I'm still scratching my head trying to understand what principle drives the distinction between an abstract idea and a non-abstract one. It seems to me that the very definition of a computer program is the description of an idea - in pedantic detail, yes, but it will always come back to an idea. If that is true, and if, as many say, it is only a matter of time before "software eats the world", is there anything left that will be patentable?
The court decision this week seems to entangle two things; the fact that the abstract idea was not novel, and the fact that it was abstract. It is not clear to me: had the abstract idea itself been truly novel - would that have changed the outcome? A lot of people are saying this ruling is against patenting abstract ideas but it seems to me that it is mostly about the fact that directly transcribing an idea to a computer doesn't ADD concreteness (or novelty?), NOT a judgement about the veracity of the original idea. Would love to hear any insight about this.
I am not a lawyer but the way I understand it is, in the old days you could get a patent for a machine that separates cotton fibers (a complete and specific design for creating something that does this), but not for "a machine that separates cotton fibers" (the very concept of such a thing, and all possible implementations of said concept, accompanied by a crude description of what such a thing might look like, given a sufficiently smart compiler).
What it sounds like to me is that the court is leaving open the possibility of software patents that are more like the former case, although the vast majority are more like the latter.
It almost seems then that written into the definition of non-abstractness is that the idea could be implemented multiple ways. In other words, If there was no possible other way to make a windscreen wiper then the abstract idea itself would be sufficient to specify the actual implementation (or the implementation can be derived as an inevitable logical consequence of the abstract idea). But if we can postulate a different way (perhaps less optimal) then we can say there is an abstractness about the original idea.
I still have trouble with this though because the point where you define the abstractness seems a bit arbitrary. There are many ways to make a wiper blade, a rotating mechanism, a motor, etc. So aren't those abstract too? If the patent specifies it exactly then it is truly concrete, but then it's too specific - competitor can just use a slightly different material.
And then at some point we invent a generically programmable army of nanobots that swarm to accomplish any physical task using a goal oriented AI type language ... and now merely specifying the idea IS the implementation for ANY task.
Well, that's the fundamental problem of patents. They restrict the free dissemination of ideas. In a free market, the idea of a patent is fundamentally flawed. Soft patents are the reductio ad absurdum of the whole concept.
You have it completely backward. Patents were introduced because it is in society's interest to disseminate ideas as widely as possible. This is why patented IP reverts to the public domain after a certain period.
Were it not for patents, inventors would resort to the only tool at their disposal to limit free implementation of their ideas: the trade secret.
What's interesting I think is how the web has proven this to be utterly demonstrably false - we see staggeringly valuable and complex webapps written entirely in trivially reverse-engineerable Javascript and HTML. I can't count the number of times I saw a cool technique and within minutes figured out how they did it using only tools built into just about every browser. Now it's true that given the option companies will absolutely obfuscate to the maximum extent possible (eg: for a while you would see stuff built in Flash purely to obfuscate it) - but the idea that people won't develop products at all merely because they can't protect their secrets just seems to be utterly false. In fact competition seems to thrive beyond all measure when revealing secrets is forced by the technology.
I think that's still not the point. A popular website's trade secret is not necessarily its source code. You can copy facebook's code from soup to nuts but you are not going to be nearly as successful.
On the other hand, imagine someone found out a way to cure cancer. If he just releases his formula, then there's a very good chance that he will not make any money off of it at all. Without patent protection, he will probably hoard the discovery for years while building a company of his own to sell the drug. Finally, the drugs will be available, and yet the public will still have no knowledge of how to develop the drug, improve it even further, or apply it to other diseases.
I'm not personally a big fan of patents. I'm merely addressing your comments, which I see to be somewhat irrelevant.
I don't see a problem here. If a patent is granted for the cure for cancer, then the products will most likely be very expensive because you have a monopoly on the cure. Poor people will not be able to buy the cure, so how has that helped them?
I think that's missing the point again. The point of a patent is so that the underlying technology is public. A patent system in fact does exactly what you suggest, which is to create a monopoly for a certain period of time so that the inventor can profit. But the existence of a patent implies that the formula for the cancer drug is out, and so other companies can use that to potentially create better versions or apply it to other areas.
Let me ask you, as a rational, profit-seeking individual or organization looking to secure revenue from your new invention in a patent-free society. Do you:
a. Volunteer the details of your invention to your competitors?
b. Keep these details secret?
Moreover, what would it take to get you to choose option a?
c. Appear to volunteer the details, but obfuscate them so deeply in patentese that nobody can ever tease them out, but I still get to sue my competitors.
How is that a strawman? He said that patents were introduced to allow ideas to be disseminated. That's a faulty premise.
However, to handle your own strawman, the answer is b. I'd keep the details secret. Which is not a problem, because there have been many, many occassions when someone reverse engineers the work or comes up with the idea completely independently.
My (non-expert) impression is that this ties back into the machine-or-transformation test [1], especially given how Diamond v. Diehr was cited as an example of a patent-eligible invention, notwithstanding its software elements.
The point of Alice Corp. would then be to remove the loophole of merely writing up an abstract process as a computer program running on a generic computer; the process would have to be linked to specific hardware characteristics to pass the test (obvious examples would be an MRI scanner, anti-lock brakes, or device drivers).
I think the distinction would be practical implementation. For instance:
Abstract: A way of making video files smaller by compressing it using various mathematical procedures.
Patentable: A way of making a video file smaller by braking it down to keyframes and than ......... {detailed description of one specific way of doing it}....
