Beware: Europe's 'unitary patent' could mean unlimited software patents

The battles seen in the US over software patents could spread to the UK and the rest of Europe if the unitary patent is allowed to come into force.

This is a verbatim copy of Richard Stallman's orginal article published on guardian.co.uk on Monday 22 August 2011. We've helped the author writing this story.

Just as the US software industry is experiencing the long-anticipated all-out software patent wars, the European Union has a plan to follow the same course. When the Hargreaves report urged the UK to avoid software patents, the UK government had already approved a plan that is likely to impose them.

Software patents are dangerous to software developers because they impose monopolies on software ideas. It is not feasible or safe to develop non-trivial software if you must thread a maze of patents. (See Patent absurdity, Guardian, 20 June 2005.)

Every program combines many ideas; a large program implements thousands of them. Google recently estimated there might be 250,000 patented ideas in a smartphone. I find that figure plausible, because in 2004 I estimated that the GNU/Linux operating system implemented around 100,000 actually patented ideas. (Linux, the kernel, had been found by Dan Ravicher to contain 283 such ideas, and was estimated to be 0.25% of the whole system at the time.)

The consequences are becoming manifest now in the US, but multinational companies have long lobbied to spread software patents around the world. In 2005, the European parliament took up the second reading of a directive that had been proposed by the European Commission to authorise software patents. The parliament had previously amended the directive to reject them, but the Council of Europe had undone those amendments.

The commission's text was written in a sneaky way: when read by laymen, it appeared to forbid patents on pure software ideas, because it required a patent application to have a physical aspect. However, it did not require the "inventive step" – the advance that constitutes a patentable "invention" – to itself be physical.

This meant that a patent application could present the required physical aspect just by mentioning the usual physical elements of the computer on which the program would run (processor, memory, display, etc). It would not have to propose any advance in these physical elements, just cite them as part of the larger system also containing the software. Any computational idea could be patented this way. Such a patent would only cover software meant for running on a computer, but that was not much of a limitation, since it is not practical to run a large program by hand simulation.

A massive grassroots effort directed at convincing the European parliament to change its mind resulted in defeat of the directive. But that does not mean we convinced half of parliament to reject software patents. Rather, it seems the pro-patent forces decided at the last minute to junk their own proposal.

The volunteer activists drifted away, thinking the battle won, but the corporate lobbyists for software patents were paid to stay on the job. Now they have contrived another sneaky method: the "unitary patent" system proposed for the EU. Under this system, if the European Patent Office issues a patent, it will automatically be valid in every participating country, which in this case means all of the EU except for Spain and Italy.

How would that affect software patents? Evidently, either the unitary patent system would allow software patents or it wouldn't. If it allows them, no country will be able to escape them on its own. That would be bad, but what if the system rejects software patents? Then it would be good – right?

Right – except the plan was designed to prevent that. A small but crucial detail in the plan is that appeals against the EPO's decisions would be decided based on the EPO's own rules. The EPO could thus tie European business and computer users in knots to its heart's content.

Note that the EPO has a vested interest in extending patents into as many areas of life as it can get away with. With external limits (such as national courts) removed, the EPO could impose software patents, or any other controversial kind of patents. For instance, if it chooses to decide that natural genes are patentable, as a US appeals court just did, no one could reverse that decision except perhaps the European Court of Justice (ECJ), the highest court in Europe.

In fact, the EPO's decision about software patents has already been made, and can be seen in action. The EPO has issued tens of thousands of software patents, in contempt for the treaty that established it. (See http://webshop.ffii.org/, "Your web shop is patented".) At present, though, each state decides whether those patents are valid. If the unitary patent system is adopted and the EPO gets unchecked power to decide, Europe will get US-style patent wars.

The ECJ ruled in March that a unitary patent system would have to be subject to its jurisdiction, but it isn't clear whether its jurisdiction would include substantive policy decisions such as "can software ideas be patented?" That's because it's not clear how the European Patent Convention relates to the ECJ.

If the ECJ can decide this, the plan would no longer be certain disaster. Instead, the ball would be one bounce away from disaster.

Before adopting such a system, Europe should rewrite the plan to make certain software is safe from patents. If that can't be done, the next best thing is to reject the plan entirely. Minor simplifications are not worth a disaster; harmonisation is a misguided goal if it means doing things wrong everywhere.

The UK government seems to wish for the disaster, as it stated in December 2010 that it did not want the ECJ to have a say over the system. Will the government listen to Hargreaves and change its mind about this plan? Britons must insist on this.

More information about the drawbacks and legal flaws of this plan can be found in unitary-patent.eu.

• You will note that the term "intellectual property" has not been used in this article. That's because the term spreads confusion because it is applied to a dozen unrelated laws. Even if we consider just patent law and copyright law, they are so different in their requirements and effects that generalising about the two is a mistake. Absolutely nothing in this article pertains to copyright law. To avoid leading people to generalise about disparate laws, I never use the term "intellectual property", and I never miss it either.

Copyright 2011 Richard Stallman. Released under the Creative Commons Attribution Noderivs 3.0 license.