The Supreme Court doesn't understand software
vox.com
vox.com
> In Thursday's ruling, the court rejected a patent that claimed the concept of using a computer to hedge against "counterparty risk" — the risk of making a bargain and then having the other guy not pay up. The Supreme Court complained that in the process described in the patent, "each step does no more than require a generic computer to perform generic computer functions." Such a generic patent, the court said, isn't eligible for patent protection. The problem is that this criticism can be leveled at literally any software patent. At root, software is nothing more than a sequence of mathematical operations.
The Supreme Court in Alice came to two conclusions:
1) The concept of intermediated settlement is a "fundamental economic practice" and is so ineligible for patent protection under the judicially-created exception Section 101 for "abstract ideas."
2) The "method" and "system" claims in Alice Corp.'s patent amounted to nothing more than describing how to implement intermediated settlement on a generic computer in a generic way, and did not render the otherwise unpatentable abstract idea patentable.
The purpose of (2) is to foreclose on clever drafting that tries to get around the "abstract idea" limitation by framing it in terms of a specific implementation on a computer. What Alice says is that implementing the idea in a generic computer doesn't turn an abstract idea into a concrete implementation. It does not say that an otherwise patentable idea is rendered unpatentable because the patent describes how to implement the idea in a generic computer.
* checks *
Yep, Timothy B Lee. He's the Rush Limbaugh of Hacker News, telling us exactly what we want to hear, and the incredibly bad things that are possible, where "possible" means "doesn't defy the laws of physics."
"The change represents a fundamental shift in power in the Internet economy that threatens to undermine the competitive market structure that have served Internet users so well for the past two decades"
http://www.washingtonpost.com/blogs/the-switch/wp/2013/12/06...
"But then, in the 1990s, a patent-friendly appeals court handed down a series of decisions that opened the door to patents on software. That triggered a wave of patenting that has drowned the technology industry in litigation."
http://www.forbes.com/sites/timothylee/2013/01/17/aaron-swar...
"If Ortiz thought Swartz only deserved to spend 6 months in jail, why did she charge him with crimes carrying a maximum penalty of 50 years? It’s a common way of gaining leverage during plea bargaining. Had Swartz chosen to plead not guilty, the offer of six months in jail would have evaporated. Upon conviction, prosecutors likely would have sought the maximum penalty available under the law. And while the judge would have been unlikely to sentence him to the full 50 years, it’s not hard to imagine him being sentenced to 10 years."
I get that all these things resonate deeply with hackers. But Rush Limbaugh's broadcasts resonate deeply with his audience, too.
I don't see how you should be casting stones considering the rhetorical device you're employing right here.
He's super libertarian, for sure, and he writes about technology policy, so what did you expect? There is still legitimate content here, these are not techcrunch articles, and dismissing it like this is just weak.
It's not merely "unlikely" that Swartz was going to get 50 years. It was impossible. He could have plead not guilty, and then taken the stand, answered every question with "LAWRENCE LESSIG SAID I AM SPECIAL, SO I AM ALLOWED TO GO WHEREVER I WANT" and he still would not have gotten even half that penalty. "It's not hard to imagine Swartz getting 10 years" says more about the speaker's imagination than any real reading of sentencing guidelines.
It's like an average-quality HN commenter got a journalism job and didn't bother actually learning things that disagree with his worldview. Saying "Swartz was facing 50 years!" sure gets the blood flowing, doesn't it?
Apparently our industry is "drowning" in legislation. Given the tremendous rise in power and salary for all players, I bet other industries wish they were "drowning" the same way we are.
If you look at CLS Bank v. Alice, the Court concludes that intermediated settlement is an abstract idea because it is "a fundamental economic principle." So to use your example, data compression (replacing frequently-repeated sequences with a shorter representation) might be a "fundamental computer science principle." But Lempel-Ziv-Welch, a specific compression algorithm, wouldn't be.
