Patent 5,893,120 Reduced To Pure Math

An anonymous reader writes "US Patent #5,893,120 has been reduced to mathematical formulae as a demonstration of the oft-ignored fact that there is an equivalence relation between programs and mathematics. You may recognize Patent #5,893,210 as the one over which Google was ordered to pay $5M for infringing due to some code in Linux. It should be interesting to see how legal fiction will deal with this. Will Lambda calculus no longer be 'math'? Or will they just decide to fix the inconsistency and make mathematics patentable?"

Comments

Haskell

[taktoa]

wins again!

This doesn't change anything

[Anonymous Coward]

Given enough equations, one could model the entire universe. I built a Torsen differential once in Pro/E which effectively reduced it to a list of equations.

Re:This doesn't change anything

[wagnerrp]

That's why patents used to require a physical demonstration. Ideas were all well and good, but until you demonstrated a physical device that acted like the patent claimed, you could not get it patented. This was intended to prevent fraudulent devices that looked good on paper but could never actually be manufactured, like perpetual motion machines. It could just as easily prevent software, algorithms and business methods.

Re:This doesn't change anything

[bunratty]

You can reduce any algorithm to a physical device that implements the algorithm. So therefore all algorithms are patentable. See, I can play the silly argument game too!

Re:This doesn't change anything

[dingen]

The point is that only the device is patented, not the algorithm.

Re:This doesn't change anything

[Anonymous Coward]

"You can reduce any algorithm to a physical device that implements the algorithm."

O RLY? Here's my algorithm:

1. Build a device, call it Larry
2. Bend the universe as a demonstration of what Larry should do.
3. Sell this universe to another universe, also as a demonstration of what Larry should do.
4. Deposit proceeds into a my offshore account. Explain to Larry that this step is very important.
5. Make sure Larry knows that, for every 5th loop, he's to build a Larry clone which does this recursively, instead of just a normal iteration.
6. Set Larry to work.

Re:This doesn't change anything

[bunratty]

That's not an algorithm, because the steps are not precise. An algorithm can be performed by a mindless device. "Build a device" is a vague instruction that someone needs to figure out how to do. Not all sets of instructions are algorithms.

Re:This doesn't change anything

[bunratty]

The device is not what is patented. If you build a device out of metal and that device is covered by a patent, and then I make a device using the same plans but I make mine out of wood, it uses the patented idea. Similarly, I can write an algorithm in C or Fortran, and no matter which I use if the algorithm is patented both pieces of source code use the patented idea. It's the idea of how to construct the device which is patented. Otherwise, a trivial way to get around any patent would be to build a minimally different device.

This is actually what happens in the pharmaceutical industry. If a drug is patented, the molecule itself is patented. If someone makes any change to the molecule, no matter how slight, the new molecule is not covered by the patent.

Re:This doesn't change anything

[Dragonslicer]

You can't patent atoms. A physical device is just a collection of atoms. Therefore, you can't patent a physical device.

Re:This doesn't change anything

[CheerfulMacFanboy]

That's why patents used to require a physical demonstration. Ideas were all well and good, but until you demonstrated a physical device that acted like the patent claimed, you could not get it patented.

And oddly enough, dozens of "working" perpetuum mobiles were patented. Anything proving that they couldn't work could and can not be patented.

Re:This doesn't change anything

[jcwayne]

So grandma will finally get to see The Google.

Re:This doesn't change anything

[jcwayne]

And that, boys and girls, is why you always leave one step unspecified.

Re:This doesn't change anything

[Cyberax]

"If you build a device out of metal and that device is covered by a patent, and then I make a device using the same plans but I make mine out of wood, it uses the patented idea."

Nope. That depends on patent claims. If it reads like "a metal rod for pushing things" then your "cylindrical wooden apparatus for pushing things around" would NOT be infringing.

It gets more complex if the other patent just says "a rod for pushing things around". Then your invention might still be granted a patent (for example, if your idea of using wood allows to build lightweight rods) but you _might_ need to license another patent if you want to build wooden rods.

Re:This doesn't change anything

[BitZtream]

Sounds like what happened by accident to a certain archeologist Arthor Dent ran into with Marvin.

not relevant if reducible to mathmatics.

[rubycodez]

it doesn't matter if program is reducible to mathematics, only that a claim for a software patent might be valid if it contains "a mathematical formula [and] implements or applies the formula in a structure or process which, when considered as a whole, is performing a function which the patent laws were designed to protect" http://www.bitlaw.com/source/soft_pats/final.html

Re:not relevant if reducible to mathmatics.

[betterunixthanunix]

The problem is the definition of "implementation." If I patent, say, linear discriminant analysis, and implement it using C++, what did I get a patent on? A C++ implementation? My own personal implementation? No, in the current patent system, I get a patent on any implementation -- or in other words, a patent on the mathematics itself, to within a particular interpretation of the variables and results of the computation.

Re:not relevant if reducible to mathmatics.

[Anonymous Coward]

That is not "the" problem. There are many, many problems with our current patent system.

The bottom line, however, is this: your technical mumbo-jumbo will not prevent rich people from using the patent system to prevent other people (rich or otherwise) from competing with them.

That is all.

Re:not relevant if reducible to mathmatics.

[MightyMartian]

It is the central problem of allowing software patents. I think the parent has described why software patents are absurd as concisely as I have ever seen it done. In essence we're allowing the patenting of algorithms, which are at heart pure math. We are allowing the patenting of mathematics, because judges and juries are too fucking incompetent to understand the very basis of computational science. A fraud has been perpetrated on the legal systems of many countries, but rather than throw every single person who has sought a software patent in prison for fraud, we in fact reward them by permitting them to extort money. It's a travesty that very shortly is going to bring the entire industry to its knees. Once you start going after garbage-collected hash tables and refuse to recognize that such techniques are decades old, no one could hope to implement any kind of operating system or virtual machine or, fuck, most interpreters, without risking ending up having their asses sued off.

Re:not relevant if reducible to mathmatics.

[Anonymous Coward]

Fallacy of composition, people. Its like making the argument "This chair is of wood, which came from trees, and you can't patent a tree, therefore you can't patent this chair!". Just because all software is reducible to math doesn't mean you can't patent some of it. This is as likely to open the door to patenting mathematics as making chairs will open the door to patenting trees.

Re:not relevant if reducible to mathmatics.

[RyuuzakiTetsuya]

But mechanical patents fall into the same category. Aren't all mechanical patents at it's heart, physics, which is any number of different mathematical fields? What about chemical patents?

But you're right that the danger of software patents, and as I stated above, well, any sort of patent is that our system isn't designed to handle our modern, robust(well, relative compared to our patent law) field of science and math.

The problem is our patent system but the solution isn't to throw it out and go entirely with out some sort of protection of IP in the market place. Nor is it what we've got now. Science and technology have jumped way ahead of our patent system, and if we started to figure out what was fair, we're probably not going to find a really adequate solution, even if we include the possibility of just throwing the whole system out.

I think the least worst solution would have to incorporate the idea that innovations must be protected, but, those protections must be limited.

Last thing I'd want to do is come up with some whizbang unpatented compression technique, only to find out some MegaMondoCorp has turned it into a multi-billion dollar product leaving me in the dust. on the other hand, I don't want to be sued because my for loop looks too similar to some code sitting in some code base unused for 20 years.

Re:not relevant if reducible to mathmatics.

[Anonymous Coward]

If I patent, say, linear discriminant analysis, and implement it using C++, what did I get a patent on?

IANAL, but as I understood it, you get a patent on "Running linear discriminant analysis on a machine" (i.e., a computer). If you can somehow run linear discriminant analysis without a computer, the patent doesn't apply (since you can't patent an idea or a mathematical equation).

Re:not relevant if reducible to mathmatics.

[HungryHobo]

Last thing I'd want to do is come up with some whizbang unpatented compression technique, only to find out some MegaMondoCorp has turned it into a multi-billion dollar product leaving me in the dust. on the other hand, I don't want to be sued because my for loop looks too similar to some code sitting in some code base unused for 20 years.

so you want to both have your cake and eat it too?
to be protected yourself but have the ability to go after others?

as it stands it's impossible to tell with certainty if you've violated one of the millions of patents out there as the problem is functionally equivalent to the halting problem.

so you build your lovely wizzbang application and make a profit and someone can swoop in and take it all away over a patent you have not a chance in hell of finding.

Re:not relevant if reducible to mathmatics.

[Anonymous Coward]

"a mathematical formula [and] implements or applies the formula in a structure or process which, when considered as a whole, is performing a function which the patent laws were designed to protect"

Are the patent laws designed to protect functions that purely manipulate information? I think not, and that's exactly why Bilski was struck down, and exactly this patent which (incorrectly) cost Google money.

Re:not relevant if reducible to mathmatics.

[davidgay]

Fallacy of composition, people. Its like making the argument "This chair is of wood, which came from trees, and you can't patent a tree, therefore you can't patent this chair!". Just because all software is reducible to math doesn't mean you can't patent some of it. This is as likely to open the door to patenting mathematics as making chairs will open the door to patenting trees.

Actually this just shows you don't understand programming. The inherent nature of a chair is not a tree. The inherent nature of a program/programming language IS its semantics, which is defined mathematically if you have any desire to be rigorous (and aren't patents supposed to be precise?). The only rigorous definition of a program is a mathematical equation.

David Gay
PS: Lookup axiomatic, denotational or operational semantics for details.

Re:not relevant if reducible to mathmatics.

[MaskedSlacker]

I'm pretty sure Monsanto has at least one patent on a tree.

Re:not relevant if reducible to mathmatics.

[devent]

Such a patent should be valid if the device's only function is to perform the software and nothing more (or the device is performing the software to archive one goal). But if the software is run by a general computer, than you can't patent the computer.

The lawyers try to tell then the computer is somehow magically transformed by the running software into a device that can be patented. And that is the main problem with software patents that you get a patent that covers every possible devices, existing and future devices, that run the software. And that is why a software patent is a patent on an idea, not a device. It's the same as if someone patented the idea of a horseless carriage, which covers a car, a bus, a motorcycle, and any future automobile.

Re:not relevant if reducible to mathmatics.

[aaaaaaargh!]

a software patent might be valid if it contains "a mathematical formula [and] implements or applies the formula in a structure or process which, when considered as a whole, is performing a function which the patent laws were designed to protect"

I wonder if they ever heard anything about the Curry-Howard isomorphism? And what do they mean by "a structure or process"? A formula of the untyped lambda-calculus? A formula of the untyped lambda-calculus plus the conversion rules? A formula of the pi-calculus? An implementation of this formula on a physical machine? How do they determine when two computable functions are equivalent? LOL, by the functions having the same graph? (Hopefully not...)

