Oral arguments in the Bilski patent case were recently heard at the Supreme Court of the United States (SCOTUS). The line of questioning from the justices suggests to me that they have an understanding of computers and software that could be very problematic for those of us designing software and trying to protect that intellectual property with patents.
Their understanding seems to go something like this. 1) You can’t patent ideas in the abstract, they must be embodied in some concrete form. 2) Software programs are embodied by running them on a computer. 3) But, since the computer is already capable of running the program, simply saying that you run the program using a computer doesn’t make it patentable, because what you’re really patenting is the abstract idea expressed in your software. 4) Abstract ideas are not patentable. Ergo, software is not patentable.
I think a re-framing of the computer/software relationship could help explain why it seems obvious to programmers that their work can be novel, non-obvious and useful (the tests for patentability).
Consider the analogy: a computer is like steel. Many inventions are made of steel. That is, steel is the medium in which the idea (say a uniquely shaped tiller blade) is expressed. Steel is an expressive medium. Now, SCOTUS has no problem upholding patents for ideas expressed in steel if they are novel non-obvious and useful. That feels right. SCOTUS also has no problem upholding patents for new methods of making steel, for alloys, for ways of hardening steel, etc.. That is ways of improving the expressive medium. That also feels right.
The analogy I’d like to draw is that the computer is an expressive medium for software in the same way that steel is for tillers and many other useful things. So, many inventions are made using steel as the expressive medium, and I would argue that many inventions are made using a computer (and display and inputs etc) as the expressive medium. SCOTUS seems to be grappling with the idea that you don’t create a “special-purpose computer” to express a particular piece of software. That, to me, would be like saying that a new tiller design would have to be expressed in a new alloy specially made for the purpose in order for it to be patentable.
When thought of this way you realize that the very purpose of a computer, that is, a device designed to be programmed to perform new functions, is as an expressive medium for software. So, just like new alloys, new, faster, lighter, lower power computers are invented all the time. You’ve improved the expressive medium, and it feels right that these inventions are deemed patentable.
Computers are “ductile” like steel, and creating software is analogous to hammering the computer (the expressive medium) into a functional shape. You still need to ask if that shape is novel, non-obvious and useful just as you would when someone hammers a new tiller blade out of steel, but the use of a computer as the expressive medium does not make all software the expression of an abstract idea.
The problem I believe we’re trying to solve is that in some (many?) cases simply claiming that simple (obvious, not-novel) ideas become patentable simply because they are expressed on a “machine” known as a computer. These are the “trivial” software patents everyone decries. I’ll say two things about that. First, if these ideas really are “trivial” then they should fail the novel, and non-obvious requirements. No need to develop a new set of legal tests for that one. And second (and with only a trace of sarcasm), the definition of a “trivial” patent is the one that you’re infringing. (i.e. I never write a trivial patent myself, but every patent that I infringe is a “trivial” patent that should never have been granted in the first place.) Let’s not rewrite the patent law to handle cases that it is already equipped to deal with. If the patent is truly “trivial” it should fall under the most basic test of patentability, novelty and non-obviousness, and if it’s not, then you should negotiate a royalty and get on with your business (it’s what you expect the licensees of your patents to do).