Slashdot Mirror
Supreme Court Weakens Patents
ajakk writes "The U.S. Supreme Court, in a unanimous opinion, overturned the decades old test for determine whether a patent is obvious. The Court ruled that the Court had looked at obviousness in a "narrow, rigid manner." This should allow patents to be more easily invalidated because they are obvious."
Now that one click is not patentable...
http://www.supremecourtus.gov/opinions/06pdf/04-13 50.pdf
No. Patents can be re-examined at any time with this ruling in mind. This will apply to every flimsy patent issued because of a bad CAFC ruling made years ago. This is the first time SCOTUS has weighed in on this topic since the Graham v. Deere case that established the rules for Obviousness. By calling for more re-examinations (poor patent office might get overrun), these old patents can and should be overturned. The screaming you hear is the big pharma who are going to lose their butts on this. The people who are happy, well believe it or not, they are the software folks (and patent examiners, who will like being able to reject patents without nearly as much effort as before).
Also, people can use this case as precedent to have patents that they are being sued with overturned, showing that they do not have the strength to overcome obviousness over prior art. (Basically, SCOTUS just re-defined obviousness in prior art. Now that prior art has changed, patents can be overturned on prior art they previously were not able to be.)
At the end of the full ruling is this little chestnut:
We build and create by bringing to the tangible and palpable reality around us new works based on instinct, simple logic, ordinary inferences, extraordinary ideas, and sometimes even genius. These advances, once part of our shared knowledge, define a new threshold from which innovation starts once more. And as progress beginning from higher levels of achievement is expected in the normal course, the results of ordinary innovation are not the subject of exclusive rights under the patent laws. Were it otherwise patents might stifle, rather than promote, the progress of useful arts. See U. S. Const., Art. I, 8, cl. 8.
I'm of the opinion that software patents are not necessarily horribly bad or wrong, at least not moreso than any other kind of patent, but it's just that the way they have been implemented currently is so far from ideal that we'd be better off eliminating patent protection from software entirely than sticking with it.
What has traditionally been patentable are particular methods of solving problems. E.g., the sewing machine we're familiar with today (with two interlocking threads, one in a bobbin, etc.) is one way of solving the "how do we attach two pieces of material together" problem. It's (or rather, was) a novel solution to the problem, it was non-obvious, and it was particular. That's an example of a pretty good, justifiable patent. (Also because it's not easy to protect by other means -- once you see a sewing machine and take one apart, you realize immediately how it works and it's trivial to re-implement it, but if you hadn't ever seen one it's not obvious that two running threads is the way to do it, hence why it took so long to be invented.)
I'm not sure that there is a good argument for preventing people from patenting the solutions to problems, where the form of the solution happens to be microcode, in the same way that the form of the solution to the sewing-machine problem was milled pieces of steel.
But the problem arises when judges and patent examiners aren't skilled and selective about what's patentable. It's much easier, with software-based inventions, to get overbroad patents that negatively impact invention; rather than patenting a particular solution, what gets patented are entire classes of mathematical functions, or all possible software implementations (solutions) of a given problem. That would be like getting a patent, not on a particular sewing machine design, but on all sewing machines generally, or even "any machine for attaching two or more pieces of fabric together."
The problem, in my opinion, with software patents isn't with the fact that they're software -- in my mind, software ought to be patented, and it ought not be protected under Copyright (unless we're willing to define it completely as "speech" with all the freedoms that entails) -- but that they're typically of very poor quality, shoddily researched, and overbroad.
For this reason, I think the Europeans have done a good thing in just avoiding the issue entirely, because the cost of overbroad patents on innovation is far worse than no patents of a particular type at all. (I think this is trivially obvious but there are a lot of historical examples where overbroad patents have been problematic and basically stymied development that was otherwise ongoing -- the old internal-combustion patents are a prime example.)