  How do they know when two algorithms are identical when not even theoretical computer scientists can agree on a definition that is useful in practice and theoretically well-defined at the same time?

In other words, are patent lawyers and lawmakers really that retarded?

Re:not relevant if reducible to mathmatics.

[RyuuzakiTetsuya]

Yeah, what's so wrong with that?

What's wrong with advocating for a moderate solution where in large scale concepts and innovations like entirely new algorithms to perform tasks related to commercial grade software would be protected, but, also disallow abuse by those who would make it their entire business model to abuse the system?

I never thought moderatism was extremist.

Re:not relevant if reducible to mathmatics.

[barrkel]

What you are not understanding is that the law is not a machine, nor is it like mathematics. The law relates to humans and human conceptions of the world. It's irrelevant to human conceptions of patents and algorithms that programming is inherently reducible to mathematics; what is important is the law as it is understood by humans, and how that law is interpreted to apply to any given situation. If most people consider an algorithm to not to come under the mathematical exception, then, for the purposes of the law, it doesn't come under the mathematical exception. Trying to argue using mathematical reductions is irrelevant; all you will succeed in doing by that line of argument is causing politicians to alter and/or remove the mathematical exception so that it no longer applies to algorithms, in this hypothetically convinced state.

Re:not relevant if reducible to mathmatics.

[Anonymous Coward]

Exactly. Its not saying that you can't licence your libraries, and then sell them to people. Its not saying that people can use your library without your permission.

This is about whether because you wrote the *first* implementation, now nobody can without paying you. Nobody is allowed to write that code, even if they come up with it from scratch, having never read your code. And that is simply ludicrous.

Re:not relevant if reducible to mathmatics.

[HungryHobo]

who said anything about extremism?

You seem to want a system where you somehow cannot be attacked with patents belonging to someone else you may be infringing but if someone violates your own then you want to be able to effectively go after them.

myself I'd like a pony.

Re:not relevant if reducible to mathmatics.

[martin-boundary]

That just shifts the problem onto defining exactly what is a machine.

Many people these days would say that humans are (biological) machines, so it's not clear if that would mean people couldn't run patented algorithms in their heads. Even if you ignore that obvious case, the boundaries of what constitutes a machine are extremely fuzzy. Just think of the variety of computers we have: there are silicon based computers, analog computers, quantum computers, molecular computers, etc.

We're better off fighting patents on mathematics, than fighting patents on "mathematics performed on a machine".

Re:not relevant if reducible to mathmatics.

[RyuuzakiTetsuya]

I want a system where if you don't have a product sitting on the shelf using that patent, you can't sue.

This would eliminate 90% of the patent trolls out there.

Re:not relevant if reducible to mathmatics.

[russotto]

What's wrong with advocating for a moderate solution where in large scale concepts and innovations like entirely new algorithms to perform tasks related to commercial grade software would be protected, but, also disallow abuse by those who would make it their entire business model to abuse the system?

Because 1) This idea has been patented however 2) No one has ever managed to reduce it to practice, in any field. The typical result of attempting the "moderate" solution is a system which has all the disadvantages of both extremes.

Re:not relevant if reducible to mathmatics.

[Anonymous Coward]

is it #5,893,120 or #5,893,210?

Re:not relevant if reducible to mathmatics.

[Kreigaffe]

Well that's just it, really. A mechanical patent on a.. let's say a crane. That patent is on *that crane*, of *that design*. It does not apply to all devices by which loads may be lifted, it only applies to those that use the particular sort of device described by the patent.

In other words, patenting the use of pulleys and such to create a mechanical advantage for lifting heavy crap? Not gonna fly. Patenting a particular design of pulleys? That's cool, but someone else can make a crane of their own design, based upon the same basic principals, and so long as they didn't steal or copy your design? They have a unique device.

At least, that's how things should work in a sane world. YMMV.

What this looks like is more akin to patenting the use of gravity to carry an object downward. You.. you can't fucking do that. That's absurd.

Re:not relevant if reducible to mathmatics.

[Anonymous Coward]

Thanks for bringing that up. I'll go patent x+1=y for all Anonymous Cowards. I want to see all counters have to work on the sly, or else pay a fee.

Re:not relevant if reducible to mathmatics.

[jcwayne]

What's wrong with advocating for a moderate solution where in large scale concepts and innovations like entirely new algorithms to perform tasks related to commercial grade software would be protected, but, also disallow abuse by those who would make it their entire business model to abuse the system?

Under that system Edison would have patented the idea of a light bulb, waited around for some other sucker to figure out how to actually make one, then sued for any profits a few years later.

Re:not relevant if reducible to mathmatics.

[HungryHobo]

any groups who do research but don't have a factory producing would also get swept up.

Requiring that they have at least 1 working example would be a big improvement but requiring product sales just hurts legit R&D outfits.

The patent trolls would keep on trolling, the ones which actually have any kind of working implementation of any form would just sell one unit to the company owners son.

Re:not relevant if reducible to mathmatics.

[wrook]

I don't think it will necessarily bring the industry to its knees (patent trolls notwithstanding). What it does is require that you have a certain amount of capital before you can write software. In other words, only big companies will be able to write software. I believe this is the intent of the current software patent system. The interesting thing is that patent trolls, as much as we might hate them, throw a wrench into this scenario. If large compaies use software patents in some MAD plot to keep upstarts out, what happens when the trolls extort actual money from them?

Re:not relevant if reducible to mathmatics.

[mr_mischief]

Except that Edison wasn't the only or the first person to have the idea. He wasn't even the first one to prototype a working bulb. He was just the first one to make them last long enough to be practical.

Re:not relevant if reducible to mathmatics.

[mysidia]

But mechanical patents fall into the same category. Aren't all mechanical patents at it's heart, physics, which is any number of different mathematical fields? What about chemical patents?

No, the patents are about applications of physics using specific machines.

Mathematics itself is separate from physics; and physics itself is separate from machines built that work by applying physics in elaborate ways.

Re:not relevant if reducible to mathmatics.

[Anthony Mouse]

What this looks like is more akin to patenting the use of gravity to carry an object downward. You.. you can't fucking do that. That's absurd.

Exactly.

I feel like the problem is that software has no real physical constraints. When you invent a new kind of brake for a car, it has physical characteristics. Maybe the materials are less expensive, or light weight, or they allow the car to slow down faster, or don't fade as quickly, or don't wear out as quickly, etc. Anyone else can come up with a new kind of brake design, but their design won't have the same characteristics, and if the patented design has advantages over the alternatives then the patent is worth something.

The problem with software is that those kind of constraints just don't exist. If two implementations have the same function, that's all anyone cares about. The cost of producing another instance of the software is zero, the bits never wear out, etc. In theory some implementations may use less CPU or memory than others, but in the large majority of cases the difference is not enough to care about. So the problem with patenting software is that unless you can get some incredibly overbroad claims that basically cover any possible way of doing the same thing, the patent can be trivially avoided because alternative implementations have no material disadvantage over the patented ones. But a patent on all possible ways of doing something gives the patent holder more than he deserves and causes all of the problems we see with software patents.

Re:not relevant if reducible to mathmatics.

[Anonymous Coward]

I wonder if they ever heard anything about the Curry-Howard isomorphism?

I would guess no.

Re:not relevant if reducible to mathmatics.

[Anonymous Coward]

what about prior art?

Re:not relevant if reducible to mathmatics.

[metacell]

How do they know when two algorithms are identical when not even theoretical computer scientists can agree on a definition that is useful in practice and theoretically well-defined at the same time?

In other words, are patent lawyers and lawmakers really that retarded?

The court has to make a subjective decision on whether two algorithms are substantially different, just as they do when they decide whether two machines are substantially different. It's not based on mathematical isomorphism, it's based on how obvious an algorithm is to a human expert working in the field.

The legal system doesn't need a comprehensive definition of what an "algorithm" is - they work it out on a case-by-case basis, and base later cases on similarities to previous cases (case law).

Unfortunately, software patents lead to absurd consequences in practice, since programs practically consist of patentable algorithms. As a programmer, it becomes practically impossible to be sure you're not violating someone's patent.

Re:not relevant if reducible to mathmatics.

[frig.neutron]

what happens when the trolls extort actual money from them?

Nothing. They pay and move on. Losing the occasional trollsuit is the cost of making a killing.

Re:not relevant if reducible to mathmatics.

[scot4875]

They pay and move on

And increase the price of the final product a bit; in the end, it's the customers that pay for everything.

--Jeremy

So?

[betterunixthanunix]

This point has been made repeatedly, but nobody cares. The eHarmony patent was shown to be nothing more than linear algebra with particular names assigned to each variable. People have been pointing out the relationship between software and lambda calculus since before most Slashdot users were in high school, but it has had practically no impact on the legality or public opinion of software patents.

Re:So?

[Tom]

Politics has been indisturbed by the facts of the real world for as long as history allows us to judge. All you have to do is look beyond the Solons, Washingtons or whatever your countrys famous politicians are - the day-to-day dealings of politics is a horrible mess and always has been, and things like truth, fact or evidence are way down the list of things to worry about.

The judicial branch is often quite a bit more pragmatic, but also caught up in its own world. The most important problem being that they try hard to be consistent, so change is hard to get. Once a higher court has decided on an interpretation of law, the lower courts usually don't disagree too much, and it requires a new case with new facts to get everyone to revisit the decision.

Which is cool because it means good decisions stand and aren't easily challenged just because you have a lot of money. The problem is that it takes considerable time to get rid of bad decisions. In the long run, the system works very well, but in the short run, it often fails when new facts, ideas, technologies, etc. are involved.

Re:So?

[ATMAvatar]

If changing variable names is all it takes to dodge a patent, that makes software patents pretty worthless. The names of variables are completely irrelevant to the actual execution of the program, so there's nothing stopping anyone from scrubbing their code with a refactoring tool to change those names. Having different names doesn't magically make your program do something else.

You cannot possibly be correct in your assertion.

Re:So?

[Anonymous Coward]

Yes, you are correct - Mr. Anonymous lawyer (the GP) is confusing copyright and patent. Strange for an anonymous "lawyer" on slashdot, no?

Re:So?

[Anonymous Coward]

If context is everything in a patent claim, and irrelevant in software design, it should be logical that software design and patent claims are completely incongruent. I think this is the major communication flub between Lawyers and Programmers causing this whole huge debacle.

ATMAvatar is right though, Like in mathmatics the particular names do not change the functionality of the program whatsoever. I knew one programmer who named all of his variables dickballs, tits etc as a manner of code obfuscation.

I would love to see a patent lawsuit on that beauty go down.

Re:So?

[clang_jangle]

While I'm inclined to share your skepticism, it's also a fact that there are a lot of incompetent lawyers out there whose advice isn't worth a thing. Having done business with quite a number of lawyers over the years, I can only assume there's more than one way to pass the bar exam.