Basic Lossless compression can function like this:
Imagine a string of 1's and 0's e.g. 10010000011000101111001
This string can be trivially compressed in a losses manor using this algo, every time the bit changes to a one or zeor, record the previous run of bits. So we would compress this string to look like: [1,1][0,2][1,1][0,5][1,2][0,3][1,1][0,1][1,4][0,2][1,1]
I understand now this is a terrible example however it is good enough here.
So since most/all compression algos are just using a pre-defined set of choices on how to compress data in either a lossless or lossy manor. The only difference between two compression algos would be their rules for what data to keep and how to arrange it more efficiently into a different data structure.
I think the analogy of board games can be used here. While you are free to get a trade mark on many aspects of your game, you can not patent the actual rules or game play. http://www.copyright.gov/fls/fl108.html
The rules and game play is what makes Risk different from mouse-trap.
Replace rules with compression algo (or any software algo...) and we come to the conclusion of software is not patentable.
So at what point does a collection of fundamental computer science concepts become patentable?
The line is completely arbitrary using the compressions algo example. I honestly have not been able to reason through a real life example that hold up to this scrutiny.
Maybe I am misinterpreting your answer, so if I am, I apologize, just ignore me =D
EDIT
Thanks to person for the discussion free down vote. Why engage when you can suppress.
There's a good argument to be made that the cost of the line drawing exceeds the benefits. I don't think it does, generally, but maybe it does for software. That said, I think you should make some money if you invent LZW. I don't like the idea of an economy where you can't make money off R&D unless you package it into a product with lots of advertising and sales people. I don't think that creates the best incentives.
I think the second example provides a great illustration of the divide between software patent proponents and detractors. Some people think a thing like LZW should be patentable, because they imagine inventing something on that scale of ingenuity and want to be able to make money off of it. Other people, in the scope of a larger project, usually, come up with things on the scale of ingenuity of LZW compression and are exasperated to discover that someone else patented it a few years prior and wants prohibitively large licensing fees, rendering the technology unusable; they don't think such things should be patentable because from their perspective it reduces innovation.
I've often heard the claim that game rules are uncopyrightable but that they might possibly be patentable.
Despite your assurances, I am uncomfortable with this decision. Indeed, I am uncomfortable with Gottschalk v. Benson. I don't think a bright line exists between patent-eligible software and ineligible algorithms -- in this I agree with the Vox article. I would much rather have seen a decision that invalidated this patent on the grounds that taking an existing manual process and computerizing it is, by itself, obvious.
The purpose of patents had nothing to do with compensation. It was already presumed that an inventor could sell an invention. The real purpose was to promote disclosure of how to make things. In other words to reveal what would otherwise be kept a trade secret. This fits well with the notion of "non-obvious to a person practiced in the art", which would preclude Amazons one-click. The idea was: Tell us how it's made in exchange for a monopoly for some number of years. Spreading knowledge is how you promote innovation.
http://en.wikipedia.org/wiki/History_of_patent_law
By the 16th century, the English Crown would habitually grant letters patent for monopolies to favoured persons (or people who were prepared to pay for them).[13] This power was used to raise money for the Crown, and was widely abused, as the Crown granted patents in respect of all sorts of common goods (salt, for example). Consequently, the Court began to limit the circumstances in which they could be granted. After public outcry, James I of England was forced to revoke all existing monopolies and declare that they were only to be used for "projects of new invention". This was incorporated into the Statute of Monopolies in which Parliament restricted the Crown's power explicitly so that the King could only issue letters patent to the inventors or introducers of original inventions for a fixed number of years. It also voided all existing monopolies and dispensations with [some] exception[s]....
There has also been some historical research that suggests that the "Statute of Monopolies" was a political compromise[1]. There wasn't really a lot of thought about maximizing innovation or incentives, or whatever. However, I fully agree that the popular understanding of the patent system has more to do with incentives for innovation and disclosure. But any engineer will tell you there's far more optimal ways to explain technical devices and inventions than legalese and claims.
On a related note, the USSR had a patent system. With patents. Who would have thought? This stuff is super ingrained.
