As the courts increasingly flirt with excluding software from patent protection, a common argument from software patent supporters is that wholesale abolition of software patents is a crude way to deal with the system's problems. The legal scholar John Duffy is the latest to take this line, decrying abolition as a "brute-force 'reform'" that has proven to be "profoundly shortsighted."
But the reality is that everyone thinks certain kinds of innovation should be excluded from patent protection. The only disagreement is whether software should be on the list. For example, though you can copyright a specific movie or a song recording, you can't patent the general concept of the buddy comedy or the verse-chorus-verse pop song structure. And hardly anyone wants to change that.
Notice also that patent law's ban on music and movie patents applies even if a patent's claims are framed as a machine that embodies a particular artistic innovation. You can't patent a CD with a verse-chorus-verse pop song on it. Nor can you patent the process of viewing a buddy-comedy DVD. The patent office and the courts would recognize this kind of patent application as sophistry. The "invention" in these cases isn't a new machine or physical process, it's a creative work that happens to be stored in an electronic format.
And the law's ban on movie and music patents applies no matter how novel and non-obvious the claimed invention is. The problem with music and movie patents isn't that they would be "low quality," it's that music and movies do not represent the kind of innovation the patent laws are designed to protect.
To many computer programmers, software patents look a lot like movie or music patents. A computer program is a sequence of abstract mathematical operations. The Supreme Court has long said that by themselves such mental steps are not patentable. And just as musical innovations didn't become patentable once musicians started recording music electronically, so software patent opponents don't think sequences of mathematical steps shouldn't become patentable just because a computer happens to be doing the calculations.
Of course, others disagree with this way of looking at it. Patent attorneys have had a lot of luck re-casting software patents as patents on machines that happen to run a particular type of computer program.
But I think the analogy to music patents demonstrates why that's a bad idea. A big problem with music patents is that they'd be hugely burdensome to musicians. Creating music is an individual, expressive activity carried out by millions of people around the world. These people shouldn't have to hire a patent lawyer before they start composing a song; such a requirement would stifle, rather than promote, musical innovation.
Moreover, patent protection isn't needed for music because musicians are already protected by copyright law. It's much easier to get copyright protection than patent protection, and musicians don't have to worry about accidentally infringing someone else's music copyrights.
The same points apply to software. Computer programming is a creative endeavor practiced by a millions of people. Many of them don't have the time or resources to hire a patent attorney and perform a patent search. The result is a lot of frivolous litigation against people who accidentally infringed someone else's software patent.
Meanwhile, computer software is eligible for copyright protection, making patents unnecessary to reward software innovation.
If the courts had mistakenly allowed music patents, leading to a rash of patent lawsuits against musicians, the solution would not be to raise the quality of music patents, allowing only the most significant musical innovations to be patented. Rather, we would say that allowing music patents was a mistake, because that's not the kind of innovation that should be eligible for patent protection.
By the same token, the only way to preserve the freedom of programmers is to provide them with the same kind of categorical exclusion from the patent system that musicians and movie directors enjoy. Even if we dramatically improve the quality of software patents, it will remain the case that the costs of forcing every programmer to run a patent search before beginning work on a computer program dramatically exceeds the benefits from allowing patents on software.