Re:So?

[Anonymous Coward]

I'd love to see him write a complex program using that sort of notation.

while (jizztits >= 0) {
          fuckstain.poll(dickballs);
          jizztits -= fuckstain.entrycount();
}

dickballs.update(fuckstain.returnval());

And then try to remember, 20,000 lines later, whether or not jizztits, dickballs, or fuckstain was the object he wants to manipulate.

Re:So?

[MacGyver2210]

The code the responsible companies wrote for these technologies should be patentable(which for something written is called Copyright) - so you can not make a verbatim copy and sell it as your own - but they should not be truly patentable any more than a book is.

If I write patented code in a novel as something a hacker character came up with, but it doesn't actually run on a computer or do anything(as it is printed on a page of a book) then should you be able to come after me for a patent violation? No. You would come after me for a copyright violation. But you can't win that case, unless the code is verbatim(or close enough) to what was copyrighted originally - you can't copyright the idea behind the code.

Re:So?

[aaaaaaargh!]

Well, if lambda calculus, pi calculus, the Actor model, and so on have nothing do with software patents, then I'd really like to know what prevents someone from translating those "names" into something else entirely, which be done automatically, and thereby easily nullify any software patent. That would be great.

You see, the real problem with software patents is that there is no working definition of the sameness of algorithms, recipes, programs, or however you call it, that is both working in practice, theoretically sound, and in the meantime captures the fairly idiotic 'intuitions' of patent lawyers, lawmakers and the patent offices. And if there were such a definition, it would be clear to everyone that these patents do indeed patent pure mathematics. All of this and especially the lack of a theoretical foundations wouldn't be so bad if the patent offices would have higher requirements, would e.g. always require some concrete source code, and actually do their work and check for prior art.

But as it is these patents are simply not justifiable by any rational means, the vast majority of them is just ridiculous and an intellectual insult to anyone with rudimentary knowledge of computability theory. Turing and Church would surely throw their guts out of their graves if they knew any of this patent bullshit.

Re:So?

[The End Of Days]

You can't talk to Slashdotters about patents. Save your fingers. They all start from the wholly unproven principle that the only thing stopping them from being masters of innovation is that the law prevents them from copying the works of others, and they brook no interference with that viewpoint.

It's kinda like trying to get a toddler to understand why he doesn't own the seesaw in the playground.

Re:So?

[Anonymous Coward]

If you are implying that there is a historical precedent for not caring, thus we should continue to not care, I strongly disagree. Although, I don't think this is what you are implying, but some people might understand what you have said in that way. People need to be careful with their pessimism, it can easily become self-fulfilling.

Re:So?

[maxwell demon]

You can't talk to Slashdotters about patents. Save your fingers. They all start from the wholly unproven principle that the only thing stopping them from being masters of innovation is that the law prevents them from copying the works of others, and they brook no interference with that viewpoint.

The problem with patents is that they also apply if you don't copy.

If I implement something you implemented before, but I provably never saw your implementation, I'm safe even if my implementation happens to be very similar to yours. If however my implementation happens to match a patent of yours, it's game over, even if I can prove I didn't know that patent.

Re:So?

[gmhowell]

In the long run, the system works very well, but in the short run, it often fails when new facts, ideas, technologies, etc. are involved.

In the long run, we're all dead.
-John Maynard Keynes

Re:So?

[Yvanhoe]

Except it is the basis on which European Union declared software unpatentable. Another text (I believe it is even in the declaration of rights) lists explicitly "mathematical formulas" as being a common good of humanity and non appropriable. The point is regularly made that algorithms and computer programs are mathematical formulas and therefore not appropriable and outside the scope of patents.

Re:So?

[Anonymous Coward]

that is a REALLY bad analogy.

Re:So?

[Kaz Kylheku]

This story is about patent, not copyright. The idea behind code is patentable in the United States, and there are famous examples of ideas that are, or were patented which can be reduced to pure math.

Re:So?

[Daniel Dvorkin]

They all start from the wholly unproven principle that the only thing stopping them from being masters of innovation is that the law prevents them from copying the works of others

Whereas you appear to be starting from the wholly unproven principle that "[t]hey all start from ..." etc. In other words, a classic straw man argument. In order to back up your argument, you would have to show that every single Slashdot post on every patent-related story, including the one you just made, argues from the principle you claim is the basis of all such posts. Good luck with that.

Actually, I doubt that you can find any posts that argue from this principle explicitly, and very very few which any reasonable person would agree argue from it implicitly. Saying "software patents stifle innovation" is not the same thing at all as saying "software patents keep me personally from being the most innovative coder EVAR111," as much as you might wish it to be so. The first is an entirely reasonable statement for which there exists a massive and ever-growing body of evidence; the second is, as I said, an absurd straw man which exists only in your mass of whatever tissue trolls use for brains.

Re:So?

[Tom]

In the long run, we're all dead.

Mankind is a construct that survives beyond the death of its individual members, just like your body is a construct that survives the death of individual cells. It makes as much sense for societies to create rules that benefit it in terms longer than an individual life than it makes sense for you to invest into a future that none of the cells that make up your body today are going to live to see.

Re:So?

[Anonymous Coward]

In the long run, the system works very well, but in the short run, it often fails when new facts, ideas, technologies, etc. are involved.

Isn't the long run a series of n short runs?

Re:So?

[metacell]

But I'm not a society, I'm a person, so it makes no sense for me to support rules which only benefit society as such. The same is true for every politician, voter, businessperson, etc. Thus, the rules which only benefit society as such can't be created by individuals, they need to arise spontaneously - or, in other words, they need to be created by society as such.

Re:So?

[Tom]

Because running a marathon is exactly the same as running 420 100m dashes end-to-end? Yeah, right...

Re:So?

[Tom]

But I'm not a society, I'm a person, so it makes no sense for me to support rules which only benefit society as such.

Not in a 1:1 reciprocal calculation, but it does benefit you to support society because you benefit greatly from it. Yes, even if you're a wonky Texan with a "fuck everyone outside my ranch" attitude - your whole fucking ranch wouldn't be there if it weren't for society and its advances.

Society as a whole does indeed have a mind independent of the individual minds that make it up. Much buzzword bingo has been played with that concept, but the simple fact is that society regularily acts in ways that no individual within would. That is true for the most primitive ad-hoc societies (e.g. groups of people, say at a demonstration or rock concert) and it's true for more complex and longer-lasting ones as well.

Re:So?

[Tolkien]

Imagine if it was written in VB with EVERYTHING declared as an object. I've seen that kind of code. I told my boss there's no way I'm going to even try to port that code.

Re:So?

[turkeyfish]

But since a society is composed of individuals and can only act via the action of individuals, individuals in society must create and support rules that, at least to some extent, do not benefit the individual but society, assuming of course we are going to have civilization. Indeed, perhaps thats why its called civilization. In a civilization people act in ways that often do not benefit themselves entirely, but benefit through a shared commitment to do things collectively, even though the entire benefit is not to themselves. This may well mean that voting republican and thereby abetting their modus operandi "every man for himself", "less government regulation", and "entirely pure unregulated capitalism" results in a market place of both ideas and of goods that is inimical to civilization, which may explain why the economy always seems to do so poorly either during or immediately after republican control.

Re:So?

[metacell]

I agree.

meh

[Anonymous Coward]

He should have used erlang.

Re:meh

[Tablizer]

He should have used erlang.

Or Brainfuck. The confusion would at least delay a judgment until the patent expires.

A much needed example - hope there will be more.

[Anonymous Coward]

This is just one great example, and shows how obviously stupid patents on software are (for the same reason as patents on mathematics). I applaud its creation very much, becausewWhile I am not directly affected by US law, the indirect effects are quite visible, and there is the threat of software patents coming into european laws, too. So this actually does a lot to help defend my freedoms and sanity in the field of my chosen profession, directly.

However, I think if you really want some political or judical effects, I figure that you'd need to start a collection of these, and spread the word that it exists quite far, so that there's an actual pressure from voters. US politics and law is otherwise quite, uh, unreceptive to what people say.

http://en.swpat.org/wiki/Software_is_math

[ciaran_o_riordan]

I've collected various examples of this argument here:

http://en.swpat.org/wiki/Software_is_math

However, you have to remember that this is *not* the end of the discussion (or at best, this will result in this one patent getting invalidated or narrowed).

When faced with "software=math" arguments, judges still argue that *applications* of math to real world problems can be patentable.

What we need is legislation saying that writing, distributing, selling, and using software cannot constitute a patent violation.

Re:http://en.swpat.org/wiki/Software_is_math

[newcastlejon]

Many of us...

Oh, spare us... even if you were responsible for "spend[ing] millions on algorithm research" you'd know that if all you had boiled down to mathematics you didn't have a product at all. Patentable inventions are (read: should be) the things that are unique and inventive implementations of physical and mathematical concepts*. I'm not going to address the OP but your argument was so weak that I couldn't resist responding to it.

* The fact that one plus one makes two is obvious and cannot be seen as inventive, but a design for a processor that adds could be seen as an invention. Similarly the knowledge that non-conductive materials can become conductive under certain conditions comes out of the maths (I imagine), but designing something that uses this knowledge for a purpose is different.

Personally, I'm not willing to decide where obvious facts of nature stop and inventiveness begins, so - again, if it was up to me - I'd give the benefit of the doubt and drop the idea of patents altogether. Patents do, allegedly, promote innovation, however, so I made the decision years ago just to steer clear of the whole mess as best I can and leave such things to people who have more... shall we say... flexible standards.

Re:http://en.swpat.org/wiki/Software_is_math

[oliverthered]

is it a general purpose machine or a specific purpose machine? is it an objectifiable patent or a meta-patent?

Re:http://en.swpat.org/wiki/Software_is_math

[Thing 1]

Personally, I'm not willing to decide where obvious facts of nature stop and inventiveness begins

We are a product of nature. So, it doesn't stop. The existence of patents in this universe is the part that does not fit the physical laws. I agree with your recommendation to get rid of patents.

Patents do, allegedly, promote innovation

I think it can be shown that they retard progress. Perhaps they were useful immediately after their creation, if the environment was so chaotic that business just couldn't get done, and this helped to straighten it out. (I question whether that was the case, though.) But these days, when technology progresses faster and faster each year, patents can make a company spend a great deal more on an alternative (and weaker!) implementation. Society functions best when its actors have the least amount of wasted effort; the patent system helps to create "industrial waste" by having multiple actors repeatedly solve the same equation, rather than sharing their work.

Re:http://en.swpat.org/wiki/Software_is_math

[The End Of Days]

So with patents, I have the incentive to invest time and money in the creation of some widget, and to publish how said widget is created, because I have a possibility to get a return on that investment due to the exclusivity granted me by the law. I also have the incentive to come up with a new way of doing what someone else did, possibly improving it, for the same reason (you called this part waste).