[1] http://diyhpl.us/~bryan/papers2/Generally%20inconvenient:%20...
Yeah, I agree. The entirety is:
To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries;
That's the whole description in the constitution. At the time (1780s-1790s), other countries had patent systems, including the Empire from which the authors drew inspiration for at least other aspects of the constitution. A patent system wasn't a new idea. For that matter, the 1790 implementation of the patent system also looked very similar to the others of the time. My guess is they were thinking "hey, let's do that too".
To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.
It's clear from the historical record that this was about ensuring monetary compensation for invention, as knock-off devices were just as common then as they are now.
In fact, public access to patents was basically non-existent until the Patent Act of 1836. The "non-obvious" test came even later.
The purpose is "To promote the Progress of Science and useful Arts". This is achieved by "securing for limited Times..the exclusive Right".
Monopoly and financial compensation are not the aim of constitutional patent law, merely the means to an end.
Knock-off devices were common then as they are now, and in the absence of patent law, the best protection for so-called intellectual property is to just keep it secret. This, of course, has a negative impact on society, as research effort gets duplicated, and the state of the art advances more slowly than it otherwise could.
So, the balance is - share with the world the details of making your invention (i.e., progress the science or useful art), and in exchange, we'll guarantee that your openness doesn't bite you by aiding your competitors (we'll give you a time-limited exclusive right).
However, it's that end that's being examined here, not the means. My point was that by granting a limited-time monopoly on an invention, Congress can advance the state of "Science and useful Arts". It can be argued what is meant by "promote". As I noted, the historical context makes it pretty clear that it's meant as an incentive to invent, not necessarily public documentation of inventions, which really came later. This is exactly what is said in the article: "The idea behind patents is to create a financial incentive to promote innovation: patents ensure that the inventor of a new idea can get compensated for it, rather than seeing her idea immediately ripped off by copycats".
Regardless, what absolutely is missing from that clause in the constitution is any requirement at all that there be a public disclosure of the invention. That's what I was saying the constitution was clear on. Any assertion of public disclosure being the purpose is an interpolation, because the constitution has absolutely no requirement that that take place. All patents could be secretly filed away, for all it cares.
The US constitution does not contain the idea behind patents, it contains a view of how people wanted to encapsulate existing ideas into a legal document for a new nation.
Yeah, I find it very interesting how the US constitution is often assumed the origin of patents, but the same people wouldn't assume it's the origin of other legal concepts, like a judicial system.
What historical evidence is there that the "real purpose was to promote disclosure of how to make things"?
I agree with the general idea of this piece, and I don't believe that software be patentable (or at least, software should be protected in a very narrow range of cases), but this argument doesn't hold up for me. For a person implementing a complex payroll system or a complicated user interface, the universal underlying mathematics are meaningless. Of all the time spent developing software, algorithm development is a tiny portion, much of which occurs in non-commercial environments anyway. It's comparable to the difference between the laws of physics and mechanical devices that operate as a consequence of those laws.
The Knuth quotation in the article, used in support of this position, actually seems to invalidate it in my view. "Algorithms are exactly as basic to programmers as words are to writers." The key being words. The value in a piece of literature has little to do with the individual words it's made from, just as a piece of software is clearly vastly more than the sum of its algorithmic parts.
Footnote: I mean this for multiple dimensions of 'value', not simply some measure of economic worth. Information processing and transformation ability, for one.
Literature would be hobbled if one couldn't use certain words without fear of litigation or rent seekers suing them.
> software is clearly vastly more than the sum of its algorithmic parts.
But it's all algorithmic parts, and the whole is no less an algorithm because its parts are.
Algorithm + Algorithm = Algorithm
If you want to rely on literature as your software analogy--that the composition of words/algorithms, the expression should be protected--then you're arguing not for patent protection but copyright protection, which we already have, and ridiculously is far better than what writers get. Because we don't have to share the human readable part the source code, just the compiled products. At least with patents one was supposed to disclose something of value.