We have the legal framework to deal with software, but unfortunately we just haven't used it correctly, and until we're willing to do it correctly -- and that means we're going to need to apply a lot more resources to the task of ensuring that patents are novel, non-obvious, narrow in scope, and deserving of protection -- they're a lot more trouble than they're worth.
"Ladies and gentlemen, my killbot features Lotus Notes and a machine gun. It is the finest available."
Software is typed up stuff, written in a language or languages, and as such, is more akin to written books or articles or say like musical scores, and should only be allowed copyright, not patents. Patents should be restricted to tangible products. In addition, the software industry itself has insisted and got granted immunity from normal consumer warranties, which is clearly evidence they don't see their own typed up stuff as a "normal product". It's *special*.
They shouldn't have it both ways when no other industry can claim that. If it is patentable, it should come with a minimum implied normal warranty (suitable for use, no glaring and or dangerous defects, etc). No warranty should mean no patent, copyright only.
I hope that is linear enough to answer your question.
I am a patent litigator (meaning I mostly kill patents; I don't create them), and this of course is big news in my practice today, and will be very helpful in several on-going cases.
Here's my take: the case allows lawyers to tell a story about the prior art that makes sense. Previously, the Federal Circuit had shut you down if you couldn't point out explicit prior art for every little detail of the patent that you wanted to invalidate. That's exactly what they did in the case under review. The defendant pointed out that all of the basic problems had been solved in other patents, but the Federal Circuit responded that they hadn't been solved with the intent of solving the particular problem the patent said it was trying to solve. Well, so what? We should be able to assume (and argue) that ordinary engineers have a little common sense and creativity in determining how to use previous inventions. We shouldn't have to show the courts that there was an exact road-map for an idiot to follow and arrive at the precise "invention" at issue. That's the big help in this case.
So from my point of view, here are the two big advances from KSR today:
This should be obvious, but it will help that the Supreme Court said it. It will be quoted a lot because it shows that we can assume that the ordinary engineer can make simple inferences and doesn't need his hand held.
This is the Supreme Court's long-winded way of boosting the "obvious to try" argument. The Federal Circuit has for a long time rejected the argument that it would have been "obvious to try," instead saying that it needs to be "obvious to do". (For example, it may be obvious to try to build a time machine, but that doesn't mean the invention of a time machine would be obvious.) But I read this opinion as saying that "obvious to try" goes a long way towards showing that it was "obvious to do". In other words, if it was obvious to try and the trying used predictable methods and yielded a predictable result, then the whole thing was likely obvious. So this will help as well.
Finally, it is going to be interesting to see how the PTO itself deals with this opinion. If an examiner comes back to you and says, "no I think this is obvious," it's very difficult to "prove" otherwise, especially early in a product's life cycle when you don't have market data to show how successful it might be, etc. At least under the previous test, the examiner would have to point out all the explicit references in the prior art. Now they apparently just have to point out the basic elements, and then say, "in their opinion," a person of ordinary skill and creativity would have been able use these elements to make the invention. How do you argue with that? "No, I don't think ordinary people are that creative"?
So, you fix one problem and possibly create another. The opinion shifts the debate to help prove obviousness, but it doesn't solve the underlying problem of what obviousness means. So it goes.
For example, take a simple algorithm which decides the next best move in a chess game... do you really think that the standard mental process is to recursively examine the next several moves and potential counter moves for a given board layout. Similarly, is the typical mental process for using language based off of some well-defined CFG?
Algorithms are designed according to the availability of several functional units not offered by the human brain. Software based off of mathematical algorithms exploit the availability of an ALU and/or FPU. Hardware might even be developed to accelerate frequent, specific tasks required by software (such as designing an accelerator to perform matrix multiplications) in which case the software is optimized according to the availability of such hardware.
The truth is that, despite some superficial similarities, both the underlying design and functionality of a brain and computer are quite different and serve different purposes. To argue that there is no difference in the way each is used (which is what you are saying when you claim that software really isn't any different from thought) is to ignore the very structure of a computer.
"Is not a sentence" is not a sentence. Well damn.