What are my incentives without patents? I can come up with "do nothing, since I don't have to" and "hide my work, since I have no protection."

Can you point me to the positives? Stick to reality, please, and don't assume that all humans will magically conform to some utopian ideal of behavior.

Re:http://en.swpat.org/wiki/Software_is_math

[hweimer]

What we need is legislation saying that writing, distributing, selling, and using software cannot constitute a patent violation.

Unfortunately, this is not allowed under TRIPS. The only way to prevent software patents is by not granting them to start with.

Re:http://en.swpat.org/wiki/Software_is_math

[Anonymous Coward]

What are my incentives without patents? I can come up with "do nothing, since I don't have to" and "hide my work, since I have no protection."

Can you point me to the positives? Stick to reality, please, and don't assume that all humans will magically conform to some utopian ideal of behavior.

You can choose trade-secret and use the head-start you have by being the first to come to the market with a useful implementation. On long term, it's likely that the implementation will be reverse-engineered, thus it's likely that the trade-secret alone would be a better incentive to keep inventing (run faster then your competitors).

irrelevant

[t2t10]

Yes, the mathematical formula is not patentable, and you can compute with it all you like. But when you actually apply it to perform information retrieval in a physical computer, then the patent applies.

I'm not saying that the patent is valid (there is tons of prior art), or that such a distinction ought to be made. But it is certainly a distinction one can make.

And if people keep insisting that such a distinction is meaningless, it is more likely that the patentability exclusion for mathematical formulas/facts will get dropped than a lot of patents will become invalidated.

Re:irrelevant

[SpiralSpirit]

The question is more-> if you have a patent that is a formula derived from "open-source" mathematics that anyone can use, how valid can any patent on it be? It's not a technique, per se. It's not new. It's a mathematical manipulation that a million people before you have thought of, learned, and used. Could engineers have patented the formulas used to create objects like engines, etc? How many automakers would we have now if the formulas used in the design of cars were patented by those making the cars? They're pure math, we know, but hey - they're being applied.

Re:irrelevant

[91degrees]

What should be patentable then? Nintendo's Wii controller? All the hardware used to make it already existed, the actual wiring should be fairly obvious once the purpose was described. I'm sure the technology for identifying a bright spot in an image is already well known.

The innovation was in putting this all together and combining it into a means of controlling video games.

Re:irrelevant

[maxwell demon]

What should be patentable then? Nintendo's Wii controller? All the hardware used to make it already existed, the actual wiring should be fairly obvious once the purpose was described. I'm sure the technology for identifying a bright spot in an image is already well known.

The innovation was in putting this all together and combining it into a means of controlling video games.

Something shouldn't be patentable just because it's an innovation. If it's obvious how to do it as soon as you have the idea to do it at all, it's clearly not patent worthy. Patents should be for inventions, things that are not obvious how to do it, even when you know that you want to do it.

Re:irrelevant

[91degrees]

So, devils advocate time, can you suggest examples of things that should have been patentable? I can certainly find a lot of things that by this argument shouldn't but have been.

Re:irrelevant

[maxwell demon]

Well, for example Edison's patent on the light bulb (which was not about the idea to use a glowing wire for making light, but about how to get that wire last for an extended time; the idea of using a glowing wire would IMHO not have been patent worthy, and additionally wasn't exactly new back then anyway). It was far from obvious how to make that damned wire last long enough to make the light bulb worthwhile for actual usage.

Re:irrelevant

[t2t10]

What's patented is the computer system that implements the formula, not the formula. As long as you just compute with the formula (numbers in, numbers out), you're fine and non-infringing.

Once you use the formula to achieve a physical effect in the machine, you have a physical mechanism, and that is patentable. In this case, the physical effect is how the blocks are arranged on disk.

So, formulas for engine design are not patentable; you can calculate with them by hand or machine, as long as all you ever get out is human-readable information. However, the engines designed using those formulas are (including specific dimensions of engines if novel and non-obvious). Furthermore, an engine controller that uses the formulas to control valves, timing, etc. may be patentable.

This will ruin my day...

[Tasha26]

Software patents are the kind of thing that make me sick in the stomach... what's the point of even going to university if the end result is discovering that your idea has already been patented ages ago.

Re:This will ruin my day...

[PCM2]

what's the point of even going to university if the end result is discovering that your idea has already been patented ages ago.

Well, if your goal is to create innovative new work, I imagine the point of going to university is learning what has been done before.

On the other hand, if you enroll in an undergraduate program with an amazing, brilliant idea in your head, only to discover that it had already been patented ages ago, maybe it's a wake-up call that you aren't the most brilliant person ever born, after all, and that you still have some studying to do.

Re:This will ruin my day...

[newcastlejon]

Real engineers go to university too, you know, and the point in going is to learn the skills one needs to come up with one's own ideas and to articulate/instantiate them.

Baiting aside, the argument that software patents make a degree in the computer sciences worthless isn't why people are so up in arms about them. The problem is that comp. sci. can cover areas exempt from patents - like mathematics - and other areas that are subject to patent law. For my two penneth, legislators have drawn the line too far into the former's camp.

Re:This will ruin my day...

[Anonymous Coward]

what's the point of even going to university if the end result is discovering that your idea has already been patented ages ago.

Well, if your goal is to create innovative new work, I imagine the point of going to university is learning what has been done before.

University is unable to do that. There has been waaaaayyyy too much created to learn it all in 4 or even 8 years of study.

On the other hand, if you enroll in an undergraduate program with an amazing, brilliant idea in your head, only to discover that it had already been patented ages ago, maybe it's a wake-up call that you aren't the most brilliant person ever born, after all, and that you still have some studying to do.

Just because it has been invented before doesn't mean one isn't brilliant. I would think someone who "invented" a transistor on their own and who has never read a scientific journal and has been completely cut off from modern civilization as being brilliant.

Then again, if you keep studying you start to think more and more like the status quo so that when someone who is truly innovative comes along, you can poo poo them for not thinking correctly - which is what happened to all of the greats.

Then there are folks who didn't study that much formally and have done tremendous things.

Formal education is an asset but it is also overrated. And for some subjects, I think formal education handicaps one's thinking - Business for one.

Re:This will ruin my day...

[Anonymous Coward]

what's the point of even going to university if the end result is discovering that your idea has already been patented ages ago.

Well, if your goal is to create innovative new work, I imagine the point of going to university is learning what has been done before.

University is unable to do that. There has been waaaaayyyy too much created to learn it all in 4 or even 8 years of study.

I wouldn't need to know the ins-and-outs of every steam engine invented to understand thermodynamics. However if we take the parent as read then strictly speaking no, I wouldn't be able to learn about everything done in the field in the space of even a week if I had 4 years to do it.

On the other hand, if you enroll in an undergraduate program with an amazing, brilliant idea in your head, only to discover that it had already been patented ages ago, maybe it's a wake-up call that you aren't the most brilliant person ever born, after all, and that you still have some studying to do.

Just because it has been invented before doesn't mean one isn't brilliant. I would think someone who "invented" a transistor on their own and who has never read a scientific journal and has been completely cut off from modern civilization as being brilliant.

Of course, they're brilliant in the sense of being a genius - they discovered from first principles what took a whole body of research decades. Then again, they aren't brilliant in the sense of producing something that ends up outshining other products of research. We aren't talking about what makes a person intelligent or insightful or whatever; we're talking about what makes an idea (brilliant or not) patent-worthy. If I were the parent with a great idea that had already been done then I'd expect the world's smallest violin at most and chalk it up to "that's life" or realise that it might not have been such a brilliant idea to start with.

Then again, if you keep studying you start to think more and more like the status quo so that when someone who is truly innovative comes along, you can poo poo them for not thinking correctly - which is what happened to all of the greats.

Perhaps you're right, but suggesting that all of the world's great thinkers have been the victim of a - possibly unwitting - conspiracy isn't so solid an argument. Word to the wise: don't use the word all unless you really, really mean it.

Then there are folks who didn't study that much formally and have done tremendous things.

So? The odds say that once in a while a stupendous intellect will be born: this says nothing about the relative merits of formal education

Formal education is an asset but it is also overrated. And for some subjects, I think formal education handicaps one's thinking - Business for one.

Hmm, "creative" economics is a double-edged sword at best, and I'm trying to be tactful putting it like that.

Re:This will ruin my day...

[Anonymous Coward]

Can't troll worth shit, too stupid to make an intellectually honest point. Yup, it's an American.

Re:This will ruin my day...

[VortexCortex]

what's the point of even going to university if the end result is discovering that your idea has already been patented ages ago.

Well, if your goal is to create innovative new work, I imagine the point of going to university is learning what has been done before.

On the other hand, if you enroll in an undergraduate program with an amazing, brilliant idea in your head, only to discover that it had already been patented ages ago, maybe it's a wake-up call that you aren't the most brilliant person ever born, after all, and that you still have some studying to do.

On the other hand, If your goal is to create your own software using your own ideas while ignoring the whole field of CS research, prior art and patents as their information is not accessible to you, and it matters not if wheels get re-invented because you are inventing them to learn how they work: The patent system can and will prevent you from profiting from your original works.

I was exposed to computers at a young age (7?) I taught myself BASIC on the Apple IIe, and on an IBM 386. Learning things by doing, I had needed more functionality and therefore learned Assembly and C by the age of 12. To solve a problem I would just think, how would I ___ in C? Then write an outline, and code it... I had no access to any other source code, or even higher mathematics.

EG: In Junior high -- age 13, having been taught about algebra and line slopes ratios I was able to render lines. For a vector graphics space battle game, I actually "invented" the unit circle, and sin/cos tables (I called them angle tables) by realizing line slopes could be used to roughly determine the degree of angle between two lines. I built a table of all 5.625 degree increments to line slopes, and interpolated between the line slopes to derive my "decentAngle()" function. Three years later the American public school system introduced me to trigonometry, and I told the teacher, "Gee, that Pythagoras guy wasn't that much of a genius, anyone can figure this out..." while showing her my sin() approximation functions and distance theorem... Oh, so that's called a "hypotenuse"?

Yes, the C library had a sin() function, but I didn't know what to use it for -- The compiler manual didn't teach trig...

As a teenager I built my first interpretor and virtual machine, without ever hearing those terms (I needed command language for my Doom .WAD (map) editor to build stairs, align textures across "curved" walls, etc.)

I had an expensive Borland C compiler, it's manual, a book on QBasic, and a book on x86/x87 Assembly. Understanding that a C compiler must somehow generate assembly like language in order to make .COMs & .EXEs, I wrote my first compiler by reverse engineering complied assembly instructions -- A "command to machine code table" was used to directly piece together an executable... it was ugly, but it worked, and I know better now (thanks to the existence of the Internet).