It's unreasonable to ask them to understand software as well as a software engineer.
You could say that anything at its root is simply a mathematical algorithm. These days it is possible to do a great many specialized tasks with either dedicated hardware, or software running on general purpose hardware. Should you be able to patent an asic design, but not a piece of software that does exactly the same job? Should incredibly complex and novel pieces of software not be eligible for patents simply because they're made out of 1s and 0s instead of nuts and bolts? Doesn't make sense to me.
What does make sense to me is to reform the entire patent system, in many cases drastically increasing the threshold for what is considered novel and non-obvious, and decreasing - again sometimes drastically - the time periods of monopolies offered. I don't, however, believe that this is solely an issue of software vs. not-software.
Nonsense. A brick is not a mathematical algorithm. You can use mathematical algorithms to calculate things about bricks but that's not the same thing at all.
Or yesterday there was a frontpage story about a guy who spent 13 years of his life building a (pretty cool) freeware game. I'm not arguing that he should be able to patent the mechanisms of the game, but not because it's just one big "mathematical algorithm". That's ridiculous; it's a work of art.
To use the example from the article, any design must be described with a sequence of words. You can't patent a sequence of words, so all patents should be invalid. To me, that's not much more of a stretch than saying that all software is simply a collection of mathematical algorithms. Perhaps literally true in a sense, but essentially meaningless.
Edit: By the way, since the perfect patent system is probably a pipe dream, I expect the world would be better off if software patents were abolished. I just don't believe this particular argument holds much water.
It's definitely nonsense. Things with corporeal existence are clearly distinguishable from algorithms. It is unambiguously true that a brick is not an algorithm.
> A photograph can be saved as a file on a computer. The file essentially just consists of instructions for displaying the photo. Does that mean photography === mathematical algorithms?
Traditional photography is a chemical process. The shutter opens and the light causes a chemical reaction on the film. Digital photography is an electrical process. Light strikes the sensor causing electrical charges that can be measured and recorded. Both of those are physical processes. Neither of them is an algorithm, although the data either of them produce could be processed by algorithms (e.g. JPEG compression). Moreover, a camera as a product is the sort of thing you could patent. You can tell it isn't an algorithm because you can't load software into a general purpose computer and have the software cause the computer to be able to take photographs without a camera.
> I'm not arguing that he should be able to patent the mechanisms of the game, but not because it's just one big "mathematical algorithm". That's ridiculous; it's a work of art.
That's exactly why it's not patentable. For art you get a copyright, not a patent.
> To use the example from the article, any design must be described with a sequence of words. You can't patent a sequence of words, so all patents should be invalid.
This appears to be the source of the nonsense. You aren't distinguishing between the words (or algorithms) that describe a thing and the thing itself.
> To me, that's not much more of a stretch than saying that all software is simply a collection of mathematical algorithms. Perhaps literally true in a sense, but essentially meaningless.
All software is simply a collection of mathematical algorithms. That isn't meaningless, it's the reason it's impossible for any software to exist that you could load into a computer and cause it to be able to take photographs without a camera.
> This appears to be the source of the nonsense. You aren't distinguishing between the words (or algorithms) that describe a thing and the thing itself.
That's exactly my point. By saying that software isn't patentable because it is simply a collection of mathematical algorithms, you aren't distinguishing between the algorithms that describe the thing and the thing itself. Software is literally a collection of data and algorithms, yes, just like a digital photograph. (Although you could stretch further and define data as simply a single-purpose algorithm.) And legally a "collection of algorithms" cannot be patented. But in reality, much like the photograph, there is a great deal more to software than that; there is thought and creativity involved in "collecting" those algorithms. Saying software is a collection of algorithms is akin to saying a book is a collection of words. Literally true, but missing the bigger picture.
And yes, I understand the difference between patent and copyright. I was simply giving an example of something else - a digital photo - that is exactly like software: a collection of data and algorithms. Does that mean that exactly the same laws should apply to software as to digital photographs (and everything else that is simply a collection of data and math)? Of course not, because these things are fundamentally different, and are each more than the sum of their parts.
