Eudoxa

Background

While software is not patentable per se in the EU due to the Munich accord, there is a number of patents awarded to CII - Computer Implemented Inventions, devices where computers and their programs play a key role. This corresponds to a de facto regime of patentability, but not well protected or defined by EU law. The recent debates surrounding the CII patent directive hinge on whether to accept the patentability or to push it back only to physical inventions.

Intellectual property is critical for EUs hopes to grow into a knowledge economy. There is a need to harmonise the existing patent systems of EU to ensure cooperation and fair competition. The industry seeks transparency and efficiency. More deeply, when the basis of the economy is knowledge and creativity, then intellectual property corresponds to the means of production. At the conference there was a repeated worry that competitors would use a loose IP regime to get ahead themselves, gaining the fruits of European innovation while keeping their own inventions locked away. Competition from low wage, high tech economies with a stronger basis in IP would be most unwelcome; EU doesn't want to become a western suburb of Shanghai.

But the CII directive has triggered a fierce debate about software patents, with an unfortunate polarization of opinions and emotions. What would a few years back have been a rarefied legal discussion has become a hot topic among techno-pundits, free software advocates, politicians eager to save money and the traditional left seeing a chance to strike at capitalism.

The worry in the industry is that the value of software components is going to be reduced to zero due to weak IP protection. If there is no protection beyond copyright (which is easily circumvented), then there is either no incentive to develop advanced products since development costs cannot be recouped or products need to be designed as black boxes with strong DRM technology to keep their secrets safe. The consumer or producer benefits in neither situation.

This might be too pessimistic, since there are clearly unprotected goods that have finite value and are developed anyway (such as free software). But the inability to recoup development costs is still a major problem for any hi-tech industry: it is easy to copy results once they exist, and R&D costs can be extremely high for the kind of technologies envisioned for digital Europe. Limiting patentability might force much software to remain "handmade": rather than being developed industrially, it would remain produced individually at higher costs. Development would be slow and uncompetitive.

CII and the Nature of Inventions

One part of the discussion centred on the key issue of the increasingly blurred line between machines and the software they run.

Peter Dreiert from Siemens pointed out that the point of CII is to see software as an integral part of a device, rather than a separate thing embedded into it. This makes the software inseparable from the device, and hence patentable

In the old modernist paradigm, software was seen as something embedded inside the hardware. Now we are seeing them as inseparable parts of the same system. In the future, it is not inconceivable that we will see hardware as being embedded within software, a bit like how today’s computers often are just platforms to run and process our software and data, with the hardware secondary to the software side. But for the present debate, CII is problematic enough.

Several CII systems were showcased. On one end of the technological spectrum was the Scania truck, containing over 80 microprocessors running tasks such as brake control, steering, fuel injection, navigation systems and adaptive headlights. More than 50% of the patented inventions in a Scania truck are CII Especially the new functions and improvements are solely implemented as computer programs. Not being able to patent them would limit them severely. A Scania representative pointed out that a likely consequence of having software more weakly protected would be for the company to divide its inventions: a first class group of patentable mechanical inventions and a second class of computation inventions. The company would be reluctant to share information about the second class with other companies, reducing interoperability.

On the other end of the spectrum was the skin cancer detection system from Astron Clinica, a Cambridge start-up. Using a camera connected to a computer, the system could "look into" the skin and help doctors determine whether a mole was merely a mark or melanoma. Here the key was how the software tied together image input and medical decision-making.

The idea that an invention must "control a force of nature" to be patentable (suggested by the rapporteur Michel Rocard) is problematic even for the Scania trucks. An electronic traction control system literally controls forces, but what is innovative in a particular design is how they are controlled. This is where the software aspect becomes crucial and cannot be taken away. In the cancer detection system there is an interaction with the outside world, but the truly innovative parts that have developed from research and development is how to tie together the input with the output - again the software.

As Rudy Provoost from Philips put it, "Software at heart of the tech revolution". He exemplified with the development of magnetic resonance imaging, a research work that spans many decades and tens of thousands of researchers, papers and patents. Many of the key developments have been software enabled: the basic physics has always been in the open, but how to fully make use of it requires clever software. Had patents not existed the research would not have been as well funded (since much academic research is funded by patent revenues), and second the incentives for the key software R&D would not have existed. Patents help hold together the large web of technological interdependence that underlies advanced technological systems.

