I strongly believe that the recent trend to patenting algorithms is of benefit only to a very small number of attorneys and inventors, while it is seriously harmful to the vast majority of people who want to do useful things with computers. When I think of the computer programs I require daily to get my own work done, I cannot help but realize that none of them would exist today if software patents had been prevalent in the 1960s and 1970s. Changing the rules now will have the effect of freezing progress at essentially its current level.

Donald Knuth, letter to the Commissioner of Patents & Trademarks, USPTO, February 1994

Inventions must be tied to a particular machine or transform something. Useful, concrete, and tangible result of State Street is inadequate.

David Kappos, Federal Circuit, in re Bilski, 30 October 2008

Patents aren’t bad. They’re an essential element of our economic system. The patent system and the 1952 shift to the protection of a process is the problem.

The 1952 Patent Act expanded coverage to include industrial processes. With the increasing importance of manufacturing to the economy, Congress had been successfully lobbied to provide a layer of protection around industrial manufacturing processes. With the Act, however, it introduced the framework by which to patent business processes – patents such as one click buy. The patents have muddied the water and introduced an expensive and chilling sense of uncertainty to business and information-oriented innovation.

Patents start with a basic tension. Economies benefit from the dispersion of ideas accompanied by sharing the details of inventions. Inventors, however, have little incentive to share the details if it results in no more than a roadmap for competitors to follow. Patents offer inventors a simple trade-off. Make public the details of your invention, and the government will in exchange grant the exclusive right to exploit it.

The 1952 Patent Act was designed to advance manufacturing quickly. Industrial processes, without the protection of a patent, might remain trade secrets, and these trade secrets were perhaps too valuable to the economy to be kept private. A patent system designed to protect them would open them up and accelerate growth. Or maybe it was more facile than that. Patents would provide legal protection and accrue enterprise to those that developed them. Either way, the patent system shifted from one organized around physical designs to one accommodating of processes.

A patentable process began as an industrial process. Though they may be anchored in a sequence of steps, these generally ensured some kind of physical transformation. The expanded coverage was designed to includes the innovative process of “floating glass on molten tin” to manufacture panes of glass, for example.

Patentable processes, however, would begin a migration from an area circumscribed from the physical to the immaterial. Patentable processes would eventually occupy the ground held by software processes, pricing systems, and general business processes. Notable among these include Amazon’s one-click-buy or the decision in State Street v. Signature Financial. The latter protected Signature Financial’s broad methods of pooling mutual fund assets into a partnership for tax purposes and economies of scale and closed the long-held business method exception. The Court decided, “a numerical calculation that produces a useful, concrete and tangible result, such as a price, is patent-eligible,” so Signature Financial’s methods must also be protected. Patentable processes went from a concept rooted in protecting the tangible innovations of the industrial economy to one oriented toward the more abstract innovations of an information economy, such as business methods and software patents. These could be called information processes.

The main architect behind the efflorescence of the patent system was Judge Giles Sutherland Rich. Judge Rich began his career at the United States Court of Customs and Patent Appeals (CCPA) and later moved to the United States Court of Appeals for the Federal Circuit (CAFC). The movie Patent Absurdity describes in Judge Rich a figure who held a singular, persistent and systematic influence on the expansion of the patent system. Judge Rich played a key role in the development of modern patent law through rulings that allowed for the patent of living things (Chakrabaty v. Diamond), software based inventions (Diamond v. Diehr), and business methods (State Street Bank v. Signature Financial Group).

The shift in orientation brought with it a new challenge. A patent of an industrial process produced a material result, so the boundaries of the patent should circumscribe the process by which one arrives at a specific physical product – a verifiable, physical transformation was required. Anyone could float glass on molten tin, but if they did so and made panes of glass, they would be in violation. Infringement could be narrowed to those who produce a similar transformation by the same process.

A patent of an information process, however, doesn’t yield the same transformation. It differs in two key respects. First, the information process often reflects a practice in the real world that has been codified for computers to track, analyze, and manage. Someone might realize, for example, that the ability of a known customer to go into a store, find a nice book for the ride home, and say, “put it on my account” on the way out the door, was valuable. They might introduce an online analog and patent the ability of a known customer to click from a good advertised on the screen directly through a transaction, without stopping to confirm payment or shipping details – one-click-buy. And they did. Amazon patented exactly that.

The claims are also often so broad that it is impossible to consider their limits without resorting to adjudication. Facebook’s most recent patent award provides a sprawling example of breadth in this area. Greg Sterling of Search Engine Land suggests that Facebook may have “patented social search.” The patent originated with Friendster, which was acquired in August 2010 for approximately $40m, but it promises to underpin much of Facebook’s search strategy. As one might expect, it insinuates a sense of novelty, but on reflection, it may only seem like a codified way of asking your friends for advice. The patent describes a system, comprising variously tasked servers, in the barest detail, that looks at popular results within a particular social network: clicks and social proximity determine relevance. OK, so if it’s popular with my friends it’s great. If it’s popular with my friends’ friends, it’s probably great. And if it’s popular in the successive networks beyond those immediate connections, it will register, but perhaps not weigh so heavily. Mind-blowing stuff, really. But Facebook’s award may have patented social search, and no one will know for sure until it is tested in court.

Facebook’s recent award dresses up common sense with all the verisimilitude of computer-system diction, replete with a shrill register of server types, hash algorithms and message authentication codes. Amazon’s one-click-buy cloaks the established practice of a charge account with the novelty of the internet. Of course my friends are valuable sources of recommendations. Of course it would make sense to bring the convenience of a charge account to the world online. But are these patentable?

No one in social media knows if Facebook patented social search, but everyone in the industry is exposed. Amazon received a patent for what appears to have been an established practice in the real world but a meaningful innovation on the internet. Imagine the chilling effect on an entire industry. Imagine the desperate sense of duty to patent any process carried over from the real world to the internet, if only for defensive purposes.

Patents are good. They encourage innovation by incenting inventors to share and invest in their inventions. The patent system, however, is the problem. The boundaries of what’s acceptable or ultimately awarded are so poorly defined that they don’t really exist until one reaches court. And it’s the cost of figuring out those boundaries, through both applications and litigation, that has the capacity to sap its strength and chill an industry.