion of programs for computers as such, and is reproduced in the Annex below) but only “technical” features can count towards an inventive step.
Following criticism of the EPO case law by the UK Court of Appeal, the EPO Board 3.5.01 defended its position in Duns Licensing/Estimating Sales Activity issued in November 2006, which was strongly endorsed in a later Opinion of the Enlarged Board of Appeal in G3/08. Different Boards have consistently followed the approach set out in the “Trio”.
Thus, a mere recitation of a technical means, e.g. that a method is carried out by a computer, is sufficient to avoid exclusion by Article 52(2) EPC. However, in considering inventive step, a technical problem must be solved and non-technical features in the claims are disregarded. As we explain further below, in the EPO’s “Problem + Solution” approach to inventive step, non-technical aims and features are deemed “constraints to be met” that are applied in a preliminary phase prior to any inventive activity.
The EPO’s examination of inventive step almost always follows the so-called “Problem + Solution” approach in which the invention is regarded as a solution to a problem. Duns Licensing and the related cases emphasise the requirements that the problem must be an “objective technical problem” and the solution must be a technical one. Furthermore, the invention must be inventive from the point of view of a person skilled in a technical field. Thus objections to non-technical inventions in the EPO are now more often based on lack of an inventive step than on exclusion by subject matter (e.g. as a program for a computer as such).
In its most basic form, the EPO’s problem and solution approach assumes that the making of an invention is a two stage process: identification of a problem and creation of a solution to that problem. An inventive step can be found in either stage; most commonly the inventive step lies in solving the problem, but an inventive step can be recognised if the problem was not obvious (even if it was easy to solve once identified). In the case of many computer-implemented inventions, the case law seems to imply that a three step process may be needed: 1) formulation of non-technical constraints to be met, 2) identification of a technical problem, 3) creation of a solution to that problem and implementation of the solution in technical means.
Into the “non-technical constraints to be met” go all the non-technical issues (e.g. any business method to be implemented), which are considered “givens” for the person skilled in a technical art. No inventive step can be found there, no matter how clever or non-obvious the non-technical constraints are. To find an inventive step in the second and third steps it is necessary to identify a problem much more specific than “implement the business method in technical means”. In Pension Benefits and Duns Licensing there were no such specific problems and the claims simply recited “means for” carrying out the steps of the respective business methods, hence there was no inventive technical solution. In Hitachi a technical problem was identified but the solution was not technical, hence no invention. In Comvik the solution was technical, but held to be obvious on the facts, ignoring issues relating to the non-technical purpose behind the invention.
The above approach would seem to rule out a class of perfectly good inventions: where a technical solution to a non-technical problem is found or a solution is implemented in a technically clever way. However, it should be possible to protect these types of invention by suitable redefinition of the problem and Boards of Appeal have in the past said that if a technical invention is clearly inventive, the existence of a technical problem is implicit and it does not need to be explicitly identified.
One significant absence in the current EPO case law is the lack of a definition of what is technical. There does not seem to have been a decision where the applicant has argued against a Board of Appeal’s assessment of certain features as technical or non-technical. The best that has been said is that a technical field is one that is not one of the excluded fields listed in Article 52(2), such as a business method or computer program, but because this is a non-exhaustive list, it cannot be taken that everything else is technical. The reasoning of Duns Licensing implies that one guide to whether a field is technical would be the skills and qualifications of the practitioners in that art.
In Philips/Designing Optical Systems the point was made that a method of “mere design”, even of a technical thing, is purely abstract and thus not patentable, unless the claims are limited to “physical, technical implementations”. Such a limitation can be achieved by specifying that the method is carried out by computer or including a concrete step in the real world such as making the thing designed. UK case law has taken a similar line.
Infineon/Circuit simulation held that simulation of an electric circuit of a particular form and more generally “specific technical applications of computer-implemented simulation methods” are inherently technical, even without some step interacting with the real world.