CII Patents and SMEs

Small and medium enterprises (SMEs) are at the core of the patent debate. Anti-patent activists claim that software and CII patents pose major risks for SMEs by raising barriers to entry, enabling major companies to control the market through defensive patents (often trivial ones) and forming patent sharing oligopolies. Several SMEs have also spoken up against CII patents.

But is this due to widespread concern among SMEs, or just a few activists? It should be noted that about 60% of SMEs have business models dependent on patents. When discussing with SMEs, it becomes clear that the overall views are much more in favour of IP than might be expected.

The barrier entry problem might actually be reverse: in order to get venture capital funding, SMEs need patent portfolios. Investors need the security they bring, and they help clarify the investment needs of the SMEs. A typical example was Astron Clinica, who gained 7 million dollars of venture finance based on its patents. Without protection of the patents many SMEs would have a hard time entering the market in the first place.

There are many other market entry problems for SMEs, such as certifications needed for medical applications. Given the costs of certification, small scale software developers will not have the chance to compete anyway, regardless of the software type. Larger projects will have the resources, but they would also have the resources to pay for patent licences.

The solution to this is not to ban CII patents (or certification requirements), but rather to make them cheaper.

Patents and Innovation

The key point in the debate is whether CII patents are good or bad for innovation. The anti-patent lobby claims they limit innovation, the pro-patent lobby claim lack of protection limits innovation.

If software patents are so bad for innovation and small companies, then one would expect America to lag behind Europe or other areas with no or weak software patenting. This does not seem to happen: quite the reverse, the US has a dynamic software sector and the EU does not seem to compete as strongly in pure software.

Patents were after all created to support innovation by ensuring disclosure of innovations in return for a temporary monopoly. Historically there is strong evidence in favor of this working well in general: patents produce a pool of available ideas that does stimulate further invention. Practical issues such as length of protection, the limits of patentability and how they are administered of course have to be addressed. In fast developing fields such as software the traditional 20 year length is seen by some as unduly long, while gene patents have been criticised for not holding a sufficient level of inventiveness.

These are all issues that can be decided by political compromise - if there is an open debate about them, and a realisation that patents serve to stimulate innovation. If the assumption from the start is that patents serve to retain the status quo then there will be no attempts to find constructive legal frameworks and instead efforts will be spent on dismantling or protecting the system.

Conclusions

The industry vision is a strong digital Europe, made competitive through incentives to innovators and protected from copy-cats. This is an appealing vision to the EU politicians and fits in well with the Lisbon agenda. But the vision needs to be shared beyond Brussels. The patent debate is heavily polarised, far more than it needs to be.

One reason for the polarisation is the lack of communication between technologists, the legal professions, industry and politicians. Misconceptions about what is involved in patenting, patent enforcement, patentability and especially the motivations of the other players are very common in the popular debate. Much of the open source movement has become painted with the same brush as the more radical free software movement.

Another reason is that the anti-patent movement has become too linked to anti-free trade movements and the traditional left (there are many other components of it, of course). What was originally an entirely independent question gained allies more easily on the left than on the right, which in turn opened political and organisatorial channels for it to influence NGOs and politicians. This in turn tended to "reward" the anti-patent groups that organised in ways that fitted in, turning it more and more into a standard left-right issue.

Many of the concerns voiced by IP-sceptics are concerns of industry too. There is a need to prevent abuse of patents such as trivial patents granted too easily, patents with prior art or drawn out and uncertain processes. A commonly stated concern was to avoid the "American extreme", where patents are "allowed for nearly anything" and litigation becomes the norm. This is nothing the EU industry wants, and there was great support for strict demands of global prior art checks, technical contribution and transparency in the process.

Everybody agreed that merely accepting the directive in itself would not guarantee a healthy CII sector. The patent system needs to be a healthy tool to support innovation and competition. If the CII directive passes, the next step is to make sure the EPO is able to handle the new patents.

If the directive ends up in a third reading and is further delayed, there is risk that the uncertainty caused will hurt European innovation. The only group who might benefit from an unclear or anti-patent directive would be the patent attorneys, who would be in greater need than ever in such a regime, and competitors from regions with more clear IP regimes.

The software patent struggle is not over yet. Regardless of the parliament decision in July, the deeper issues of patenting and other forms of IP will be debated for years to come. But industry cannot be complacent that the political sphere understands their concerns and needs. They have to be articulated and expressed creatively.

As Kuan-Tzu wrote:
"If you are thinking a year ahead - plant seeds,
If you are thinking 10 years ahead - plant a tree,
If you are thinking 100 years ahead - educate the people."