A couple weeks ago NPR's Planet Money and This American Life had some really great episodes about the broken patent system. These are great stories for people who don't understand why patents are a problem, but they overlooked a couple of crucial points. First, go check them out. Planet Money: When Patents Hit the Podcast, and This American Life: When Patents Attack... Part 2. This American Life's part two was largely a rehash and update on part one, available here. In many ways, the earlier story did a better job of illustrating the absurdity and obviousness of software patents, so if you have time, listen to that one too. Three points to make on this topic:
- Software patents have regularly been granted on completely obvious ideas.
- Patents have been applied to completely different innovations that generate the same result.
- In software, value comes from the network, and not the code or idea being patented. Let's take a step back and discuss the original purpose of patents.Ius Mentis has a good summary: > The historical purpose of the patent system was to encourage the development of new inventions, and in particular to encourage the disclosure of those new inventions. Inventors are often hesitant to reveal the details of their invention, for fear that someone else might copy it. This leads to keeping inventions secret, which impedes innovation. The patent system was created so that inventions would eventually make it into the public domain where they would provide benefit for everyone, and it did this by allowing the inventor to control manufacture and distribution of their inventions. Over the years, patents came to have 3 principles, tests to determine whether they were patentable: they must be Novel, Non-Obvious, and Useful. I'm not going to argue whether or not any "invention" is novel -- it's really hard to say who came up with the idea of a podcast first. There's a famous quote, "Good writers borrow, great writers steal" -- but we don't even know who originated that! Was it Oscar Wilde? Picasso? T.S. Eliot? But patents are about engineering, about putting together ideas into a functioning machine. Copyright is about expression. If you're spending a huge amount of time and effort putting together a functioning machine that solves a problem, I think you should be able to benefit from patent protection for a machine that uses your design. Not for one that uses a different design.
Non-Obvious is a different story. This is the largest problem with the patents being granted in software. What is non-obvious about being able to purchase from an e-commerce site with a single click? Apparently a huge number of patents are being granted to cover pretty much exactly the same software "inventions" -- and that to me seems like proof of their obviousness -- if they were non-obvious, why would hundreds of people come up with exactly the same idea and file very similar patents on it? In the early 1990s, I came up with a novel idea, actually for a novel (I was a wannabe fiction writer at the time). I had dreamed up having a ubiquitous grocery delivery service. I dreamed up a bunch of innovations that have since come to pass: my protagonist worked at home, logging into something pretty much exactly like Second Life, worked in a virtual office. Never left his house -- he used a service very much like Home Grocer.com back in the Web 1.0 days, or Amazon Fresh today. People were afraid to go outside because it was too dangerous -- sun exposure, kidnapping, all sorts of things to be afraid of, compared to the virtual world. So the people who went out and about were the delivery people, who were actually trained spies working for something similar to the NSA, or CIA, keeping everybody fearful in their houses so the people in power were free to run everything as they wished... but I digress. Did I invent virtual worlds? Certainly not. Did I invent the idea of home grocery delivery? I had never heard of it before -- if I had patented that idea, could I now be raking in a license fee from Amazon and Safeway? This is all absurd. Of course not -- obviously these were obvious ideas at the dawn of public use of the Internet, and imagined by dozens of science fiction authors before me. And yet these kinds of patents are being granted all the time, and used by trolls to extract licensing fees from people and companies who do the hard work to... make things work.
Useful, or idea?
The NPR stories touch very briefly on patents being useful, but completely miss the far bigger point. Right on the Patent and Trademark Office's web site, it states that: > A patent cannot be obtained upon a mere idea or suggestion. The patent is granted upon the new machine, manufacture, etc., as has been said, and not upon the idea or suggestion of the new machine. A complete description of the actual machine or other subject matter for which a patent is sought is required. This is the meaning of Useful in the patent tests -- is there sufficient detail in the patent to use it to obtain the intended result? And this is where Planet Money completely missed the boat on the Podcasting patent. Jim Logan is claiming he has a patent on the idea of podcasting, not the implementation. The implementation involves technologies called RSS (Really Simple Serialization), media enclosures, codecs to encode the video and audio, hypertext transfer protocol (http), and a bunch of other technologies Logan had absolutely no hand in creating. And the patent office itself says you cannot patent a "mere idea." So why exactly does any podcaster owe anything at all to some guy who invented something completely different (a series of audio programs on cassette tapes), that failed in the marketplace? I do think there are cases where software patents are appropriate, possibly. These would involve coming up with very specific, intricate algorithms for doing specific things. Codecs and image pattern matching comes to mind -- coming up with unusual methods for taking our analog world and creating a meaningful digital representation. It would still need to pass the "non-obvious" test -- generally I would think that this means it's the result of substantial engineering, and incorporate a new technique that current professionals in that field wouldn't think of at first glance. But if somebody else comes up with a different method for accomplishing the same end, I don't see how that infringes on a different patented method -- it's different. And even for these complex innovations, there's another factor in software that makes patents less valuable, less important.
