Saturday, June 09, 2007

Blocking Innovation

As Vonage and Verizon having been battling in court since early this year over Verizon's claims of having patented some faily obvious elements of VoIP used by Vonage, software patents have again fallen under scrutiny. Tim Berners Lee, the inventor of the World Wide Web, has an OpEd in the New York Times arguing against software patents. He quotes Bill Gates writing in 1991 that if early computer pioneers had understood how to patent their works "the industry would be at a standstill today."

Software has always seemed poorly suited to patents for a number of reasons. Most of what software does is fairly obvious. The really innovative stuff tends to fall into either broad conceptual ideas (like the web) that would be inappropriate for patents or clever algorithms (like quicksort) that are basically mathematical discoveries, abstract intellectual concepts of the sort not traditionally patentable. The rest of it, however complex and difficult, tends to be fairly mundane implementation, requiring little originality (but often a huge effort in coordination).

Existing patent law, in theory, should prevent unoriginal things from being patented. But in reality it does not seem to do a good job of this, particularly in the initial granting of patents. And even if questionable patents can still be challenged and overturned in court, this is a remedy of limited value to small developers and open source projects. Going up against Microsoft or IBM's legal team is often an insurmountable barrier to entry. And both patent examiners and judges appear to struggle with the concepts of computer science in their efforts to determine what is or is not an obvious development. As the OpEd points out, two of Verizon's patents in this case cover the painfully obvious concept of converting between phone numbers and internet addresses.

From the economic analysis side as well, software development is ill-suited to patents. For one, there is, as the OpEd discusses, alternative protection available for software developers in the form of copyrights. Any direct copying already gives rise to a legal claim. And copyright is, as the article notes, considerably less cumbersome and expensive. Also crucial, copyright does not create liability for independent invention, while patent does. So if Verizon has valid patents on basic elements of VoIP they can bar anyone else from offering a VoIP service until those patents expire. With a copyright they could only prevent people from actually copying their system, but they would have no claim against a competitor who developed their own, even if it worked the same way.

It is important to recall the purpose of intellectual property protections: to allow inventors and developers to profit from their investments in new technologies and innovations, thereby promoting more such investments. The need for strong protection is much lower in the software space than for, say, pharmaceuticals. Development costs are relatively low and innovation occurs rapidly enough in software that there is a significant first-mover advantage even without patent protections. As long as competitors cannot rip off a new program wholesale (which copyright prevents) and have to independently develop their own version, the first innovator will be rewarded for her efforts. Look at how YouTube still dominates over other video-sharing systems. Some of its competitors are offering technically superior systems, but YouTube got there first. The software market, before these big companies realized they could patent everything under the sun, was wildly innovative and fiercely competitive. They didn't need patents to achieve that. Patents impose signicant transactional costs and barriers to entry without adding any meaningful beneficial incentives to the software business. This system is a mess, and I hope the Verizon case will help to spark reform.

No comments: