Paul Graham’s essays are usually brilliant, but I found this essay on software patents of software patents to be rather short of his usual standard. He actually asks two separate questions: First, given the state of the law, is it evil for companies to seek software patents? And second, is permitting patents on software good policy?
I agree with him on the first question–a technology company that doesn’t play the patent game opens itself up to the risk of extortion by patent trolls. So I don’t blame innovative companies like Microsoft from acquiring software patents in self-defense. And like Graham, I fault companies that attempt to use their software patents offensively against competitors, as Amazon has done.
However, on the merits of software patents as public policy, his defense strikes me as rather weak:
One thing I do feel pretty certain of is that if you’re against software patents, you’re against patents in general. Gradually our machines consist more and more of software. Things that used to be done with levers and cams and gears are now done with loops and trees and closures. There’s nothing special about physical embodiments of control systems that should make them patentable, and the software equivalent not.
There are, in fact, a couple of things that are special about software. First, the capital required to develop any given “invention” is typically very small. True, a lot of money spent developing software, but that’s because a given software product can easily have thousands of “inventions” embedded in it. What makes software development difficult isn’t coming up with any given invention, but building a complex system out of many well-understood components.
Secondly, and more importantly from a policy perspective, software enjoys copyright protection. Without patent law, there’d be nothing to stop me from copying the “levers and cams and gears” in somebody’s machine. But if I copy somebody’s “loops and trees and closures,” the owner of the software can sue me for copyright infringement. So to justify software patents, you have to show not just that software needs IP protection, but that software needs protection above and beyond that available under copyright law.
I largely agree with the rest of the article: software patents probably don’t directly prevent all that much innovation. But I think he misses the broader lesson. Just because innovation is still possible in a world with software patents doesn’t mean they’re harmless. As Graham admits, software patents don’t particularly promote innovation either. What they mostly do is encourage a lot of pointless rent-seeking. Big companies like Microsoft apply for thousands and thousands of frivolous patents to give them ammunition in patent battles. Patent trolls like NTP harrass legitimate businesses in the hopes of extorting licensing fees. Patent lawyers spend millions of hours looking through old software documentation looking for evidence of prior art for “inventions” that had been common knowledge for decades, but everyone thought too obvious to write down.
If software patents don’t promote innovation, then every hour billed by a patent attorney–and every hour an engineer spends describing his work to patent attorneys–is a deadweight loss to the economy. And as long as patent trolling is a profitable activity, that deadweight loss will continue to increase. Like any large settlement of a frivolous lawsuit, NTP’s $612 million payoff is going to attract a lot of imitators. Perhaps they won’t put a stop to software innovation, but by draining capital away from successful businesses, they will certainly reduce investment in new technology. That’s guaranteed to reduce innovation in the long run.