Software patents: yes or no?

The IPKat is pleased to note that certain members of the anti-software patent lobby have been paying attention (for examples see comments on the posts here and here). He (or rather one of his amenuenses) has been pondering the issue of "software patents" and their effect for some time, without coming to any definite conclusion.

Right: The IPKat tries his paw at a bit of software coding.

The anti-software patent lobby, exemplified by the (rather pretentiously titled) Foundation for a Free Information Infrastructure, have resisted attempts to harmonise patent law throughout Europe by opposing, and ultimately killing, the Computer Implemented Inventions (CII) Directive. The general view, which caused the death of the CII, is that patents should not be granted for computer software. Extensive lobbying of MEPs resulted in a very confused situation, and the CII was abandoned largely because it turned into a bit of a farce with lots of unworkable and contradictory amendments. This was, in the IPKat's view, largely because nobody really understood what they were trying to do. Practically the only MEP who had any degree of understanding on the issue was Sharon Bowles, the only MEP who is also a qualified patent attorney (see her comments on the issue here).

The problem was nobody (even those with some knowledge of IP) could agree on exactly what needed to be changed, let alone how it could be done. While the law could be changed to exclude computer software (more than the current 'as such' exception), how could this possibly be achieved without having effects worse than the current law? Would it be possible to exclude the practice of claiming a computer program (as became possible after this EPO decision)? If so, would method claims that cover computer programs in part be allowed? Should it be a contributory infringement to supply a computer program that was essential to putting an invention into effect (as was found in Menashe v William Hill)?

It is possible that everyone was barking up the wrong tree, and that it is not in fact possible to define the boundaries of what should and should not be patentable in a way that would please everyone (or at least most people, it being impossible to please everybody). Perhaps Section 1(2) is not the place to be looking for a solution.

Section 60 of the UK Patents Act 1977 sets out what does and doesn't constitute an infringement of a patent. In subsection (5) there is quite an extensive list of exceptions to the standard ways of infringing a patent (eg making, using, disposing of). These are meant to be equitable ways of limiting the power of a patent holder against those who should not be pursued for infringement. These include farmers, private non-commercial users, medical researchers and others.

Perhaps this is the place to look for a solution. Why should the development, use, distribution etc. of computer software itself be classed as infringing? If it wasn't an infringement, wouldn't this solve the problems of those who only want to make new software or modify existing software and don't want to risk infringement? Wouldn't making an exception of this kind help the development of new and innovative software solutions?

Here is a proposal then that might just achieve this. In Section 60(5), insert this subclause:

(j) it consists of the making, use, disposal, offering, keeping or importation of computer software.

This little change would exempt software itself from patent infringement. It would not prevent computer implemented inventions from being patented, nor would it prevent companies from suing others for infringement based on real embodiments requiring patentable software solutions, for example the sale of music players or telephones incorporating patented software-implemented inventions. It would, however, prevent software developers from being sued under a patent for merely developing and distributing software they have developed themselves.

Of course, big companies like Microsoft and IBM would doubtless complain that their rights would be severely restricted by such exceptions. However, a large part of the protection afforded to software such as Microsoft's operating systems is due to the combination of copyright protection together with the almost impossible task of reverse engineering publically available object code into source code that can be made sense of. So, copyright protection (which, to remind readers, lasts for 70 years after the last author's death - a very very long time) together with the law of confidential information, seems quite adequate to protect the investment needed to develop software. After all, a return on investment is what the IP game is all about, is it not?

The IPKat would be interested to see if anyone has any comments about this.