The software patent mouse trap

At ZDNet I wrote a lot about software patents, to the point where they seemed to be a problem mainly for open source.

After the Roberts Court's refusal to make a decision in Bilski vs. Kappos, Google's competitors decided they could box in Android, a Linux distro Google's Open Handset Alliance had developed for mobile phones, by filing patent suits.

NOTE: This is the second in a series on open source subjects I'll be publishing here until I get a new paying gig around the subject.

What we've seen since is a bit like the opening of the First World War, everyone suing nearly everyone over absolutely everything. Florian Mueller has been following the action over at his FossPatents blog so you don't have to.

Right now it's a phony war, with all sides manuevering around various courts and claims. Shots will be fired if some court has to come to a decision, endorsing claims that Android is in violation of a patent and fixing damages. This would essentially make Linux in violation of patents.

But this is not really about open source at all. It's about the nature of innovation when machines are reduced to software. It's about the desireability of software patents, and the price we may be paying for them.

Consider the mouse. Xerox' Doug Engelbart won the first patent on a computer mouse in 1970. But he patented only one physical design for a computer mouse interface. 

There has been enormous innovation since then. We have had one-button mice, two-button mice, three-button mice. We have had trackball mice (mice turned upside down), mice for those with carpal tunnel syndrome from working with other mice, and a succession of laptop mice, culminating in a touchpad with buttons. Which led directly to touching the screen.

But what if Engelbart had been able to patent the whole idea of a mouse interface? What if his patent had covered, not a specific mouse design, but the whole idea of a point-and-click controlled by software? What would have happened to innovation then?

Well, you might say, couldn't they each licese the Engelbart patent? Sure, but the game defeats itself. The second mouse maker licenses Engelbart, what does the third guy do? And the fourth? Do they have to license all those patents? Lawyers might well say yes. And small inventors would be forced out of the market in that case.

I've called this the mousetrap test. Patenting a mousetrap forces you to show your design and lets people design around you. It protects the innovation but also the right of future innovators to innovate. Patenting the idea of killing mice has the opposite impact on innovation. It slows it down, and eventually prevents it, because you can't get around basic concepts.

Which brings us to the problem of software patents. Because, at their heart, software patents protect ideas, not implementations. That's what copyright is supposed to be for.

This story is copyrighted as soon as I write it. If you cut-and-paste the whole thing and republish it under your own name, or on another site without express permission you can prove, that's a copyright violation and I can sue you. Same thing with code.

But software patents don't protect code. They protect the idea behind the code. You can't write around them, as with copyright. For instance, you can write your own story on this issue, citing the same sources I did, even paraphrasing me, or quoting a bit of it with credit and/or a link. No problem. Many authors don't even give the credit, if their employers don't want it given to a competitor, although I try to get a link in myself when one inspires my work.

The mousetrap test, or something like it, has been around ever since software patents were first granted -- by a court, not by Congress. In its Bilski decision, in fact, the Roberts Court merely affirmed a lower court ruling. It is entirely court-made law, something "originalists" claim can't and shouldn't be done.

In this case, they're right. We've seen that since the original pro-software patent cases were decided. And, as Bitlaw notes, courts are now trying to pull back. But it's like diplomacy in the face of the Guns of August. Armies of lawyers are on the march.

Only Congress has the power to recall those lawyers, but that won't happen in a divided Congress unless the industries which are impacted agree on a solution. Getting such an agreement means going to all industries impacted -- especially the medical industry -- and hammering something out, then presenting it and getting it through both Houses. Hard to do, but not impossible if the initial business agreement is in place.

I have hope this can happen. The medical industry is now starting to feel the impact of software patents, as more-and-more devices become software-driven.

Meanwhile, billions of dollars will be wasted here defending the indefensible.