Software Patents... still crap

Today I read on Dairy of an x264 Developer that Tandberg had patented, or rather applied for a patent on, some algorithmic techniques that Jason had used in his work on x264 and ffmpeg. While interesting, to me at least, the real point of this is to point out that it is really the flawed patent system that leads to this sort of application being made.

Not even entering into the idea that software patents are a bad idea, there is plenty in the media about that already, this illustrates some inherent flaws in the system. While obviously Tandberg in this instance have acted in a very bad manner, they have done so because the patent system not only allows such actions but even goes some way to encourage it. It is only because the developer working on this area, had the patent application brought to his attention that it was noticed. 

The problem, I feel, really lies in both the prior art and obviousness determinations made by the USPTO and the cost of challenging them. While I do believe that this patent will be stopped, similar patents copied from other peoples work have been granted in the past and have gotten through the approval process. The fact that the infringing patents have not been picked up is related to how prior art and obviousness are assessed. Ideas that are completely obvious, to the point of not being deemed worthy of patenting, to someone in a particular field, may not be at all obvious to someone outside that particular field.

How is it then that the USPTO is supposed to ensure that obviousness is tested thoroughly? Well, in all honesty they can't possibly ensure that some don't fall through the cracks. This is where the second and more insidious problem comes into the fore. A big company with a scary team of lawyers is basically always going to win against the little. If Joe Bloggs from Farmville Tennessee comes up with an idea that is stolen by a large company his small amount of funds to fight a patent battle will be swallowed in briefs and other such things from the large company. To the point where he can no longer challenge the validity of the patent for fear of not being able to feed his family.

The excessive cost of challenging patents has led to a situation where large companies, like Tandberg, feel comfortable attempting to patent things like this. The risk reward payoff is worth it for them. They are going to get a slap on the wrist at worst and at best they have a patent. This patent will then be used primarily as a negotiation tool with other big companies that think Tandberg has stolen some of their technology.

Basically, they come up and say, 'hey you're using this which we patented', then Tandberg comes back and says 'well you're using this which we've patented' and then they both agree to license the other to use their ideas and they live happily ever after.

Companies are willing to patent obvious things (touch screen computer anyone?) so they have a drawful of patents to wave at their competitors should the need arise and when they get desperate to start suing people in a vague attempt at perpetuating the companies existence. However, the big patent wars never really get anywhere in the long run and just seem to hinder the growth of new and exciting technologies.

So I'd written this whole ramble and then stumbled across a thread on Slashdot which was started by another coder that had the same thing happen to him. There is some pretty good discussion in there as well.

Patents are Screwed! US Patent App: 0080270152

Ok, so there are plenty of articles floating around in the ether about the uselessness of the patent system in the US. I'd like to point out a couple that really bring it home for me.

The first is United States Patent Application: 0080270152 a patent on how to patent something that a competitor needs for a competing product and then use this to obtain a cross licensing agreement. Not only is it dodgy business practice at best but it hardly seems patentable.

Then there is the admission from James "Father of Java" Gosling that the engineers at Sun pretty had a competition to get the goofiest patent, here's a quote from his blog (via Zdnet).

There was even an unofficial competition to see who could get the goofiest patent through the system. My entry wasn't nearly the goofiest.

And I have to admit that we had similar competitions at my old place of employ. We tried to get a lot crazy patents idea through, and we also spent a lot of time discussing the silliness of the entire process.

And finally there are the continuous stream of stories of software and other patents coming out of Apple such as this one from Go Rumours. It's not that I think Apple is doing the wrong thing, in fact it would be stupid of them not to be out there patenting everything that they can think of.

The problem is that too many patents that should be chucked out get approved, too many patents that have prior art, are so obvious that everyone else with less patent lawyers sitting around have skipped past and too many patents that are on things that should be covered by copyright and not patents. Copyright is another legal construct that needs a serious overhaul.

So, what should we do? If you've been paying attention recently you would have seen a whole lot of stories about a P=NP proof, that was released. The merits of the proof seem to have been beaten down now but the process by how this occurred shows a lot of promise. It was peer reviewed by the most savage group of peers in existence, every mathematician with an internet connection. This led to the paper being quickly reviewed in a way much better than the old peer review via journal process.

How about for patents we get a peer review system? Obviously privacy needs to be considered so that ideas are stolen. Perhaps there is a pool of respected representatives in a given field. They each have the ability to review the patent application, subject to NDAs of course, with a system in place that allows them to be selected or opt in and out of reviewing a particular patent based on collection of categories, products etc that their companies work in. There are some obvious kinks to be worked out, how does one get the patent application reviewed by peers in the field without those peers getting influenced by ideas etc that they have seen in the patent. How about we start with retired people in the field, or people that have moved on to different areas of expertise.

Obviously this is all a work in progress, the ideas need to be mulled over by people smarter than me. In the mean time I am going to see if I can patent a patent vetting system so I can make some cash of it when the Patent Office starts using it!