At the IT conference I am attending, the organisers conducted a straw poll of who opposed the impending Euro-software patent legislation. Everybody I saw was against it; The unambiguous conclusion is that software patents are harmful and the ‘economic majority’ of Europe’s IT industry who oppose software patenting.
In essence, a patent grants a monopoly on an idea. Patents were invented hundreds of years ago to protect the investments made by inventors. In return for publishing the designs of an invention (e.g. the plans of your contraption), and agreeing to license the idea for a standard fee, the government would protect that monopoly. This worked back in the 18th century because other inventors could study the designs in the patent-office and invent better machines.
The assumption of many MEPs appears to be that since patents appear to work quite well for contraptions they must also benefit software inventions. However there are critical differences between the world of software and the mechanical age that spawned the patent.
A central principle of computer-science is idea-reuse: Software languages are built to encapsulate ideas and re-use them. Any programmer can download ‘modules’ - collections of working functions. Some functions implement simple, single-minded ideas. Others amalgamate thousands of the simpler functions and can do complex things. The programs that you or I use in our day to day work are made of thousands of these complex functions, so it’s not unusual for any substantial work to contain millions of potentially patentable ideas.
Unlike a car or a steam-engine, a useful program is almost never a totally original work because re-use of ideas is so fundamental to what we do. The programmer who works at the leading edge of technology will almost certainly not be aware of which of the millions of ideas in a program are patented, and therefore any substantial program is likely to blunder into the patent minefield.
Software patents drive up the cost of innovation. Since any program is potentially infringing then all commercial software must be scoured for potentially infringing code before it can be sold. Of course, companies could opt to remain ignorant of their potential patent infringements, but if convicted of a patent violation the costs are enough to bankrupt most firms. Most companies would rather spend 10% to 20% of their annual profits on lawyers to audit their products.
In addition, rather than re-investing their remaining profits back into development, companies will feel pressured into building their own patent arsenals. A patent typically costs between £30k and £100k, so the more money we send to the patent office, the fewer productive computer-scientists we can afford to employ. This is why software patents cost jobs.
According to a friend who moves in parliamentary circles, there is at least one LibDem MEP who has not yet made his mind up about software patents. He told the FFII that if they can present two hundred signatures opposing software patents he would vote against the legislation. We managed to collect more than 250 from the delegates in the space of half an hour: a unanimous response against software patents
I believe that a vote for software patenting is misguided and harmful because it makes the best-practices of software development into something that is legally risky. In theory, any program that I write that is substantially useful has the potential to infringe somebody’s patents. The risk of accidental patent infringement will turn today’s playing-field into a minefield.
Software patents will harm me, and they will harm my customers. If any MEPs are readinf this, I request that you against software patents.