It's not the idea that is valuable, it's how many people use it
The Internet, and open source, has changed things in some profound ways. Jim Logan's attempt to monetize his patent by suing podcasters illustrates this. Podcasting has become such a large and thriving thing because the barrier to entry is so low. Get a microphone, record you and your buddies talking about something, throw it up on the Internet with an RSS feed, and presto, you have a podcast. Anyone can do it. And thousands of people do. The reason podcasting is a thing is because lots of people do it. If every podcaster and every listener had to pay a license for all the technologies involved in podcasting, it never would have taken off in the first place. The reason the Internet took off the way it did is because after you have your connection and your computer, it's pretty much free. We have a thriving ecosystem of web hosts who got started by downloading open source software for free, and figuring out how to use it. We have hundreds of millions of websites because the financial barrier of entry is extremely low for somebody willing to figure out how to make one. And this has led to the Internet being successful -- because there are people there. On the Internet, ideas are worthless. Nobody pays for an idea, there are countless ones out there. People pay for services. They pay companies like mine to help them put stuff out so they can interact with other people. People may pay for good writing, for good content, but rarely for the ideas in them. Even the online marketers with "information products" are selling ideas you can easily find elsewhere for free -- the main value they bring is in packaging them up for easy consumption. People pay for getting things done. That means for an idea to succeed, it has to have the best value proposition -- if there's any cost to it it needs to be substantially better than what you can get elsewhere for free. And that's why for even those software machines you could patent, you might be better off leaving unprotected. Because unless you can invent something substantially better than everyone else out there putting stuff out there for free, if you try to charge for your idea, you'll just get ignored. The Internet is powered on free. All the major technologies that power most of the main sites are completely free. Free web servers dominate the web world. Free operating systems and databases power them. You didn't pay for the web browser you're reading this story with, or the search engine you found this page with. And yet there are lots of people making money here. Google, which led 2/3rds of the people reading this story here. Rackspace, which happens to host this site at the moment. Comcast, that connected me so I could post it. Apple, if that's who you bought your device from. In the open source world, there are two major kinds of licenses: Copyleft, and BSD. The Copyleft licenses are epitomized by the GPL, the GNU General Public License, and its main purpose is to protect the rights of people using GPL software to be able to put it to work however they choose, with almost no limitations, other than you cannot restrict what other people do with it if you make changes to it. The BSD (Berkeley Systems Distribution) licenses are even more free in that you can add restrictions. These days the BSD style licenses are becoming far more ubiquitous, because there's pretty much no restriction in any way. Often one restriction people do add is requiring you to maintain credits, attribution to those who created it in the first place. When the marketplace is the Internet, when all else is equal, free wins. BSD wins over the GPL because it's more free. GPL wins over anything patented for the same reason. If you want your technology to get widely adopted, you have to make it as free as possible, or people will choose a freer alternative. The one thing many of the freer licenses preserve is credit -- they preserve your name. And in this world of free, that's far more valuable than patents. Because it proves you can do the work, get the job done, and that is what people are willing to pay for, not your silly idea.
Ok, I've ranted long enough. But this really illustrates why the patent system is broken, at least for software. Far too many patents are granted on obvious ideas, without any useful implementation details, in direct contradiction of the tests the patent office is supposed to use. And even when there are patentable software methods, using patents inhibits adoption and limits the potential market for the inventor. The only people that benefit from software patents are trolls and lawyers. And as This American Life concluded, we're all worse off because of how much patents stifle and deter would-be entrepreneurs who actually want to put an idea to work.