2945th session of the Council of the European Union - Com...

Karsten, great: define "software patent" first.

What OS activists appear to have huge trouble acknowledging that the EPO actually refuses quite a lot of patent applications on the grounds that that they refer to computer programs or business methods "as such". And many other applications aren't even filed in the first place because patent attorneys warn their clients that their software-related ideas have very little chance of being granted a patent.

The issue is not so much whether "software patents" should be patented (although that's another debate), but what a "software patent" actually is.

Patents have been granted on technical processes since the dawn of the patent system. A patent claim on a technical process reads as a set of instructions on how to perform that process (for example: "heat up substance to 150°; mix it with substance B; cool it down to 50°; if there is a solid deposit, extract it by centrifugation..."). Such a set of instructions is an algorithm, so such a patent could fit some people's definition of a "software patent".

Of course, you could try to respond to this by defining software as a set of instructions to be performed by a programmable computer. But this overlooks the problem that most technical processes nowadays are computer-controlled.

Again, you could try (as some FFII-backed MEPs did during the CII patents directive debate) to restrict patentability to only those technical processes which make use of "forces of nature". The first trouble here is that even "pure software" makes use of "forces of nature" within the microprocessor. Even marketing processes make use of forces of nature governing our brains. So this could become quite a meaningless restriction, and even if it was applied restrictively it would leave, for instance, sound or image processing methods in a grey area.

The current approach of the EPO's Boards of Appeal, which is confirmed by the judges of most EPO states, is that for a computer program to be patentable there must be a "technical effect" going "beyond the normal physical interactions between the program and the computer". It is an interpretation which of course leaves a lot of room for debate, but do you have a better alternative?

Of course, considering how difficult it is to pin down what a "software patent" actually is, it is difficult to avoid that other debate: why bother? why shouldn't pure software be patentable? what's so special about it in comparison to other areas of technology?

PL Hayes wrote: Why shouldn't technical culinary inventions informed by chemistry and physics, such as Heston Blumenthal's stuff or my spooning hot butter over the yolks of frying eggs be patentable?

Why indeed? As a matter of fact, food products and food preparation methods are patentable, and have been so for a long, long time. While Mr. Blumenthal does not appear to have attempted to patent any of his techniques, his no less famous "molecular chef" colleague Ferrán Adriá has pending patent applications (for a dry-frozen coffee foam) in a number of countries.

Indeed, a few important lessons can be learnt from history in that particular field. First and foremost: it is not because something is patentable that people are going to patent it. If patents are uncommon in the field of "haute cuisine", it is not because they were not available as a means of protection, or because chefs are particularly generous with their knowledge. Indeed, chefs do jealously guard their intellectual property. However, they've generally preferred to protect it though trade secrecy (like...um...proprietary software), rather than through patents. Why? Well, patents are expensive to obtain and even more expensive to enforce. For patenting to be worthwhile, you need an industrial-scale exploitation of the invention. Enforcing your patents against a handful of other chefs is hardly going to be profitable. For starters, you'd have to prove that your competitors are using your patented techniques, something that may not be all that easy without access to their kitchens. And enforcing it against the legions of amateur cooks is not only going to be extremely unprofitable, it'd also be a PR disaster (just like suing amateur programmers). Better to sell recipe books and cash in the copyright royalties...

Patenting food products and food preparation methods is thus interesting almost only for large food companies, like Nestlé, Unilever, Kraft or Sara Lee. And, by Jove, they do patent their processes and enforce their patents against each other.

This choice between patent, trade secret, and publication is one to be considered by inventors in all fields of technology. For most software-related inventions, I wouldn't necessarily recommend the inventors to file a patent application.

So if the calculation speed and efficiency improving software invention I posted in the Meaningless Banter section of an obscure science-related forum recently can be patented

But it can't. Since you published it, it is no longer new, and thus cannot be (validly) patented. Not even by yourself (except in the US, where the inventor can benefit of a one-year "grace period" between publication and patent application).

Even if your publication was so obscure that the patent examiner wouldn't find it (and patent examiners do a better job of searching prior art than they are generally credited with), you can place quite a safe bet that, if a patent was granted and there was litigation, somebody would find your publication. Even more so if you were yourself the alleged patent infringer.

I want to see some damn good reasoning backed up with solid empirical evidence before I'll accept that the negative consequences of software patent eligibility are justified.

I also would like to see an open, informed debate on the matter. I certainly keep an open mind the patentability of "pure software" or "computer programs as such" (I used to be very much against). However, I must also note that a lot of people who express their opinion in this debate (including quite a few of those economist) are very ignorant about the patent system, and appear to have reached their conclusions before even starting to consider the facts.

“what's so special about it in comparison to other areas of technology?” I think that should be obviousi>

I don't see it's quite so obvious. Could you explain why?