I think that's the line the supreme court is drawing.
It seems to me that in almost all cases, the "specialty" Courts tend to rule on the side of what they were meant to "judge". If it's a "patent Court", then they will almost always rule in favor of the patents. If it's a spying Court (FISC), then they will rule in favor of the spying, and become a rubberstamp Court for surveillance requests.
So maybe we should get rid of specialty Courts, no?
It is rather unfair to lump the "FISC" system in with all the courts, because it is more of a board, panel, committee, in camera hearing, or administrative proceeding. "FISC" lacks the essential qualities of a court, as the defendant has no right to appear, they are not represented by counsel, charges are not presented, evidence is not disclosed, and there is no right to trial by a jury of the accused's peers.[1]
Yes it's dreadfully unfair. Little children are crying. And yet GP's observation that "specialty" Courts tend to rule on the side of what they were meant to "judge" is true, without consideration for particular procedural differences.
Part of this is due to the fact the courts like CAFC are appeals courts.
By the time CAFC sees a patent, a lot of experts have already looked at it and said it is good. These include:
1. The patent attorney or agent who advised the inventor during the application process.
2. One or more examiners at the patent office.
3. When the patent owner was considering his lawsuit, he would have hired patent attorneys, non-attorney patent experts, and subject matter experts to evaluate the patent and the alleged infringing product, in order to determine if it was worth going through with a lawsuit.
It would be surprising if patents that get through this gauntlet did not tend to be approved by the CAFC.
Actually, the meme that CAFC tends to rule too much on the side of patent owners is more of an urban legend based on grumblings of some patent lawyers, which was then picked up and spread by certain "journalists" (hi, binarybits!) and regularly parroted by others, like TFA. But this is belied by the actual data -- see this older post of mine: https://news.ycombinator.com/item?id=7849386 It cites a number of empirical studies and statistics that shows that the Supreme Court and the District Courts agree with the Federal Circuit way more often than not.
Further, if you look at the EU, patent cases there AFAIK are handled by specialty courts, and they tend to result in much cheaper, faster and more consistent resolutions. This seems to be because they are decided entirely by the judges, who not only are already intimately knowledgeable about the laws they are dealing with, but most of whom seem to have technical backgrounds. This means no messing around with juries and expert witnesses. (More knowledgeable people, please correct me.)
Based on the evidence from the EU, I actually think the solution to the litigation problem is the opposite of what you propose: move the role of the CAFC much lower down the patent litigation chain, i.e. all patent cases go directly to the Federal Circuit or some equivalent thereof. Not only could this have the same benefits as the EU system, it would neutralize the whole game of forum-shopping.
> Further, if you look at the EU, patent cases there AFAIK are handled by specialty courts, and they tend to result in much cheaper, faster and more consistent resolutions.
The "cheaper and faster" aspect is more likely to be an artifact of civil law systems vs. adversarial systems (especially the absence of a jury in civil cases), since it is not specific to patent cases. And it's very much debatable whether they are more consistent.
Also, European specialty courts are just as prone to regulatory capture. See, e.g., the fascinating holding of the EPO Board in "Auction Method/Hitachi": "The Board is aware that its comparatively broad interpretation of the term 'invention' in Article 52(1) EPC will include activities which are so familiar that their technical character tends to be overlooked, such as the act of writing using pen and paper." A holding that flew in the face of decades of European patent jurisprudence and was a fairly blatant end run around the requirements of article 52 of the EPC.
I know the author didn't write the headline, but trolls are not "mortally wounded;" they're barely scratched. This ruling will have little effect. The Supreme court declined to offer any guidance about what "abstract" means, so courts will go on making it up according to the feelings of the presiding judge.
Legislation like the Innovation Act that could be brought back next year would be much more important.
Keep in mind: a lawyer tried to attack the crypto credibility of Whit Diffie!
Remember, that lawyer who attacked Diffie won the case! The jury decided that the Patent Office knew better than Whitfield Diffie about encryption and awarded millions to the guy who claimed to invent asymmetric transport cryptography in the 1990s.
Well, the jury decided on the facts that the lawyer presented, which were facts that Diffie could not deny. He was introduced by the defense as the "inventor of asymmetric key exchange". But the plaintiff simply pointed out that, since somebody at GCHQ had already invented it secretly, he was not the true inventor. Yet because the prior GCHQ invention was secret, it did not count as prior art, which was why the DH patent was still valid!
And here's the kicker: the defense was actually arguing that the plaintiff's invention had already been invented by DH and as such was invalid, and the plaintiff had already pointed out that DH's work was not published and hence could not invalidate their patent. Exactly like the GCHQ work did not invalidate DH's patent.
Given that the plaintiffs argument was "secret prior art does not invalidate patents", it probably wasn't the best idea for the defense to bring in living proof of one of the better known examples of the plaintiff's argument to support them.
Most of the worst software patents do exactly what the court said they couldn’t: present an idea (like hedging risk, one-click shopping, etc.) and say that, because it’s executed on a computer, or a generic piece of hardware, it is not just an idea anymore.
Does this mean Amazon will lose their one-click patent?
The court decision this week seems to entangle two things; the fact that the abstract idea was not novel, and the fact that it was abstract. It is not clear to me: had the abstract idea itself been truly novel - would that have changed the outcome? A lot of people are saying this ruling is against patenting abstract ideas but it seems to me that it is mostly about the fact that directly transcribing an idea to a computer doesn't ADD concreteness (or novelty?), NOT a judgement about the veracity of the original idea. Would love to hear any insight about this.