Slashdot Mirror
EFF Asks Supreme Court to Protect FOSS Innovation
euice writes "The Electronic Frontier Foundation supports KSR International in a fight against obvious patents. They filed an amicus brief (PDF) yesterday, a short summary is on their news page (August, 23). FTA: 'The Electronic Frontier Foundation (EFF) has asked the United States Supreme Court to overturn a dangerous patent law ruling that could pose a serious threat to Free and Open Source Software projects. [...] In a recent decision, the Federal Circuit Court of Appeals affirmed its own 'suggestion test' as the main method for determining when a patent should be found obvious over knowledge in the public domain. Under this test, even the most obvious incremental advances and add-ons can be patented unless the Patent Office or a defendant in court produces a document that shows someone else suggested it prior to the patent being filed. [...]
In its amicus brief filed Tuesday, EFF shows how this 'suggestion test' has led to a massive surge in bogus patenting, especially in software. These bad patents then become weapons against legitimate innovators — especially those working on Free and Open Source Software projects.' For me, this sounds like a really good shot in the right direction."
...since most of these shoddy patents get through because the patent examiner doesn't realize the applicant has just fancied up something obvious, another major improvement would be to require patent summaries to be easily readable. How to enforce? How about this: allow patents to be invalidated on the grounds of obfusctating terminology. To test this, a defendant could be allow to offer an alternative phrasing, and if the judge|jury finds that a) the alternative phrasing describes the same invention (i.e., the plaintiff can't think of something that would fall under one but not the other), and b) the alternative phrasing is "significantly easier" to understand, the patent is invalidated. That would have the added benefit of a kick of harsh reality to those who deceive themselves about their own inventiveness. "No dude, you just put a scroll wheel on the side. You didn't provide an 'integrated mind-user-machine interface', whatever that is."
Apology to Ubuntu forum.
How about this idea?
The idea behind patents, or so they say, is that every patent is supposed to be so unique that no one else who is "skilled in the art" would ever think of such a clever idea. In order to protect the brilliant guy who thought of this ground-breaking technique, we give him a 20 year monopoly on that idea.
Of course, unless there's a good chance that the guy really is 20 years ahead of his time, we shouldn't grant him the monopoly, because he'll be holding everyone back with his monopoly until it expires.
Right now, we're giving away too much. We give people patents for things that are too obvious, and we give them away for too long.
So, why not give a bit less, but still reward true innovation? How? Ensure that no one else files the same patent for five years, and only then issue the patent. The inventor still gets 15 years of monopoly, but now skilled experts in the same field have a chance to compete. If someone else thought of it, it wasn't original enough.
It also eliminates the "first to file" vs "first to invent" problem by a large margin: if two people skilled in the art were able to discover an patentable idea within a five year period, then no patent is granted. This is pretty much the statement of the uniqueness critera for patents in the first place.
It will also eliminate people from filing spurious patents: five years is a long time to wait for something that doesn't matter. Prior art and other challenges could still be filed, but at least duplicate patents won't get issued, and if they do, both patents get invalidated...
It's not perfect, but it would be a good start...
on my opensource site, atomic-ptr-plus, and put a list of patent applications by other companies for that stuff. I have to add in another one by IBM for some stuff I worked on, with them no less. I'm working in stealth mode from now on.
Indeed the problem of judging the obviuosness in hindsight is difficult. However, the Federal Circuit's solution is ridiculous. They have said that the only way to make sure that a patent was obvious at the time of invention is if there existed written public record at the time describing the idea. This creates a perverse situation: if an extension of an invention is so immediate that no-one would bother to point it out in writing, then it's legally a "non-obvious" idea and you may patent this extension. At the same time people who think of less obvious extensions (in the everyday sense of the word) would tend to write them down. Along the way, an important element of the patent system has been written out of the law by the court: the "Person Having Ordinary Skill In The Art" (PHOSITA). The test of patentability in the law is whether the invention would have been obvious to a PHOSITA [at the time of invention and not in hindsight, of course].
One way to show the idea would have been obvious to the hypothetical PHOSITA is to show written work suggesting this literal idea. The Federal Circuit says this is the only way they'll accept. In effect, they have eliminated PHOSITA's skill in the art -- his own expertise. The EFF is arguing that this is especially bad for F/OSS developers. F/OSS is a collaborative low-budget enterprise. Most projects can't afford a patent department, or filing for defensive patents. Even worse, F/OSS developers are people or ordinary skill in the art of software development. If something seems obvious to them in view of the prior art they know then they assume it can't be patentable -- that's what the law says, after all. Also, software people don't tend to document every silly extension to every idea they see. Then it turns out that their skill in the art is irrelevant -- what's important is whether the small extension they made was litterally written down by another developer.