With software there is no separate thing. The software is the only thing. If there was a separate thing you could identify it.
You can clearly distinguish between, for example, a physics text describing how bricks are affected by gravity, and an actual brick that you can pick up in your hand. The physics text is not patentable completely regardless of how creative it is or how much work went into producing it. The brick may be patentable, e.g. if you make it out of some novel weather-resistant material.
The problem with software patents is that people making texts and not bricks are being accused of infringing patents.
> But in reality, much like the photograph, there is a great deal more to software than that; there is thought and creativity involved in "collecting" those algorithms.
Now you're talking about how the software is produced. The end product is still an algorithm. Are you suggesting that if I write a computer program that can produce other computer programs, the programs produced that way should not be able to infringe any patents because they were created deterministically without any human thought or creativity? Patents have no requirement of creativity. Again, the protection for creative works is copyright, and it applies to software as well as it does to photographs and books.
More to the point, analogizing software to several other things that are all indisputably not patentable is a very poor way of arguing that software should be patentable.
Every single bit of physics is math at it's core. If you come up with a new physical theory you will be ignored unless you also include the math for it.
Eg, has anyone filed a patent on something like "swing hammer to drive nail." Or "Spin wheel to turn car?"
I personally have not heard of such ridiculous claims however it wouldn't surprise me to learn some have slipped through the "filters" that are supposed to catch this stuff.
(http://www.google.com/patents/US6004596)
I don't think the swing patent was granted, but this one certainly was. And was only challenged in court when the multimillion dollar company thought it could stop other companies making crustless peanut butter and jelly sandwiches. In 2005.
paywalled WSJ article which I can't read (http://online.wsj.com/news/articles/SB111298192348602162)
No they're not, at all. Software, as the name implies, is a commodity, which typically includes textual and graphical elements for human interaction and which interacts with multiple different mechanical and electronic subsystems. You might as well argue that mechanisms involving gears are not patentable because their behavior is expressible as a set of mathematical ratios. I'm certainly not in favor of all software patents or even software patents in general, but this notion that program = mathematical algorithm needs to die off.
Some hackers are always going on about how the judiciary fails to understand software, while assuming erroneously that they themselves have a perfect understanding of law.
The obvious answer is that our definition of what is an algorithm and what is abstract differs from the court's.
Moreover, your formulation is erroneous. Concrete things can do abstract things. The fact that a particular braking system can slow down a car does nothing to establish that "slow down a car" is not an abstract idea. It clearly is an abstract idea.
The software "only" tells the hardware what you want it do? That is the most important thing! Without software the hardware does absolutely squat.
Here's another way to think about this: imagine you invented the lever. The thing that allows you to move heavy objects you could not before, is the prior art of a log and a rock. The beam-and-fulcrum arrangement "only" tells the rock and log what to do.
> Moreover, your formulation is erroneous. Concrete things can do abstract things.
Non sequitur. Anything can be abstracted to an arbitrary degree. A very specific type of a screw with a very exact shape made of a very specific alloy can be abstractly defined as a "fastening component." That does not mean nothing is patentable.
> The fact that a particular braking system can slow down a car does nothing to establish that "slow down a car" is not an abstract idea. It clearly is an abstract idea.
Yes, but it's clearly not an abstract mathematical algorithm, which is what my parent post was questioning. Parent was wondering how software consisting of algorithms that are "abstract" can be patented. My point was precisely that the "abstract" that the court has in mind is very different from the "abstract" that we have in mind when talking about algorithms. "Slow down a car" is a different type of abstract from "E = MC^2", even though they are both abstract.
I found a small rock and a stick and I told them to move a big rock but nothing happened. I think you need to do something more than communicate information to them to get them to produce leverage.
