Software Patents are Like Smoking
Society crossed an invisible but decisive border when it started to allow patents not only on tangible inventions but also on computer software. It must have felt good at the time, but today it is becoming increasingly obvious that software patents do no good. In that regard software patents are like smoking – it started with an experiment to improve health. It tasted quite good and it soon became a fashion statement. But today smoking kills not only those who smoke but also those who breathe nearby. And yet it is very difficult to completely eradicate smoking. Those who smoke stubbornly defend their right to do so. Perhaps there is even an argument to be made where smoking is good in some circumstances. But overall it is a habit that would deserve to be stopped. The same applies to software patents.
~ by Marten Mickos,
CEO, MySQL AB
The rights of an inventor or author need to be properly protected. Otherwise we will soon have no inventions or new creations. But when it comes to software, copyright is quite sufficient protection. In my mind software should be treated like architecture, music, literature, and journalism and not like mechanics, electronics or biochemistry or other traditional industries. Here is my logic.
First of all, in traditional industries producers produce and users only consume. But with software, users produce even more than dedicated producers. The Citibanks and Coca-Colas of the world collectively write more software than the Microsofts and IBMs. Similarly, music and literature is produced both by professional producers and by private consumers. When production is so heavily tilted towards users, patents make no sense. Or are you saying that a 15-year old who develops a great software algorithm at home should do a patent search (among tens of thousands of patents) to verify that he/she is not infringing?
The second argument has to do with determinism. For a system of patents to be functional, there needs to be an easy way for inventors to verify that their inventions do not infringe on other patents. Either the number of patents has to be very limited, or the terminology needs to be so well defined that you can search by keyword. Neither of this is true for software patents. As a result, it is absolutely prohibitive to make a reasonable patent search to verify non-infringement. The consequence is that nobody can be safe. With copyright you will always know if you infringe on someone else’s right because you can do so only by actually copying code. But with patents you cannot know. The system is not practically deterministic.
The third argument has to do with the parameters of the software industry. A patent can take long to get approved, and it may be in force for 20 years. But technology cycles in the software industry are much much shorter. In traditional industries, you may have a very limited number of patents for each product you produce. But a software product may extend over areas that are covered by hundreds if not thousands of patents. In other industries, patenting and patent defense is a small fraction of overall R&D costs. But in the case of software, patents take up a much bigger portion of the R&D budget. As a result, patenting and patent protection becomes prohibitively expensive in the software industry. The only ones who gain are the patent lawyers.
Many companies apply for software patents for defensive reasons, thinking that if someone challenges them with a patent, they can retaliate with their own patent portfolio. But today the software industry is seeing a new breed of companies - so called patent trolls – that have no other business than acquiring patents and then extracting royalties from other businesses. No patent portfolio will help against a troll, because they have no production or sales of their own that you could threaten.
This leads to the “3 Gazillion Peril”. A gazillion in this context is a unit for patent cost in a company. In a small company it can be a million; in a large one it can be hundreds of millions. The first gazillion gets spent on defensive patenting, because you don’t want to be any worse off than anyone else. When that’s done, you will have shareholders requiring you to make the best possible use of your assets (and probably they have heard that IBM makes a billion a year on patents). So you approach your competitors and tell them they have to pay a gazillion in patent royalties to you. But of course they have built their own patent portfolios, so you end up cross-licensing your patents with no money changing hands. In essence, you have now lost your second gazillion. Then a patent troll appears on the scene and demands patent royalties from you. You try to defend yourself with your own portfolio, but to no avail. In the end you pay the troll a gazillion in royalties. By this time you have spent 3 gazillion of your valuable currency, and you have achieved absolutely nothing.
This is why I believe that software patents are bad for anyone developing software, whether user or producer, large or small, closed source or open source. Open source companies and organisations have been the first to smell the danger, but everyone will be hurt. Fortunately we have seen a good debate on the topic lately, and we scored an important victory when the European Parliament rejected the software patent directive in July this year. For more information on this critically important campaign, please visit www.nosoftwarepatents.com which is sponsored by MySQL AB and other companies.