19 April 2010

Is software really comparable to books/films?

In a recent decision the parilament of New Zealand voted for a major patent reform bill that tightens the standards of patentability, especially with respect to software-implemented inventions, which are now broadly eliminated in New Zealand (see my earlier post and, e.g, here and here). The bill was drafted by the Select Commerce Committee after consulting the country's (open source) software industry and their associations. After the decision speculations arouse as to whether the committee might have been swayed more by the strength of lobbyists than by rational arguments and it was even spread that the committee accepted an argument that there is no "inventive step" in software development anyway. Presently, the various lobby groups are in still heavy arguing for and against the software patent ban.

Now, according to this post, Mr Jay Daley, Chief Executive of New Zealand Domain Name Registry, defended this decision with some remarkable thoughts regarding inventive step of software-related inventions and the apparent similarities of software and books/films as well as the conclusions to be drawn from this finding:
The actual argument is that the "inventive step" is in the algorithm; the code is simply the implementation of that "inventive step". As algorithms cannot be patented in New Zealand, it follows that code should not be either.

A good way to think about this is in comparison with books/films and plots. There are multiple books/films with the same plot - Hollywood depends on that. Each book/film has copyright, whereas the plot of a book/film cannot be patented. Software is just the same - an encoding of an underlying algorithm, which should not be patentable, just as movie plots are unpatentable.

Software attracts copyright as does books/films and copyright is a perfectly clear and defensible legal right. In jurisidictions that allow software patents, software is the only art that can attract both copyright and patents, an unnecessary duplication.
There may exist good arguments for copyright protection of software and against undeserved patent protection. However, I have the feeling that the argument that "the plot of a book/film cannot be patented" might not be one of them. In fact, the plot of a book would not be considered statutory subject-matter as required by patent law in any known jurisdiction, since apparently there is absolutely nothing technical about a film plot.

With software, or rather software-implemented inventions, it is slightly different. Mr Daley argues that since algorithms cannot be patented in New Zealand, software (i.e. code) should not be either. He is perfectly right insofar as algorithms "as such" cannot (and should not) be patented - not only in New Zealand but in all relevant industrialised countries in the world. He is also right to demand that software should not (and cannot) be patented either. Even thought the term "software patent" is somewhat misleading, however, software patents are neither about patenting algorithms nor about patenting software but exclusively about patenting inventions, independent of the concrete form the invention became manifested in. That is, when talking about inventions, the substance of the invention has to be considered over the particular way it was claimed, which is, for instance, totally in line with the current practise of the UKIPO as expressed in the Feb 2008 Practice Note in reaction to the Astron Clinica decision of the UK High Court.

Such misunderstandings, by the way, are the reason why the term "software patent", which is popular among anti-patent campaigners as a political combart term, should better be substituted by the more precise and clearer term software/computer-implemented invention.

Further, as to the bemoaned unnecessary duplication of IP rights available to software-implemented inventions, the subject of protection of copyrights and patent rights relates to different intellectual levels, since the program code is always copyright protected, while the technical invention implemented by that code may be patentable, depending on as to whether or not the invention is considered novel and inventive over the relevant prior art. That is, copyright protects the expression of the invention, while a patent protects the invention itself.

Such dualities are not only limited to software-related inventions but relate to almost any invention that is described in an operation manual, technical specification or the like, the latter being a copyright protected specification of the former, much like with program code and inventions. The only difference is that program code itself may in turn be interpreted by a computer and thus represents an implementation of the respective invention (i.e. the code transforms the computer to realise the intended invention), which makes it a bit harder to recognise the structural difference between code and invention as compared to the rather apparent difference between specification and specified machine. 
The intellectual problem of IP duality of software is also addressed in a rather entertaining post of Mr Gene Quinn in his IP Watchdog blog, though from another perspective.