Groklaw's high definition on software patents

With the European Commission heading towards a second reading of its patent legislation, there's still more heat than light in discussions on the subject. The UK Patent Office (UKPO) has of late been an active force for good, reacting to the confusion and double-talk by setting up workshops this spring for interested parties, working through the issues, trying various test cases and potential definitions, and listening to what people say. You might almost accuse the organisation of pragmatic common sense and a willingness to do the right thing, were that not the civil service equivalent of treason against the state.

One of the things that came to light during the workshops was that while the concepts under discussion needed tight, unambiguous and accurate definition, actual and proposed legislation was imprecise and unclear. Some of this was due to two highly technical professional groups attempting to converse in their mutual second language — plain English — but both lawyers and software writers found that everyday ideas are remarkably hard to pin down. In particular, the answer to the UKPO's question "Should software be patentable?" was often "What do you mean by software?". Other phrases that promised much but fell apart under scrutiny included technical contribution — which is what an invention must create if it is to be patentable.

Over at Groklaw, the SCO-watching site that is gently expanding into a general talking shop on the legal realities of free and open source software, such questions strike home. Taking as a starting position that pure software should not be patented, a group of Groklegals set to work on a legal statement to that effect. One public consultation and thirty drafts later, the statement is available for comment.

It is a solid piece of work, written in the light of the previous definitions that the UKPO workshops found unacceptable. Those interested in what defines and separates software, information and their carriers should go and read through the closely argued details on Groklaw, but the end result is clear. You can only patent things that make physical changes in the world: these can contain software, but that software is not of itself patentable.

For the purposes of the patent, the software is a magic box that does things: the contents of the box are not covered and cannot form part of the patent. If I were to invent a device for finding mushrooms in a forest — as a mycophile, this an idea I've actually spent some time on — then it would doubtless contain software. For example, I could develop an algorithm that sorted and counted spores in an air sample through video analysis, but all the patent would care about would be a black box that produced a spore analysis to ring a bell, point an arrow or whatever. That would be strong enough to protect my invention — even if someone came up with a different way to sort spores — but wouldn't stop anyone else from using video analysis to check for dandruff or cocaine particles at parties. That seems to sensibly encompass the balance between protection and disclosure that the patent system is designed to strike.

A key point — and one already enshrined in European Patent Convention Part 52 — is that information itself is never patentable. As MEPs Lichtenberger & Frassoni said, "The dividing line between the material and immaterial world, and hence between what is patentable and what is not, can be defined with legal certainty. Once a physical signal is digitised, it becomes symbolic information, which can be manipulated in an abstract fashion in software, with no possible technical effect."

The whole definition does excellent work in pegging down software with solid axioms, and delineating the boundaries of what can and cannot be patented. Even if you believe that software requires some form of patent-like protection but that the details should be different to reflect its unique properties, the definition stands as an essential first principle. Without that, discussions about reform and renewal of intellectual property as it applies to information technology will be doomed to circulate in the spin and blather of those who'd rather see the status quo continue than risk losing what power they have.

What happens next? With luck, the Groklaw definition will go through another round of public analysis and fine tuning, and will attract attention from the patent offices and legislators of Europe and beyond. It is not and will never be perfect; there is always ambiguity in definitions of the real world, but it clearly does a better job than anything that has gone before. And, fortunately, it is there for us all to build on. There is no patent on having good ideas.