At age 16 I realized I could make money selling my programs. I distributed my software as share-ware and trial-ware for free on BSS systems -- after a the trial period a pass-code was needed to unlock the program. It was easily bypassed by re-installing or setting back the system clock, but was just enough of a deterrent that people would pay to get an unlock code. Hey, I would have released it as source code, but I needed $$$; Dating girls, driving cars and buying new compilers for MS OSs isn't cheap (besides, few others would likely compile the source, and if people asked I would show them how I achieved things).

Almost exactly 2 years after I started selling MY software I received my first cease and desist notice. I didn't have the money to dispute the claim in court (and neither did my family), so I stopped distributing the application. My simple CAD program was apparently infringing

Re:This will ruin my day...

[kubajz]

Not only an undergraduate program is there to show people what had already been discovered. In the Czech Republic, I had completed my undergrad and grad program, and when I enrolled in postgrad (Ph.D. in Information Technology) I was required to spend a full year on FINDING a topic that had not been researched before. Most people do not realize that their "brilliant new idea" may have been discovered 7 years ago, by a team of 8 people working on it full-time, and organizing annual conferences to share further developments in that "brilliant new idea".

Re:This will ruin my day...

[swilver]

Exactly how I feel.

Programmers invent their own algorithms all the time. Sure it may be reinventing the wheel, but that's part of the fun -- most wheels don't take that much time to invent anyway and it gives a much deeper understanding of what you are doing than having the solution just handed to you.

Contrast it with looking through millions of patents, obtaining a (costly) license for your (free?) software and implementing something you only have a borderline understanding of... Reinventing the wheel suddenly doesn't look so bad anymore.

Re:This will ruin my day...

[Theaetetus]

On the other hand, If your goal is to create your own software using your own ideas while ignoring the whole field of CS research, prior art and patents as their information is not accessible to you, and it matters not if wheels get re-invented because you are inventing them to learn how they work: The patent system can and will prevent you from profiting from your original works.

Yeah, and that's kinda the point. The patent system isn't about rewarding your hard work. What, you think the government has an interest in making sure you get a pat on the head? No, it's about economic efficiency in innovation. It's a waste if you and others spend time re-inventing the wheel. You should be making wheel improvements or wheel replacements, not the same wheel over and over again.

The point the GP is making is that you can't even make use of your ideas. I don't care if I'm not the most "brilliant person ever born". That shit is for e-peen strokers, and mental masturbation. I JUST WANT TO BE ABLE TO USE MY IDEAS in my own software, and maybe even sell software implementing my own ideas. -- Fuck right off man, you have NO IDEA what it's like to pour your life's energy into a project that you built from SCRATCH (including sin() and cos() functions), and later have someone swoop in and gavel you on the head with a patent infringement suit and keep you from even distributing it for free. SOFTWARE PATENTS STIFLE INNOVATION. END OF FUCKING DISCUSSION!

Yes, it's the end of the fucking discussion, because, as you admit your ideas were not innovative. They were already known and used commercially. You recreated that known work, but that's not innovation. And nothing of value was stifled.

Oh, and since you like pointing out how not special people are: Before you jump into a text book, next time think it through first -- "How would I do _____ ", when you do look it up -- and find that method is patented, or highly praised as "genius", you'll have your own fucking "wake-up call" and realize just how OBVIOUS, EVERYTHING IS.

If everything is so obvious, then build us a flying car. It's obvious, so shouldn't take you long. Say, full plans by the end of the day? Excellent.

Re:This will ruin my day...

[badkarmadayaccount]

The technical parts are obvious. First, I'd like to see a legal framework surrounding the hypothetical vehicles. Maintenance, traffic laws, safety and fuel requirements. Everything. And a business analysis on how sellable are these, long-term.

Re:This will ruin my day...

[Theaetetus]

The technical parts are obvious. First, I'd like to see a legal framework surrounding the hypothetical vehicles. Maintenance, traffic laws, safety and fuel requirements. Everything. And a business analysis on how sellable are these, long-term.

Really? "I could totally build a flying car... it's just that I'm sooooo concerned about traffic. So, I'm, uh, not going to tell anyone how to build it."

Concern trolling doesn't count as innovation either.

Re:This will ruin my day...

[badkarmadayaccount]

While I agree that the GPP overstated, your way of bringing him down to earth is no more incorrect or rude than what I did. Truce?

Re:Wheel was patented too recently...

[Anonymous Coward]

Parent NSFW, annoying too.

Re:Wheel was patented too recently...

[Anonymous Coward]

You fucking piece of shit!

Re:Wheel was patented too recently...

[Anonymous Coward]

Troll link, do not click.

As compared to the Manhattan Project...

[JoeMerchant]

I'm pretty sure that in the late 1930s, the process to make Little Boy and Fat Man would have been patentable under any purview that considered any kind of patents to be valid.

And, in the end, all they are is the bringing together of big lumps of elementary material, albeit highly refined from what is found in nature, to bring about a reaction that, itself (after decades of study), can now be reduced to a set of mathematical equations.

As for software patents, some wiener patented the use of XOR to draw a cursor... if that can stand, why not anything else, like one-click order taking?

I truly hope we have hit bottom of the patent well, but I'm sure there's an army of attorneys out there willing to dig with ever sharper tools.

Re:As compared to the Manhattan Project...

[Anonymous Coward]

I look at it this way. Sure all the crap will be patented. But patents expire.

Lets get the stupid junk into the system. That way it does not pop up later.

And I wrote a system in 1997 that did exactly what that patent says. As I had a slow system with 'old' data. Remove old data and system performs again.

Re:As compared to the Manhattan Project...

[citizenr]

I'm pretty sure that in the late 1930s, the process to make Little Boy and Fat Man would have been patentable under any purview that considered any kind of patents to be valid.

They were patented by Leo Szilard, US government stole/took over for free those patents and screwed the inventor.

Re:As compared to the Manhattan Project...

[MacGyver2210]

That is a horribly ignorant way to look at it. Physical materials - even relatively simple ones like base elements - cannot be simply described as a mathematical value. You can say "Oh, it's element___ with this many protons, neutrons, electrons, quarks, leptons, etc" but you'd still be leaving out a MASSIVE amount of data which would be required to completely describe(mathematically) the object's form and composition.

This is as opposed to trying to patent the chemical chain reaction equation of Uranium, Plutonium, or any other fissile material. You can't do that...which is essentially what this argument is over.

You should be able to patent YOUR SPECIFIC, ORIGINAL, BUILT-AND-TESTED device, if no prior work has beat you to it. You should NOT be able to patent the theory behind something such as a chemical reaction equation, or even a simple algorithm. When it comes to software, even the built-and-tested device is still just mathematical numbers floating around in the computer, so the line is seriously blurred.

All around, Copyright law in the US and other developed nations has gotten seriously out of control. I think a giant revision is needed pretty badly to avoid the ultimate downfall of society at the hands of patent giants and megacorps who all just want to own people's imaginary property.

Re:As compared to the Manhattan Project...

[devent]

How is bringing together material and build something is math? Do you know what it means to build something? You have get a hammer and screwdriver in your hand and put nails or screws into holes and you have to cut metal or wood. In the end you have some device, a product, which either works or not.

Math is just a bunch of formula on the paper with freely defined meaning. There is no device, no product at the end. It doesn't work and it can't work because it's just formula on the paper. The same is with software. Only if you run the software on a device you have a product. But how can you patent a general purpose computer that the whole point is to run software? You can't.

So what do you want to get a patent for in a software patent? The software is no product unless it's run on a computer. So you can't patent the software and you can't patent the computer.

If you build a special device that only runs your software and that device is your device (meaning it's not a general embedded computer, like the ARMs) sure you can get a patent on that device.

Re:As compared to the Manhattan Project...

[drb226]

As for software patents, some wiener patented the use of XOR to draw a cursor... if that can stand, why not anything else, like one-click order taking?

New patent idea: one-wink order taking. My webpage activates your camera, and it only takes one wink to take your order.

Re:As compared to the Manhattan Project...

[JoeMerchant]

The patent system is expanding so quickly that no one person can ever know everything that's in there, even when focused to their field.

Re:As compared to the Manhattan Project...

[JoeMerchant]

That is a horribly ignorant way to look at it. Physical materials - even relatively simple ones like base elements - cannot be simply described as a mathematical value. You can say "Oh, it's element___ with this many protons, neutrons, electrons, quarks, leptons, etc" but you'd still be leaving out a MASSIVE amount of data which would be required to completely describe(mathematically) the object's form and composition.

I'm pretty sure that most atomic weapon design and testing has been done by mathematical (granted, probabilistic Monte Carlo type) simulation. If the reports are to be believed, those programs describe the end result very accurately.

Re:As compared to the Manhattan Project...

[multipartmixed]

Did Leo build a working model?

Re:As compared to the Manhattan Project...

[Anonymous Coward]

If your device is not turring complete you will have a hard time running software on it that is of any utility. If it is, any other Turing complete machine will be able to run the software given enough time and memory.

Math is symbol manipulation by a set of precisely defined rules. All software is symbol manipulation by a set of precisely defined rules.

Not sure I understand this argument at all

[PCM2]

Sure, software is math. I'll accept the point as stated.

Likewise, books are language. Can books be copyrighted? No one owns language.

A power drill is metal and plastic. Can a power drill be patented, then?

Patents aren't limited to ideas that have never been remotely observed before in all of human existence. You patent innovations. Many innovations are incremental, probably even the vast majority of them. I don't see how saying "but that's just math" invalidates the significance of a particular innovation. There's a big difference between saying "you could probably do that with a computer" and actually doing it with a computer, which is what a software patent would cover.

Re:Not sure I understand this argument at all

[ciaran_o_riordan]

I'm not a fan of software=math arguments, but an example of their value is in the US Supreme Court's "Flook" ruling:

http://en.swpat.org/wiki/Parker_v._Flook_(1978,_USA)

That's the starting point I'd use if I had to argue based on software being math, but I'd rather not rely on that.

Being equal to math isn't the reason why software patents are a problem for society. Blocking software development, forcing incompatibility, stifling competition, and being incompatible with the development models (lots of individuals and SMEs write software) and certain popular distributions models (freeware, free software) are the reasons, and we should focus on them.

If we say software=math, the court could say that only some software is math, and then patent drafters just have to formulate their claims such that they fit the court's definition of non-just-math software. It's not a path to victory.

Re:Not sure I understand this argument at all

[Draek]

Sure, software is math. I'll accept the point as stated.

Likewise, books are language. Can books be copyrighted? No one owns language.

Wrong. Words, when put together, form sentences which form paragraphs which form stories, which in turn can be copyrighted. But stories are not paragraphs, paragraphs are not sentences, and sentences are not words so the fact that words cannot be copyrighted by itself is not relevant to stories.