In G3/08, the Enlarged Board’s comments imply that a method of design that can take place in the designer’s mind would be excluded. Thus prudence suggests that in an application for an invention relating to design or simulation, at least dependent claims to steps of making the thing designed or acting on the results of the simulation be included.
The substantive aspects of patent law of the member states of the EPC are supposed to be harmonized to the relevant provision of the EPC and National Courts are supposed to take note of Decisions of the Boards of Appeal of the EPO.
Decisions of the German Federal Supreme Court (Bundesgerichtshof or BGH) are said by some commentators to be in harmony with EPO case law. However decision Xa ZB 20/08 seems to impose a stiffer test than EPO case law for avoiding exclusion from patentability. Rather than the mere presence of a technical feature, the BGH appears to require an invention to relate to the operational capability, and/or provide improved performance, of the computer system.
The UK courts have continued to follow UK Court of Appeal precedent rather than the EPO approach as set out in the “Trio” and Duns Licensing. The UK approach was reviewed in detail for example in HTC v Apple in 2013, which confirmed that assessment in the UK courts of whether or not subject matter is excluded from patentability depends on whether an invention has made a technical contribution to the known art. Novel and inventive features which relate solely to non-technical subject matter (e.g. a novel and inventive business method) does not count as a technical contribution. Thus, if the “technical” elements of a computer implemented invention are known, a UK tribunal will conclude that the claimed subject matter is excluded from patentability. At the EPO, in contrast, as explained above, such inventions will tend to fail on inventive step.
In Halliburton’s Applications it was suggested that it didn’t matter whether the UK Courts followed the EPO approach or not, as long as the end result – grant or refusal of the case - was the same. However, despite this theory that the same end result should be achieved, in practice Decisions of Hearing Officers of the UK IPO seem regularly to reject applications with seemingly good technical content. Therefore, our view continues to be that the EPO is a more friendly forum for computer-implemented inventions than the UK-IPO and possibly also than the German PO.
Many computer-implemented inventions are most clearly and concisely claimed as method or process claims though care sometimes has to be taken to ensure that the method is properly limited to execution by a computer or other technical device. For infringement purposes it is important to ensure that the method contains only steps carried out by a client or a server and not a mixture unless unavoidable for novelty or inventive step reasons.
Having written a satisfactory method claim, it is tempting to simply present an apparatus claim consisting of “means for” carrying out each of the method steps. However such a claim seems to invite the conclusion that the invention lies in the method only and there is nothing to the technical implementation of it. Therefore, an apparatus claim that emphasises any specific structures or architectures required to implement the invention is greatly preferred.
EPO practice has long recognised that given a patentable method, a claim to a computer program comprising “code means” for carrying out the method is allowable. Such claims can often be formulated in dependant form, referring to a higher numbered method claim, but only where the method claim contains exclusively steps carried out by the computer. Claims to computer readable media are also permissible but often do not add any useful protection compared to a claim to the computer program per se.
Given the territorial nature of patents, it is often suggested that infringement of some patents to computer-implemented inventions might be avoided by situating a server carrying out some or all of the method in a jurisdiction where there is no patent. This situation was considered by the UK Court of Appeal in Menashe v William Hill in 2002. In that case, the Court held that situating a server in Antigua did not avoid infringement because a UK-based user still put the invention into effect in the UK.
We summarise by giving some practical pointers on the subject of computer-implemented and business method inventions:
 T931/95 of 8 September 2000
 T641/00 of 26 September 2002
 T258/03 of 21 April 2004
 T154/04 of 15 November 2006
 T471/05 of 6 February 2007
 T1227/05 of 13 December 2006
 Available in German at www.bundesgerichtshof.de via link “Entscheidungen” or Google “Xa ZB 20/08”
  EWCA Civ 451 of 3 May 2013 available at http://www.bailli.org
  EWHC 2508 (Pat) of 5 October 2011 available at http://www.bailii.org
 T 1173/97 (Computer program product/IBM) of 1.7.1998
  EWCA Civ 1702 (28 November 2002)
28 August 2013