Devil's Advocate: Patents are killing software innovation

Individuals should be rewarded for their accomplishments, but patents aren't the way to do it in the software industry, says Martin Brampton. All they do now is prevent small organisations from competing with the big guys.

There is a risk that patents will strangle the life out of software innovation. While IBM, the biggest holder of IT patents, is holding back on obtaining more, Microsoft is making aggressive threats about patenting its developments. Yet the value of patents is doubtful at the best of times.

A parallel debate is being fought in science. Governments have sought increased commercial involvement with university research, but the price has often been proprietary control over new ideas. Many academic scientists are opposing this trend, believing that the advancement of knowledge is a collaborative and public venture.

Inventions are rarely as innovative as the simplistic argument for patents would suggest. They nearly always involve the application of ideas that have become widely known. Practical applications for radio emerged in the wake of developments in theoretical physics, for instance. And with major breakthroughs, there have nearly always been several rival inventors.

Patents are brutal. Even though a number of people may have independently had the same idea, only the first to register a patent gains protection. The rest lose everything. Copyright is a much more flexible right, with no penalties for mere lack of originality.

Software was long thought to be suitable for protection by copyright as opposed to by patent. Now, software companies are filing patent after patent, with Microsoft aiming to move from about 2,000 in 2004 to more than 3,000 next year. The claim by proponents of patents is that without them, there will be no economic mechanism to drive software innovation.

It is hard to take this argument seriously. Imagine patents being applied to works of literature. If the works of Shakespeare or the Bible were to be patented, it would become extraordinarily difficult to write anything. We all commonly use phrases from both without even realising their source. Or in modern literature, would we be prevented from talking about Catch 22 without paying a fee to Joseph Heller?

Without patents, there are plenty of routes to making profits from software. Companies with large resources can compete in fast-moving areas that are less easily supported by collaborative developments. Well-established free software still needs support, and there is huge untapped potential for commissioned development within the open-source envelope.

Thankfully, we have at least some protection against software patents. The redoubtable Donald E Knuth carefully documented a great many algorithms. Some are simple and others are highly ingenious. Knuth was not the creator of them but he recorded and gave penetrating reviews of their efficacy. They can be regarded as in the public domain and are easily referenced through Knuth's works.

We also know, of course, that many ideas long preceded the currently dominant implementations. User interfaces with windows, icons and mouse were dreamed up long ago at Xerox's Palo Alto Research Centre. Most of the mechanisms of the internet were created in the public domain by university staff.

Big software companies have excelled at refining and implementing ideas, then pushing them to a mass market. And there is no reason to suppose that software patents are needed for this to happen. Patents are a means to stifle competition, while much of the real innovation comes from individuals and groups outside the major companies, which are already well placed to withstand competition.

The trouble with having thousands of software patents is that small organisations, whether software companies or in-house development shops, will find it increasingly difficult to write new code without infringing on patents. The argument that the new software was invented independently is no defence against a patent-infringement claim.

Evading the problem may be possible through insurance, which is a direction pioneered by Open Source Risk Management. But the real solution is a fundamental rethink of the role of patents in a world of global corporations. We want creative individuals to be justly rewarded but a system that protects giant companies against competition is in few people's interests.