Mathematical formulae however, when put together they're still formulae, merely a longer (and probably more descriptive) one, and mathematical formulae cannot be copyrighted or patented.

There's a big difference between saying "you could probably do that with a computer" and actually doing it with a computer, which is what a software patent would cover.

Except this isn't saying "you could probably do that with a computer", this is saying "you can do that with a computer like this". Allowing software patents but denying math is much like allowing copyright over hexadecimal numbers but disallowing it for decimals: there exist a way to transform one into the other that's formally proven to work for any and all elements of either set, so anything that works for members of one set *must* work for the other and to declare otherwise is to automatically fall into a contradiction. If you'd study Math or CS it'd be much clearer as to how and why, if you must know, but that's kind of outside the context of this post.

Re:Not sure I understand this argument at all

[Appolonius of Perge]

Nobody is arguing against protecting a specific implementation of an algorithm (although copyright already provides this protection).

The problem is that when you patent an algorithm, you don't just patent that expression of that idea, like you do with a book or a power drill. You patent all the expressions of that idea. It would be like writing a book on some topic, and then owning the rights to all books on that topic until the patent runs out. You own a whole chunk of the language, every possible expression of you idea, not just the particular one you came up with.

Likewise, when you patent an algorithm, you have patented an entire (admittedly fairly confined) branch of mathematics, having to do with expressions of that algorithm. This is distasteful, and, as the current software patent climate has shown, has terrible chilling effects on the software industry as a whole.

Re:Not sure I understand this argument at all

[Anonymous Coward]

because it would kill the whole industry. Patents unlike copyrights are intended to promote and protect innovation, however they're being used as a weapon against anyone trying to produce, and fear monger people from ever joining the industry in the first place. Software patents don't cover actually doing it with a computer, they cover the idea of actually doing it with a computer and all kinds of implementations thereof.

It's killing the industry, and giving money to people who shouldn't get it.

Re:Not sure I understand this argument at all

[maxwell demon]

Likewise, books are language. Can books be copyrighted? No one owns language.

There's a difference. Copyright covers only the specific expression. Patents cover the whole idea.
For example, the copyright on Harry Potter covers the story on Harry Potter, and derived works. A patent on Harry Potter OTOH could look like this:

Claims:
1. A story about a normal, underprivileged boy who turns out to be special.
2. As 1, where the specialty is that he actually is the son of a magician.
3. As 1, where the boy lives in England.
4. As 2, where the boy himself gets educated in magic. ...

You see, it would cover a lot of possible books, most of which would have very little relation to Harry Potter. Even worse, it would even apply to books of authors who never heard of Harry Potter (unlikely in case of Harry Potter, but the same would be true for quite obscure books as well). Or imagine that someone else had filed such a patent before, without actually writing such a book, then Harry Potter would not have been a success story, but a nightmare for J. K. Rowling.

Patents have to be much stricter in what they can be applied to because they are much broader in scope.

Re:Not sure I understand this argument at all

[Nemyst]

You can derive mathematics in isolation. Case in point, many mathematic proofs and "discoveries" have been done independently and simultaneously (namely differential calculus with Newton and Leibniz).

You cannot, however, "derive" a book or end up with the exact same arrangement of words, apart from chance. Likewise, your drill will not have the same configuration, elements, inner workings, etc. It'll be different enough to be considered unique in its own right.

Re:Not sure I understand this argument at all

[Schadrach]

Your book example is a failure, because you switched which variety of IP you were talking about.

Software should fall under either copyright or patent, not both. Personally, I'd favor copyright despite it's ridiculously long terms, simply on the whole "pure math isn't patentable. software is an expression of math in a machine readable form, and nothing else; therefore, it should not be patentable."

Let me put it this way, if I added "in French" to something that was explicitly not patentable, it would not make it patentable. Adding "in machine language for some computer device" shouldn't either.

Re:Not sure I understand this argument at all

[Schadrach]

You actually brought up something I've wondered about: When Sony, or the DVD consortium, or whoever it is *this* time start crying about the release of their respective signing key and claiming copyright to the number, why not just start representing it in decimal, or as the multiple of two primes, or some other form that is clearly not subject to copyright/patent?

For exqample, the so called "09 F9" AACS key is 13,256,278,887,989,457,651,018,865,901,401,704,640, so that decimal number in and of itself is now copyrighted? Or patented?

Maybe we should make a website that generates sequential numbers from 1->arbitrarily large, discounting those that are known to be things like the CSS key. If any other number is used as a cryptographic key in a commercial product, they would clearly be violating our copyright on that particular arbitrarily large number, right?

Re:Not sure I understand this argument at all

[Anonymous Coward]

Exactly.You got it in one. Software should not be patentable

Software is language. A creative work, like a book. You should be able to copyright it - and you can.
You can't patent a book, thats just stupid. So why can you patent software?

Re:Not sure I understand this argument at all

[Longjmp]

The best comment I ever read about this issue.

Sorry, no mod points ;-)

Re:Not sure I understand this argument at all

[skywire]

Copyright covers only the specific expression.

This was long one of the cornerstones of US copyright law. You must have been in a Rip van Winkle nap while that principle was being excised from the law by clever judges over the last few decades. Sadly, you can't even write a completely novel imagining of an older Holden Caulfield's behaviour without falling afoul of the current 'copyright' case law.

Re:Not sure I understand this argument at all

[skywire]

Whereas, as you rightly point out, a copyrighted work will not, due to the laws of probability, be duplicated without copying it, such is hardly the case for an invention like a drill, as so many have to their chagrin discovered.

Re:Not sure I understand this argument at all

[oliverthered]

Likewise, books are language. Can books be copyrighted? No one owns language.

Copyright is not a patent and a breach of copyright may make you a parrot but it does not make you a thief. If it's a book of facts, then no you can't copyright those facts.

A power drill is metal and plastic. Can a power drill be patented, then?

I've made one from wood and gas, a plastic and metal one may be patented, but I can still make a wood and gas one.

it's the implementation of the mechanism not the power and the drilling, which can be done in a number of ways.

Re:Not sure I understand this argument at all

[MacGyver2210]

Rowling made so much money off the Potter series that even if she had been sued successfully for patent infringement, she still would have come out stinking rich...

Re:Not sure I understand this argument at all

[Anonymous Coward]

False argument, sorry you are the weakest link.

Copyright != Patent

Re:Not sure I understand this argument at all

[The End Of Days]

Yeah, the tech industry is totally dying... it's only bigger than it's ever been, and growing faster than it ever did before. Clearly being eaten from within by this scourge.

Re:Not sure I understand this argument at all

[Jeremi]

Rowling made so much money off the Potter series that even if she had been sued successfully for patent infringement, she still would have come out stinking rich...

Either that, or the person who sued her would have come out stinking rich.... I don't think there is any cap on how much a person found guilty of patent infringement might be required to pay.

Re:Not sure I understand this argument at all

[Anonymous Coward]

Books cannot be patented - you can't patent 'Sci Fi'. Indeed, you can only copyright the work you produced. Many people, including myself, hold that software should follow exactly the same standard. You should only be able to copyright your particular implementation, i.e. your code. If another programmer is able to implement the same functionality, using their own code, this is akin to someone else writing a book about a time travelling robot who is bent on destroying mankind. As long as he doesn't use the same characters, places and plot, and doesn't call it 'Terminator', it's not breaking any laws.

Re:Not sure I understand this argument at all

[maxwell demon]

The key is definitely not patented, and I'm not sure if it is copyrighted. I think it is, however, a trade secret (well, the actual trade secret is the fact that this number is the key). And as such, I'm pretty sure that it doesn't matter which form you put it in, but only whether the original binary form can be recovered from whatever you provide. Putting it in different forms would probably make it harder for them to find it, but I strongly doubt it would change the legal implications.

Let's make an analogy here: Say the admin password in some company is "god". The admin password is most certainly secret, that is, you may not reveal it to anyone else. This doesn't mean you are not allowed to say the word "god" at all, however you are not allowed to tell anyone the fact that this is the admin password of that company. And it doesn't matter if you say "the company's admin password is 'god'" or "you get the admin password from 'fnb' the same way as you get 'IBM' from 'HAL'". All what matters is that a person who didn't know the password before now knows it.

Re:Not sure I understand this argument at all

[icebraining]

Maybe we should make a website that generates sequential numbers from 1->arbitrarily large, discounting those that are known to be things like the CSS key. If any other number is used as a cryptographic key in a commercial product, they would clearly be violating our copyright on that particular arbitrarily large number, right?

No, because those numbers wouldn't have the right colour.

Re:Not sure I understand this argument at all

[kenshin33]

The key is neither copyrighted nor patented but rather a trade secret. Anyone entrusted by SONY with that secret cannot divulge it .. but if it is obtained by any other mean it is not protected by anything. The copyright claims in regard to the PS3 suit was about circumvention --or rather potential for it-- (which is by US law equated with infringement) not the key it self.

Re:Not sure I understand this argument at all

[donscarletti]

Likewise, books are language. Can books be copyrighted? No one owns language.

Books most certainly can't be patented, not sure why you are bringing copyright up. I think _most_ of slashdot agrees with copyright (apart from the commenters on P2P file sharing stories), kind of bashing a straw man here.

Re:Not sure I understand this argument at all

[Anonymous Coward]

2+2=4...ooops, that's patented, so 2+2=D?

Re:Not sure I understand this argument at all

[benhattman]

Likewise, books are language. Can books be copyrighted? No one owns language.

Yes, but they cannot be patented. It's important to keep in mind that patents, copyright, and trademark are not natural phenomena of a market. They are things we have decided improve the innovation of the market, at least under certain circumstances. If a class of patents hurt innovation, or harm innovation, then we should eliminate them.

Re:Not sure I understand this argument at all

[Anonymous Coward]

Do they still hand out mod points? I haven't seen any since the last site redesign.

Re:Not sure I understand this argument at all

[Jason Levine]

mathematical formulae cannot be copyrighted or patented

Try telling that to Weight Watchers who patented Points = (Calories/50) + (Fat/12) - (Min(Fiber, 4)/5)

They've sued "points calculator" programmers for infringing on their patent.

Of course, now they've moved on to their new Points formula which is much more complex, but still is a patented mathematical formula.

Re:Not sure I understand this argument at all

[I'm not really here]

Cross Licensing by big corporations keeps them going, but how does John Smith's "Wizbang Application House" get access to these software patents so that they have a chance to compete? The cross licensing agreements among large corporations create a MAJOR barrier to entry in many software markets.
 
I say allow patenting any mathematics for which there is no prior art, but the patent expires 1 year after issuance and becomes public domain. You don't want your source in the public domain? Keep it as a trade secret instead then, but don't whine when it leaks into the public domain anyways. Coke has kept their recipe a trade secret for a very long time, and it works for them.