That's the fundamental difference with software. A computer takes an algorithm and an input and produces an output. For example, you might take an image decompression algorithm and a compressed image and produce raw pixel values. The important fact is that the algorithm and the compressed image and the uncompressed image are all purely information. It's the difference between knowing something and doing something. Computers cause information to be created, but information is not supposed to be patentable, and there is nothing else there to patent other than information.
> Non sequitur. Anything can be abstracted to an arbitrary degree. A very specific type of a screw with a very exact shape made of a very specific alloy can be abstractly defined as a "fastening component." That does not mean nothing is patentable.
No, but what it means is that if you're going to say "fastening component" instead of providing anything more specific, that should not be the only part of your claim directed to patentable subject matter.
> My point was precisely that the "abstract" that the court has in mind is very different from the "abstract" that we have in mind when talking about algorithms.
Even if that were true I'm not sure what it's supposed to prove. Algorithms are all different kinds of abstract. And as soon as you make them specific you're straight into mathematical formulas and laws of nature.
>I found a small rock and a stick and I told them to move a big rock but nothing happened. I think you need to do something more than communicate information to them to get them to produce leverage.
Yes, you configure the log and rock to produce leverage. Just as you do a computer for it to do anything useful. You may think of it as "just information", but this configuration undoubtedly produces practical results via physical processes.
This ain't necessarily true. Unless you extend math to encompass reasoning about hardware, operating systems, browsers, other apps, networks etc. Math deals with more basic things.
The real reason software patents are harmful is because the industry moves so quickly, fueled by openness, that the patent examiners don't always know the current state of the art, let alone obviousness. This basically kicks the can down the road and creates waste.
> The courts have repeatedly said that mathematical algorithms can't be patented. But many judges also seem to believe that some software is worthy of patent protection. The problem is that "software" and "mathematical algorithm" are two terms for the same thing.
This is false on multiple levels:
1) Laws of physics and laws of nature and cannot be patented. Elements of the periodic table are naturally occurring materials and cannot be patented. Yet machines applying these very same laws of physics and made out of these very same materials can be patented. As can compounds that are new arrangements of atoms of various elements.
This is because machines and compounds are specific applications of patent-ineligible physical laws using configurations of patent-ineligible individual materials. Similarly, software is an application of patent-ineligible mathematical algorithms (along with a healthy dose of I/O that translates wholly non-mathematical things into mathematical representations). Just as machines are patent-eligible, software is too.
Laws of Physics : Machines :: Elements : Compounds :: Algorithms : Software
2) As evidence of the above, note that most claims for software don't even claim specific algorithms, but rather just high-level descriptions that could be implemented in a thousand ways using a thousand different algorithms. A claim will almost never say, as one of its steps, "looking up a value using a key in a hashtable" -- it will almost always say something like "determining a value corresponding to a key", which covers all O(1), O(logN) or O(N) algorithms. (If they even get to that level of detail at all... most claims are at an even higher level, e.g. "receiving a compressed image and detecting faces in it", where each element would involve a multitude of disparate algorithms) Such language could cover any number of ways of doing something.
Some take this as a sign of abstractness. But when you look at patents on physical machines, their claims also rarely ever mention the precise measurements of the various components and the actual materials used. The number of materials and relative sizes that could be used in making a mechanical invention is as numerous as the number of languages, data representations and storage technologies that could be used in making a software invention. As such, machine patents also cover all ways in which to implement it. This is not abstractness, but rather the eliding of details that are irrelevant to the essence of the invention.
3) When the court says algorithms are un-patentable, they mean abstract formulae like E = MC^2. By itself, that formula does nothing, and hence is not useful. On the other hand, software for a web browser can cause your computer to fetch data from another machine a thousand miles away and display it on your screen. Regardless of level of technical competence, everyone can agree that this is non-abstract and useful. That's the primary reason that software is worthy of patent protection: it meets the statutory requirements, i.e. "any thing under the sun that is made by man".
Software is fundamentally no different from a physical mechanism. (Math is everything.) Software is just more tunable.
The focus should be on eliminating bad patents or patents fully. Please don't muddy the issue.