The EPO is publishing the proceedings of the 13th Symposium of European Patent Judges, where Judge Melullis has been requested by the organisers to make an intervention with European relevance. Here is his intervention (see p186 of the proceedings):
Some problems of patent law from a German viewpoint
The symposium organisers have asked me to take a particular case or a number of cases as a starting point for discussion. This puts me in something of a quandary, as the recent case law of the Federal Court of Justice does not include any landmark decisions with a European relevance that could serve as concrete examples for such a discussion even though the work of the German courts in this area is, of course, mainly concerned with patents granted under the European Patent Convention. The specific questions are seamlessly integrated into the case law of the Federal Court of Justice (FCJ), which follows the same principles when dealing with national patent law as in applying the EPC. I shall therefore confine myself to outlining a few points which would seem to be of general interest.
1. The patentability of software or, in the broader wording of the EPC, "programs for computers", has been the subject of prolonged and intensive discussion among patent law professionals and interested sections of the public. This debate has also been fuelled by the efforts of some, mainly larger, companies in the field to obtain protection for software although the Convention expressly excludes the patenting of computer programs as such by relativising the rules, even if they cannot be abolished. This trend has been reflected, unfortunately, in the legal discussion. The relevant provision of the Convention is not a model of clarity, and legal sophistry has been employed with some success to obscure its meaning still further, in a way that in effect colludes with the efforts to extend patenting to software.
In several decisions over the past few years, the FCJ has sought to address this problem by looking again at the wording and purpose of the legal provision. These decisions have brought the developments in Germany to a conclusion, at least for the time being, which, unlike the efforts of, for example, the European Commission, limits the relevant industrial property rights to the area which to us appears legitimate
The FCJ proceeds from the assumption that the prohibition on the patenting of software "as such" means what the law says. Software cannot be protected merely because it achieves, inter alia, technical effects when used in conjunction with a technical device. Conversely, the prohibition means that, despite the technical effects normally associated with the use of software, these effects alone do not imply that the software contains a teaching for which patent law offers protection or, more simply, software is not patentable merely by virtue of being used in conjunction with a general-purpose computer (FCJ GRUR 2002, 143 Suche fehlerhafter Zeichenketten). This broadly corresponds to the earlier approach of the European Patent Office (See, e.g., the EPO's decisions T 931/95 dated 8 September 2000, OJ EPO 2001, 441 Controlling pension benefits system; and T 1173/97 dated 1 July 1998, OJ EPO 1999, 609 Computer program produc), which held, albeit on the basis of different thinking, that patentability required a technical effect beyond the normal use of a standard computer in conjunction with a program. The main difference in approach consists in the fact that the EPO boards of appeal resorted to the term "technical" for the purpose of assessing software: the FCJ found this unhelpful, firstly because, precisely when assessing software as such, the program's interdependence with the technical device makes technical content hard to deny (See, for example, FCJ Sprachanalyseeinrichtung.), so that the Court had to fall back on the expedient of the "not so technical teaching". In particular, resorting to the term "technical" was unhelpful because it shifted the focus of discussion from "programs as such", a concept which was already unclear, to a further, equally ambiguous term, the meaning of which was inherently evolving and, moreover, would become further obscured as a result of the necessary distinction from subject-matter of a "not so technical" nature.
According to this FCJ case law, a computer program can only qualify for patent protection if it solves a specific technical problem that goes beyond the normal interaction of software with a general-purpose computer. "Going beyond" means that this effect cannot be confined to controlling a general-purpose computer in order to solve a non-technical problem; apart from this, it must solve a specific technical problem. Even where this is deemed to be the case, only those elements which serve to solve the technical problem by technical means will be included in the assessment of novelty and inventive step; particular developments outside the technical field may only be considered to the extent that they influence the technical development. To take a simple example: a novel program for commercial accounting with a computer is not inherently patentable merely because the business idea it embodies is novel and not readily found elsewhere: the yardstick of patentability can only be the underlying technical implementation of that idea, in so far as the implementation solves a technical problem arising from the idea, and does so by technical means. Patent protection is only available for a program of this kind if the implementation, going beyond the normal interaction of hard- and software, exceeds the abilities of an average developer.
With this approach, the focus shifts in accordance with the thrust of the legal provision from the less tractable question of the computer program as such to the issue of inventive step. For the purpose of determining inventive step, via an evaluative assessment of the subject-matter, only those elements count which are part of the solution to the problem. Software may be included only in so far as it contributes to the solution in question; taking software into account otherwise, or protecting it as such outside the context of the solution of a specific technical problem, is largely ruled out.
On the basis of these deliberations, patent protection in the field of data-processing is initially available for developments arising from a specific interaction of hard- and software. The FCJ has not advanced beyond its early decision on the patentability of, for example, an anti-blocking device formed from a braking device and the control thereof by means of software; and it sees no reason considering the requirement for a technical teaching directed to the solution of a specific technical problem, and the intention of the legal provision to deviate from its established position. With a system in which software serves to control a technical device with a technical effect, the software is not protected as such, but only as part of that system, although a sufficient basis for novelty and inventive step can be found in a novel concept of control realised especially in the software, if all the parts interact.
The same approach also effectively rules out patent protection for software-related inventions where the description of the technical implementation is unclear or vague, i.e. where the subject-matter is described in terms of a task to be accomplished with the aid of a computer. If no details of the technical implementation of the task are given, this is either left to the skilled person, with the result that, even where the non-technical problem as such constitutes a relatively exceptional achievement, there is no inventive step in this case, the teaching as formulated already assumes that the implementation cannot go beyond the abilities of an averagel competent skilled person. Or, if the implementation is not feasible, there is a lack of sufficient disclosure, since the skilled person is not enabled, on the basis of a mere description of the task, to perform an act according to technical instructions which go beyond non-technical statements and the disclosure of which alone can establish patentability.