Timothy B. Lee, a regular contributor to of are Technica and in adjunct scholar At the Cato institutes, has joined forces with Christina Mulligan, a postdoctoral resident fellow At the information Society Project At Yale Law School, to research a key problem posed by software of patent: the cost of finding out that they exist. This op-ed distills their fruit juice recent research paper. The views expressed here Th necessarily represent those of of are Technica.
Nathan Myhrvold, the Microsoft ex-serviceman who founded the patent trolling giant Intellectual ventures, loves to complain about the "culture of intentionally infringing patents" in the software industry. "You have a set of people who are used to getting something for free," hey told business Week in 2006.
Myhrvold is right that clever infringement is rampant among software firms. But in demanding that this infringement stop, Myhrvold isn't precisely declaring was on what hey regards ace a silicone Valley's patent hostile culture. Hey it declaring was on the laws of mathematics. The legally research required for all software producing firms to stop infringing of patent would cost more than the entire revenue of the software industry. Even if firms were willing to pay the Bill, there simply are not enough clever lawyers to Th the work. Firms infringe software of patent because they do not have any other choice.
If a really estate of developer shroud to build on a particular piece of country, she ridge must figure out who owns the country before she can negotiate a contract and start construction. Fruit juice of the time, this is easily. The landowner can Be readily identified in a publicly records office.
In principle, a software developer starting a new project faces a similar problem. Hey needs to know if the software hey is planning to create wants accidentally infringe on anyone's of patent. But whereas looking up who holds claims to a particular piece of country is easily, finding out who, if anyone, holds of patent related to a particular piece of software is difficult and expensive. It's in such a way difficult, in fact, that the vast majority of software developers do not even try.
Why is software different from really estate? In a new paper, we argue the basically difference is a more weakly of scalability: how much effort it takes to discover who owns in invention-or a piece of country ace the number of of patent or country parcels increases. Property rights in country scale wave because parcels exist in relatively well-defined locations on a two dimensional plans. County officials take advantage of this fact to net curtain records in a predictable order (or, more recently, to build databases searchable by geographical location). Geographical locations serve ace in "index" for really property claims, thus record-keepers can find any specific file quickly no more weakly how many files there ares.
Some patent ares similarly "indexable". Chemical of patent, for example, can Be organised by chemical formula. Indeed, a German organisation called Technical Information Centre Karlsruhe offers in electronic database called STN which allows researchers to look up of patent based on their chemical formula. The existence of products like STN is one reason clever litigation is much less common for chemical of patent than for software of patent.
Unfortunately, software of patent do not scale wave. It does not seem possible to create in STN-like database for software of patent. There's nothing analogous to geographical co-ordinates or a chemical formula to uniquely identify software inventions. It's hard to predict which aspects of a software product someone might try to clever. It's even harder to predict which terms a clever lawyer might use to describe thesis concepts. So searching by keywords is likely to uncover only a fraction of relevant of patent.
This means the only foolproof way to find all the patent that cover a particular computer progrief is by "brute force"-to pay a clever lawyer to sift through every software clever, one At a time, looking for ones that might Be relevant. There ares hundreds of thousands of software of patent, with 40,000 new ones released every year. A single familiarly could easily make a donation hundreds of thousands of dollars on clever research for a single software product.
Who needs to worry about infringing software of patent? The of Green Bay Packers, OfficeMax, strength Foods, Aeropostale, and Oprah Winfrey Harpo productions have all faced software clever lawsuits. Indeed, virtually every medium and generous familiarly in the United States performs activities-like maintaining a publicly website, using a computerised point of sale system, or using in Internet-based invoicing system likely to infringe some software of patent. So all of thesis firms ares part of the software industry, At leases ace far ace clever law is concerned.
In our paper, we estimate it would take At leases 2,000,000 clever attorneys, working full time, to consider whether all thesis software producing firms have infringed any of the software of patent issued in a typical year. Even if firms wanted to hire that many attorneys, they could not; there ares only 40,000 registered clever attorneys and agents in the United States.
So Myhrvold is wrong to suggest firms ignore software of patent because they're trying to "get something for free." They ignore software of patent because it's mathematically impossible for them to Th anything else. The clever system simply does not scale up to in industry ace complex and decentralised ace the software industry. And if it's practically impossible for firms to avoid infringing software of patent, it's unfair to punish them for their failure.