Re:Not sure I understand this argument at all

[Anonymous Coward]

It's not that you could do it any a computer, but that you could do it with any computer. (given enough time and memory). The transformation of the mathematics into machine states is exactly what you would expect a skilled programmer to be able to do. Once reduced to mathematical formulas, the implementation is obvious to those skilled in that art. Once you have a formula, the actually doing it with a computer is routine.

Re:Not sure I understand this argument at all

[Anonymous Coward]

There is a fundamental difference when it comes to software. If software is math, you may think of the following things:

1. Math is universal, which means that unlike language, is not stated by cultural development: math is an abstraction of fundamental properties of the universe.

2. Language and inventions are creative products by applying sciences, while mathematics (and software) use creativity to find ways to discover relations between quantities and space. You can also say that inventions and books can be created, but math can only be discovered.

3. Mathematics and software can be applied in real-life processes, and those processes end up having a creative effort. There's a catch, however: a mathematical altorithm is unambiguous, so to achieve a result, there is only one way to achieve it. You write a happy/sad story in a book in an infinite amount of ways, but in the case of mathematical algorithms, it would be like patenting a straight trajectory between two lines.

3. If math is already part of the reality as we know it (1), you can't "create" it. Likewise, if you can only discover(2) and apply(3) it by using your creativity, you may hold the rights to the real life processes that use it, but you can only have one algorithm for a certain output (given that the language of math is universal), then it's not recommendable.

Re:Not sure I understand this argument at all

[JasterBobaMereel]

You can copyright a book, you can copyright software

You can patent software (in the USA), the equivalent would be patenting a book, so no-one can write a book on the same subject

Re:Not sure I understand this argument at all

[Tim C]

Likewise, books are language. Can books be copyrighted? No one owns language.

You're right, comparing software to books is like comparing maths to language, and so I agree, software should be afforded all the same protections that books are.

Books are not patentable.

Re:Not sure I understand this argument at all

[Anonymous Coward]

IOW, software patents as they are now should be authomatically invalidated on "too broad claims" basis.

yes, abolish software patents.

[ciaran_o_riordan]

If software developers are prohibited from decoding your video format, the result is incompatibility.

Some people would indeed love to have a monopoly, and would spend millions for it, but these monopolies are simply not in the public interest. They're impeding software development and blocking interoperability.

There are more reasons here:
http://en.swpat.org/wiki/Why_abolish_software_patents

Re:Wheel was patented too recently...

[Anonymous Coward]

Giggity

Math?

[ThreeGigs]

Sorry, but I'm not familiar with "if" in math. What's the symbol for "if"? Or how about "next"? I'm sorry, but I don't see a pure mathematical formula in the article. I see programming code. I was really hoping to see just how something like the 120 patent looked in just math, but alas I don't see it.

Or am I missing something?

Re:Math?

[Anonymous Coward]

Piecewise functions are analogous to "if", I think.

Re:Math?

[Draek]

Yeah, your Lambda Calculus classes at your friendly neighbor University.

Re:Math?

[vlm]

Sorry, but I'm not familiar with "if" in math. What's the symbol for "if"?

http://en.wikipedia.org/wiki/Help:Displaying_a_formula#Continuation_and_cases

Re:Math?

[Anonymous Coward]

Sorry, but I'm not familiar with "if" in math. What's the symbol for "if"? Or how about "next"? I'm sorry, but I don't see a pure mathematical formula in the article. I see programming code. I was really hoping to see just how something like the 120 patent looked in just math, but alas I don't see it.

Or am I missing something?

You are missing something; it is important. There are several symbols for "if" in math: http://en.wikipedia.org/wiki/Material_conditional

Re:Math?

[SpiralSpirit]

there are lots of ifs in math and physics. They're just, you know, in your head. If you're calculating the axial deformation of a block of metal under load, you calculate the stress. IF the stress is greater than proportional limit, you have to use a stress-strain curve to find strain. If it's under the proportional limit, then you can use the load and modulus of elasticity to find it. but of course, if someone had made a program to do that hundreds of years ago, there would have been an entire generation of people who would have had to pay him not because he made up the math, but because his program meant he 'owned' the concept of the 'if', even though it was self evident to anyone who knows anything about it.

Re:Math?

[turkeyfish]

Yes, you are missing something. Concepts such as "if" imply that the range of outcomes is different over different elements of the domain of the universe of all possible inputs. If statements are in essence equivalent to distinguishing between outcomes or range within or outside of a given portion of the entire universe of possible outcomes. Hence, an if statement will be true or false depending upon whether the elements within the domain of elements implied (or defined) by the statement fall within or outside of its range. See the definition of Cartesian product for an understanding of concepts of relation and implication. Keep in mind that the relations and implications need not necessarily be restricted to numbers, but any potential elements of a set, so long as the set is not "too large" as to lead to inherent contradictions (ie the set of all sets that do not contain themselves).

Re:Math?

[turkeyfish]

Piecewise functions are a good example, but the concept can be applied to continuous functions as well. The issue is how the domain and range of the relations are defined.

Re:Math?

[Anonymous Coward]

Yes, you're missing something. Piecewise defined functions? Dirac delta and its derivatives and antiderivatives? Even in high-school algebra (which is apparently all you include as math), conditionals are not an issue.

Try writing 5 programs in your choice of functional language sometime. Maybe it'll help you realize the fundamental truth that computer science is mathematics.

Re:Math?

[turkeyfish]

Its typically presented as the implication symbol or rule in languages such as Mathematica (->), which can refer to a variety of functions, although is actually more general than the concept of function since they may be multivalued for a given element of the domain. You can think of it as a relation that will be true or false depending upon the elements that make up the input and their "relation" to the elements that make up the output. a -> b will be true if an element a or an element in a set a implies the element b or an element in the set b.

Re:Math?

[ThreeGigs]

But that's the point. There aren't any "pure" formulae in the article. It's all pseudocode. Show me the _pure_mathematical_function_ that is patent 120. If someone doesn't understand English, they should still be able to look at a math function and understand it. The article fails that test, and as such I can't call it a math function.

Re:Math?

[Dynetrekk]

I agree. You can write them as products with the Heaviside unit step function: http://en.wikipedia.org/wiki/Heaviside_step_function

Re:Math?

[mark-t]

Based on what is evidently your limited understanding of mathematics (you did not realize that iterations and conditions exist in math, for example), I would be inclined to speculate that if someone were to humor your request and show you a legitimate mathematical formula that was isomorphic to the patent, that you would not understand it, and proclaim it to be invalid presentation on that basis.

Re:Math?

[Anonymous Coward]

Your reply reminds me of something my father, a man born in the early quarter of the 1900s who never graduated past the 9th grade, said about algebra when he glanced at my homework when I was in high school (paraphrasing): "that's not math! Math is numbers, not letters! You can't add and subtract letters!"

http://en.wikipedia.org/wiki/Lambda_Calculus#Logic_and_predicates

Re:Math?

[beaker8000]

Define a function:

f(x) = 1, x > 0, and 0 else.

there is an 'if' in there but its too obvious so its usually not written. Note, 'for all' does this too (and the 'for all' notation is everywhere in analysis). Define a new function, domain is the real line:

f(x) = 1, for all x belonging to the rationals, and 0 else.

As to 'next', for integer n, the next is n+1. Pretty common (see mathematical induction).

As to the concept of 'next' for irrational and rational numbers: 'next' doesn't exist.

Re:Math?

[fatalGlory]

There is a video lecture I watched recently called "What does Linux prove" from an old linux.conf.au event. In this lecture, the speaker shows how to use lamda calculus to implement the C language from scratch - including implementing the "if (boolean)" functionality using only pure functions with parameters and return values.

It's not something immediately obvious to every high school kid, but conditionals apparently are available in pure mathematics.

Link to lecture: http://lca2007.linux.org.au/talk/215.html

Re:Math?

[Anonymous Coward]

Have you never heard of case analysis? 'if' is a function defined by case analysis, as is 'next'. Case analysis is otherwise known as pattern matching in functional languages. Look in Haskell's standard prelude.

All software is math, period. There is no question, no debate, except by those who understand neither software, nor math.

Re:Math?

[UncleFluffy]

there's a whole bunch of "if-like" things in mathematics ... multiplication over GF(2) for one such example. Haskell can be considered an alternate syntax for System F. In other words, that "programming code" that you see is isomorphic to a "pure mathematical formula" written in a particular variety of lambda calculus. It's just written in a format that is easier for programmers to digest.

Re:Math?

[Anonymous Coward]

uhh... lambda calculus?
Computers aren't some newfangled thing. We had worked out how to program them long before they even existed, just waited for engineering to catch up and frigging build a working version.
Programming was and is applied mathematics all the way.

Re:Math?

[Anonymous Coward]

No, he has a point.
Mathematics is supposed to be pure and unambiguously understandable by all mathematicians. If it includes words / expressions from some spoken language or symbols that are not well-defined, than it is not a pure mathematical formula.

Re:Math?

[Anonymous Coward]

omg, have you studied math at all?

Re:Math?

[Anonymous Coward]

in physics, the use of the theta function is customary. If the argument is >=0, the result is 1. Otherwise it's 0.
So 5*theta(x-3)+2 is:

  7 for x>=3

  2 otherwise

Re:Math?

[Anonymous Coward]

Or am I missing something?

Yes, your brain.

Re:Math?

[alendit]

This must be the finest piece of trolling on /. i've ever seen in a while. Playing string of a nerd's soul so virtuously, faking ignorance on a subject, which, while not common knowledge, is quite familiar to most people around here. And, of cause, everyone is compelled to show just how wrong you are and how much they know.

You, sir, are a genius!

Re:Math?

[osu-neko]

If someone doesn't understand English, they should still be able to look at a math function and understand it.

Only if they understand mathematics. Since you obviously don't (beyond some rudimentary basics, no doubt), this would not be the case for you.

Re:Math?

[ustolemyname]

So, until everybody understood what the integral symbol was calculus wasn't math?

Patent Office's new policy

[Anonymous Coward]

It will be interesting to see what happens with the Patent Office's new policy regarding computer implemented functions. They appear to have increased the written description requirement. Previously only means-plus-function limitations that claimed software were required to disclose the algorithm. However, the latest policy is to require algorithms for all claimed software limitations unless it is so generic that any computer is already programmed to do it. That will likely make it much more difficult to obtain a software patent.

The thing is

[trifish]

There is a very thin line between "a lot of mathematic formulas" and an algorithm. For example, the AES encryption algorithm can be written down as a very complex set of mathematical equations.

And it is, of course, patentable. Because, for the purposes of patents, there are no real differences between algorithm/method and technique/technology.

