Qu'est-ce qu'une "Invention mise en oeuvre par ordinateur" ?

Est-ce qu'une machine a laver peut être une "invention mise en oevre par ordinateur"? Seulement les Programmes sont "mis en oevre par ordinateur". Selon l'Art 52 CBE les programmes pour ordinateurs ne sont pas des inventions au sense de la loi des brevets.

The term "computer-implemented invention" is not used by computer professionals. It is in fact not in wide use at all. It was introduced by the European Patent Office (EPO) in May 2000 in Appendix 6 of the Trilateral Conference, where it served to legitimate business method patents, so as to bring EPO practise in line with the USA and Japan. Much of the European Commission's directive proposal is based on wordings from this "Appendix 6". The term "computer-implemented invention" is a programmatic statement. It implies that calculation rules framed in terms of the general-purpose computer are patentable inventions. This implication is in contradiction with Art 52 EPC, according to which algorithms, business methods and programs for computers are not inventions in the sense of patent law.

Programmers speak about "implementing a specification". A patent claim usually specifies a sequence of events (i.e. a program) which could be implemented by a programmer or by an engineer. "Implementation" in this context means "working it out in detail so that it can run without errors". The implementation requires human work. It is never done by a computer, and in fact computers will usually not play any role in it at all, beyond that of an auxiliary tool for checking the validity of logical concepts, similar to pencil and paper.

Moreover, the programmer would not say that she is implementing an "invention", no matter how new and admirable the specification is. Rather she might be implementing an algorithm, an idea or, more often, a complex combination of such abstract elements. Real advances in the art of programming are generally of abstract nature and therefore would usually not be referred to as "inventions" by the person skilled in the art.

A term like "computer-executed innovations" or simply "data processing innovations" would have been less conducive to misunderstanding. However misunderstanding was exactly what the proponents wanted.

The proponents of patents on "computer-implemented inventions" sometimes say that it refers only to "washing machines, mobile phones, intelligent household appliances ..." and "not computer programs as such". However this is untrue. The term as defined in the EPO's Trilateral Appendix 6 and in the Commission's Directive Proposal refers to nothing but data manipulation processes running on general-purpose computing equipment. Even if the extremely rare cases where a conventional washing machine is involved in a patent claim which EPO/CEC would subsume under "computer-implemented inventions", the gist of the "invention" will lie in data processing, not in applying heat and reactants to clothes. Otherwise this would, according to the EPO/CEC definitions, no longer be a "computer-implemented invention" but rather an ordinary technical invention (which does not match the CII definition, because its "prima facie novel" aspect is unrelated to computers).

The European Parliament has redefined the term "computer-implemented invention" in such a way that general-purpose data processing does not qualify while a washing machine, where data processing has only an auxiliary function and is not constitutive for the invention, would qualify. This amendment is fiercely opposed by the European Commission and by all those who earlier claimed that they want only washing machines and the like but not programs as such to be patentable.

While the redefinition of the European Parliament constitutes a clever way of correcting a central flaw of the proposed directive, it does not make the text much clearer. The misleading term "computer-implemented invention" continues to be used, and it will continue to mislead all those people who have not carefully read and memorised the EP definition.