Empirical Examination of Software Patents

Via IPCentral, I just finished reading “Patents and Business Models for Software Firms.” The authors assemble a large data set of patents, classify them as software and non-software, and do some statistical analysis as to which type of firms are most likely to take advantage of patents. They conclude, not surprisingly, that product-oriented firms are more likely to patent than service-oriented firms.

What they don’t do (and they acknowledge it) is determine any kind of causal connection among software patents, R&D spending, and innovation. And it seems to me it would be difficult to draw any conclusions about the impact of software patents on overall industry innovation using data of this sort. Software patents clearly benefit firms at the margin, or they wouldn’t seek them. But we can’t conclude from that fact that software patents benefit the industry overall–that would be a fallacy of composition.

It seems to me the best way of evaluating software patents empirically would be at the micro level: that is, look at individual patents and try to estimate the likelihood that the covered invention would have been created without the availability of software patents. Obviously, some will be hard cases, but there are also many easy cases.

It occurred to me that this is the sort of task that could be accomplished in a decentralized, peer produced manner: set up a web page where the user can look at a patent and rate it for obviousness, prior art, etc. There are probably enough geeks out there who hate software patents that you could analyze far more patents in far more detail than a traditional research team could hope to accomplish.

I just registered AmIObviousOrNot.com. I could set the site up, but my web development skills are rather rusty, so it would take me a while. Are there any PHP gurus out there who’d like to help out with a project like this?

SHARE: 8 comments