Kevin Kelly recently wrote an interesting article he titled Simultaneous Invention. I highly recommend that you read it and draw your own conclusions, but if you are to lazy for that, I’m going to give you short summary and my un-informed commentary. The whole bit is essentially a commentary on Malcolm Gladwell’s article in the New Yorker which talks about a curious company called Intellectual Ventures which is essentially a patent mill. The company employs bunch of smart people with expertise in various scientific areas and pays them to sit in a room all day and brainstorm novel ideas. The best ones are fleshed out to the point where they are patentable and then submitted to the patent office. Their income comes from licensing these patents to various companies at competitive prices.
The whole business works because big inventions and discoveries usually happen simultaneously around the same time. The popular notion about scientific progress is that the great ideas are usually work of a single man – a genius of some sort, who comes up with a revolutionary concept ahead of his time, and then faces the hard task of trying to convince skeptical scientific society that his/her claims are correct. This is however far from truth. Much more often, scientific research is more like a race. In academia we call this getting scooped:
An event when someone publishes a paper on the same exact topic you were working on is not uncommon – in fact it happens very often, especially in highly saturated fields. Unless your working on something really, really obscure or esoteric, you can get scooped. And few people hand out grant money of obscure stuff – so you can pretty much bet your sweet ass that most of your peers are working on very similar stuff at least half of the time. Gladwell shows that this very same pattern holds of all the big, revolutionary inventions and discoveries as well:
A hundred and forty-eight major scientific discoveries that fit the [simultaneous invention] pattern. Newton and Leibniz both discovered calculus. Charles Darwin and Alfred Russel Wallace both discovered evolution. Three mathematicians “invented” decimal fractions. Oxygen was discovered by Joseph Priestley, in Wiltshire, in 1774, and by Carl Wilhelm Scheele, in Uppsala, a year earlier. Color photography was invented at the same time by Charles Cros and by Louis Ducos du Hauron, in France. Logarithms were invented by John Napier and Henry Briggs in Britain, and by Joost Bürgi in Switzerland. “There were four independent discoveries of sunspots, all in 1611; namely, by Galileo in Italy, Scheiner in Germany, Fabricius in Holland and Harriott in England,” Ogburn and Thomas note, and they continue:
The law of the conservation of energy, so significant in science and philosophy, was formulated four times independently in 1847, by Joule, Thomson, Colding and Helmholz. They had been anticipated by Robert Mayer in 1842. There seem to have been at least six different inventors of the thermometer and no less than nine claimants of the invention of the telescope. Typewriting machines were invented simultaneously in England and in America by several individuals in these countries. The steamboat is claimed as the “exclusive” discovery of Fulton, Jouffroy, Rumsey, Stevens and Symmington.
Kevin Kelly points out that even the idea of running a patent mill is neither new, not unique. He points out that Walker Digital does exactly the same type of work as Intellectual Ventures. They both were funded independently around the same time – and thus fit the very same pattern of simultaneous invention. And if there are two successful companies like that, there might be more.
Which brings me to my point. Whenever we discuss software patents we usually preface our arguments by saying “while patents may work for other fields of science, computer science is different”… The above clearly illustrates that the patent system is not really working that great for the other fields of science either. It is essentially a race – and the only way to win it, is to patent early. There are whole companies that base their income on creating and licensing patents without implementations. Big companies know this game and they patent as soon as they can figure out something is patentable. Independent inventors however, have very little chance to compete in this race – especially if they are unfamiliar with the way it works.
This is goes double for business process and software patents. Right now people are patenting trends and methodologies. Anyone keeping an eye on the software industry can look at two emerging technologies and say – hey, if you put this two together you can really improve the way we do things now. Then someone goes ahead, files for a patent and instead of a net gain for everyone we now have a hundred of independent projects being put in legal jeopardy because they had the misfortune to notice a similar opportunity to improve their process using existing technology thing around the same time.
When we are talking about genuine inventions, we can summarize that more than one researcher will discover them around the same time. But when we talk about patenting methods, or processes the chance of simultaneous discovery grows exponentially, if not faster. And I’m talking about actual valid patents which have no prior art (since IMHO more than 90% of software patents usually is just thinly disguised old technology that existed since the 80’s). Take a thousand computer scientists, present them with the same problem, and look at what they come up. While their solutions will vary greatly in implementation details, complexity, and presentation conceptually they should be an overwhelmingly great amount of overlap. They will likely use the same algorithms, same design patterns, model their data in similar way and use very similar design paradigms. If one of these scientists would patent his software, majority of the remaining 999 scientists would be found infringing.
Software patents are silly thing. The sooner we understand this, the better for the software industry as a whole.
RSS Feed
/dev/random