Software Patents as Innovation Tax, Rather Than Catalyst for Innovation

For anyone interested, I expounded at some length a while back about why method patents are well-entrenched in U.S. law and about the philosophical debate surrounding them over the years (see http://news.ycombinator.com/item?id=1171821). Software patents, of course, fit within that mold and, in light of Bilski, will not be subject to any form of categorical challenge for years to come in the courts. The issue lies with Congress, and this piece sets forth an excellent way of making the right kind of policy challenge (in contrast to the more superficial "this is ridiculous" kind of point which, though sometimes true, the patent bar is usually quite good at rebutting, at least based on existing legal standards - see, e.g., http://hallingblog.com/2010/05/26/patent-ignorance/).

In the analog era of patents, the U.S. courts went nearly two centuries in which the idea of a patent was strongly tied to that of a "machine" and where they often held that, while an original invention was in itself patentable, an improvement upon that invention was not. For example, in 1875, the Supreme Court held that "it is no new invention to use an old machine for a new purpose" (I cite and discuss this in the linked item above). It was not until 1952 that Congress overruled this line of cases in amending the patent statutes, setting the stage for what became an eventual flood of software patents as the digital age began.

The policy issue making software patents problematic in our digital age is that patentable software innovations, though supportable under current legal standards, basically are creating an ever-complex web of closely-interrelated and incremental improvements in any given area (e.g., mobile devices) such that no one can develop anything there without getting entangled in the web. It is not just tricky. It is impossible. The only way to defend is to make your own web (your "defensive patent portfolio" or your "defensive pool") and use that to catch others before or as they catch you. If you don't have that, you are in trouble.

Is this, then, what was intended by the Progress Clause of the Constitution (Article I, sec. 8, cl. 8) when it provided that the Congress shall have the power "to promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries"? Just how do these tangled webs of 1-and-0 based incremental improvements that are a normal part of all software development in the first place manage to promote the progress of science and useful arts?

The cause for reform in this area will be won or lost based on how the issues are framed and here we have a good example of how to begin to mount an effective challenge.

The amazing thing to me is the extent to which companies can't solve elementary coordination problems, and so stay in a Hobbesian state where entities just randomly attack other entities with patent-clubs and there's nothing resembling law or police entities. Step one would be banding together, not just with a centralized defensive patent portfolio, but with a centralized legal defense fund. As for the NPEs, patent the business method of patent trolling (I think somebody already did this, so just keep applying until you get another patent for the same thing), then sue them for violating it. After they're convicted by the central court of patent trolling, of course.

Astounding how the insight of law and centralized police, in just a slightly different domain, is so counterintuitive that companies still farm the lands as individuals, prey to whatever bandits stalk their way.

John Walker's PATO (described in the post) seems like it might be a good idea.

Here, just to brainstorm a little, is another idea I was kicking around the other day. I'll call it the Rational Software Patent Alliance. The basic idea is that any patent dispute between two RSPA members is settled by binding arbitration performed by the RSPA. Unlike the current system which foists technical decisions of obviousness and infringement on judges and juries who are not competent to make these decisions, the RSPA arbitration team consists entirely of experienced (and very well paid) developers and computer scientists. The arb team will form its own opinion of the validity of a patent, very much including its judgment of the obviousness of same; and everyone should know that only a minority, possibly a small minority, of software patents issued by the PTO are likely to pass the much higher bar of the RSPA.

Of course, there's some risk that a patent holder, unhappy with their patent being considered invalid by the RSPA, would quit the organization so they could sue another RSPA member in the courts. To deter this behavior, we say that an entity can join the RSPA only once; you can't get back in after leaving, and what's more, if someone acquires you after you leave, they can't join the RSPA either, and if they're already a member, they're kicked out.

The incentive to join RSPA is simple: it removes some fraction of the risk of being sued -- said fraction being a function of how many other companies join -- and it wouldn't be expensive unless you actually avail yourself of the arbitration process, and even that would be far cheaper than litigation.

I think this idea is actually complementary to PATO, as it's an idea about how members should resolve disputes among themselves, where PATO is more about how members should help one another in disputes with nonmembers. Walker suggests that PATO members should never be able to assert patent rights against other members, but I think that's going to be a harder sell than what I'm proposing, where companies do have some chance of protecting their real innovations from other members. I think there has to be some chance of that if we're going to get the big players -- Microsoft, IBM, Oracle, etc. -- to join.

So I think there should be one entity -- and "PATO" is definitely a better name than "RSPA" :-) -- that has both of these functions, to help defend members against nonmembers, and to arbitrate disputes between members.

Alas, neither of these functions does a good job of addressing the NPE problem, as NPEs would obviously never join such an alliance, and they aren't vulnerable to countersuits precisely because they're non-practicing. I think if anyone is going to fix the NPE problem, it's going to have to be Congress.

Patent insurance is the knee jerk reaction to patents, and more or less impractical.

However, the patent pool concept this article discusses sounds like a way to more or less provide patent insurance, which would be an amazing thing if it reached critical mass.

However you interpret the current situation, the patent and copyright clause is certainly not operating as intended by its authors.