Re:The thing is

[SpiralSpirit]

the question is: 1) specific implementation - because with the current system you don't have to have one, you own the concept of it. 2) were you using 'old' formulas to do something overall 'new'. In this case, the suing company didn't write the specific implementation, and didn't do anything 'new'. they just copyrighted a concept that was self evident, and done before many many times. It's only the broken patent system that allows something like that to exist.

Re:Wheel was patented recently too...

[Anonymous Coward]

DO NOT CLICK ON THE ABOVE LINK.....there isn't enough eye bleach to unsee that.

This is absurd.

[Anonymous Coward]

I fail to see the difference between mathematics and logic. Then again, I also fail to see how intellectual property is "property".

Re:This is absurd.

[Sique]

Logic is only a part of Mathematics. Mathematical logic is not expressive enough to describe all of mathematics. Set theory for instance is not completely describable by mathematical logic.

I say put up a bounty

[makubesu]

on a program that automatically reduces patents to math. Ah wait that was already patented...

A really simple argument:

[LeRandy]

When you create software, what do you do?

Do you write software or do you make software?

I think that is the fundamental question that needs to be asked. If you make it, it should be patented. If you write it it should be copyrighted. The issue here is that the software industry (in part) wants to go double-dipping.

Re:A really simple argument:

[mark-t]

By the argument that if you make it, it should be patented, then sculptures and paintings should be patentable as well.

Re:A really simple argument:

[Anonymous Coward]

I write source code and sometimes I run make.

Re:A really simple argument:

[bunratty]

When you write software, the particular implementation in source code is copyrightable. You may implement an idea in the public domain as source code and copyright the source code, and I cannot copy your source code and sell it. I can implement the same idea in my own source code and copyright my source code.

The method that an algorithm uses may be patentable. If I patent an algorithm I can require that you pay me a licensing fee to write and sell software that uses the algorithm.

The real problem with software patents is not the fundamental idea of software patents, but that the patent office grants patents on ideas that are obvious to many programmers. For an idea to be patentable, it should not be obvious to someone skilled in that technical area.

Re:A really simple argument:

[SydShamino]

Sure. The first person who did something innovative with sculpture - like, say, creating a method and product where interlocking pieces of wood are created in sculpture by carving them out of a single piece - would be patentable. 3000 years ago.

Re:A really simple argument:

[mark-t]

No no no.... not a method behind making the sculpture, but rather the sculpture itself.

Re:A really simple argument:

[JonySuede]

I do both, I first write it then I build it !

Re:A really simple argument:

[Spewns]

For an idea to be patentable, it should not be obvious to someone skilled in that technical area.

Sounds like an incredibly shaky, arbitrary foundation upon which to build a system. I'd like to see someone come up with concrete, objective definitions of "obvious" and "technical area".

Re:A really simple argument:

[bunratty]

Many laws use terms that don't have concrete, objective definitions. That's why we have judges who interpret the laws.

Re:A really simple argument:

[JustinRLynn]

Who also often have no technical expertise in the field of the case in which they are supposed to be making decisions. One week a rubber tyre manufacturer, and the next a microcomputer chip manufacturer debating about the minutiae of their patents which are almost all >10 pages long and filled with hundreds upon hundreds of "whereas"s and conditions, I'd imagine it'd be enough to make any tenured lawyer pull out their hair.

Re:A really simple argument:

[indeterminator]

I think the point is you have to make the sculpture or painting to do something, or be usable for something to make it patentable.

Re:A really simple argument:

[osu-neko]

The real problem with software patents is not the fundamental idea of software patents...

Many of us here would disagree with that. Specifically...

The method that an algorithm uses may be patentable.

As long as this statement is true, they're something fundamentally wrong with the idea of software patents. The discovery of a new algorithm is precisely that: a scientific (specifically mathematics) discovery. Such a thing should, never, ever be patentable under any circumstances at all whatsoever, period.

Re:A really simple argument:

[mark-t]

The funny thing is that software doesn't really do anything either... What we perceive a computer to do is really just an illusion created by moving bits around and performing mathematical operations on them.

Re:A really simple argument:

[SydShamino]

Almost any mechanical device that has ever received a patent is a sculpture. How are they different?

Consider, for example, each of the many varieties of barbed wire that received a patent. (One of my older relatives had a board with five or six of them attached, each with its patent number underneath.) How is that not sculpture? Do you not consider it art because it has a purpose?

Re:A really simple argument:

[SydShamino]

Also, I covered that where I said "method AND product". In other words, by patenting both, the inventor would both cover the method behind the making of the sculpture, and the innovative part of the sculpture. While the patent was valid, no other artist would be able to make a statue with that feature.

Obviously the inventor doesn't get to patent the whole sculpture, unless every little aspect of the sculpture is innovative. I don't get to patent the car just because I invented a better harness system for seat belts. No patent system works like that.

Re:Wheel was patented recently too...

[jeremyp]

You accidentally linked to one of your home movies. Is that you on the bed? It certainly looks like a huge asshole.

Has anybody asked why....

[mark-t]

... the patent office has traditionally excluded mathematics from being patentable?

I mean, really, what is the reason behind it?

Because if, in fact, there is a legitimate answer to that question, then it *MUST* follow that any patent that can be shown to be isomorphic to a sequence of math operations should reasonably be rendered invalid on that basis, and no other.

Re:Has anybody asked why....

[Anonymous Coward]

Not strictly true, at least they are close to the boundary between math and software very frequently. I deal with this constantly in my field, numerical algorithms are patented all the time. Even though it's the algorithm that's patented, the algorithm itself was always first developed mathematically by its very nature. The problems I deal with aren't discrete, so one can't see just by looking at the algorithm itself that it works, a lot of math goes behind my problems and the final algorithm implementation is really only 10% of the battle.

Given all that though, they are still patented all the time. Oil companies deal with this a lot, and even though the final implementation is the thing that is patented, if they happen to come up with something mathematically equivalent to the patented algorithm they still can get nailed.

I think this is a relatively recent thing though, as numerics at this level has only recently become widespread with cheaper hardware.

Re:Has anybody asked why....

[mark-t]

If algorithms that are nothing more than mathematical operations should be patentable, then it must follow that the underlying reason behind why the patent office has traditionally not allowed mathematics to be patented is not a very good one.

Is that your assertion?

Equivalence relation?

[Anonymous Coward]

Just saying, if you're gonna make a claim about 'pure math,' at least use the phrase 'equivalence relation' correctly. The way it's stated above makes no sense. In the strictest sense, an equivalence relation is a set of ordered pairs that satisfies reflexivity, symmetry, and transitivity.

That's at least the strict set-theoretic way to define it, the more popular approach (but the same thing) is like this:
http://en.wikipedia.org/wiki/Equivalence_relation

So given two sets A and B, it makes no sense to say "There exists an equivalence relation between A and B." I challenge anybody to make that notion rigorous in a way that won't force it to apply to any two given sets.

Just sayin'.

But yeah, patent laws need to be revised to handle ideas better. With programming has come frozen ideas, that is an idea directly translated to a product, bypassing a physical implementation. Since the product comes so close to the idea, patents are now starting to come about that 'protect' abstract things such as concepts.. which I think is totally crazy.

I work in numerical PDEs, and I hear speakers all the time from oil companies who can't use a certain algorithm because the other oil company patented it. It's nuts, because usually the other solution is the numerical PDE equivalent of "1+1=2," it's the most obvious implementation that won't go haywire in floating-point arithmetic. What do the competing companies do? They hire people like me to come up with a different solution so that they can legally run it on their computers.

Re:Equivalence relation?

[maxwell demon]

So given two sets A and B, it makes no sense to say "There exists an equivalence relation between A and B."

Actually it makes sense to say that. Saying that is equivalent to saying "A and B have the same cardinality" because that's exactly how cardinality is defined.

Of course what the poster meant was that there's an isomorphism between them.

Re:Equivalence relation?

[koreaman]

Still doesn't make sense... maybe you meant "bijection", not "equivalence relation".

Equivalence relations are defined on a set, not between two sets.

patenting math

[Anonymous Coward]

is this in the bailout for greece?

Everything is true,

[rrayst]

if 5893120 == 5893210.

Re:Wheel was patented recently too...

[newcastlejon]

DO NOT CLICK ON THE ABOVE LINK.

No problem. I don't click links on /. that use URL shorteners.

This isn't Tweeter; we don't need shorteners. It would be nice if we could have Unicode though. Are you doing something about this, Taco?

WAIT A MINUTE

[Anonymous Coward]

You just did 1+1 which is addition, which I should go get a patent on. PAY UP NAOS.

Re:Wheel was patented too recently...

[maxwell demon]

I think I should get a patent on links to goatse pictures. I'm sure I could make billions from it. ;-)

That's not what the patent is for though.

[91degrees]

This is a patent for a method of information storage and retrieval. The algorithm can be used for other purposes, although most of them are pretty abstract without violating the patent.

Elliptical curve cryptography is patentable even though the algorithms are well known for describing elliptic curves. Unless there's a mathematical notation to explain what a secret is, then this new application is patentable.

Re:That's not what the patent is for though.

[maxwell demon]

Unless there's a mathematical notation to explain what a secret is,

No problem.

Be S a sigma algebra, and P a set of parties. Then we can define a "knowledge function" K:P->(S->[0,1]), which assigns to each party a probability function over S. We define an element of s of S to be be knowledge of a party p in P if K(p)(s)=1.

In addition we define another sigma algebra M of messages, on which for each element for S we have a conditional probability function, p(m|s). After receiving a message M, each client updates his knowledge function according to Bayes' rule to K(p)(s|m). A message m is said to reveal s to p if K(p)(s) Em\in M:Es\in S:(K(p)(s)1 && K(p)(s|m)=1 && K(q)(s)1 && K(q)(s|m)=1 && Er\in P: K(r)(s|m)1)

Re:That's not what the patent is for though.

[maxwell demon]

Damn, I should have read the preview (I had unencoded less-than signs). I'm now too lazy to retype all that's missing, but the final formula is:

S(p,q) <=> Em\in M:Es\in S:(K(p)(s)<1 && K(p)(s|m)=1 && K(q)(s)<1 && K(q)(s|m)=1 && Er\in P: K(r)(s|m)<1)

Where S(p,q) means p and q share a secret (I just now also noticed I used the S twice).

I recognize Patent #5,893,210

[Anonymous Coward]

Yeah, I recognize Patent #5,893,210 as being different from patent Patent #5,893,120. If you are going to throw a number around at least copy/paste it to not mistype it.

So?

[Kaz Kylheku]

RSA crypto can be reduced to pure math. JPEG, MPEG 2 layer 3 audio encoding, ditto, etc.

Same patent?

[turkeyfish]

Are we talking about the same patent or is there a misprint here? The two numbers are not the same.

Obviousness

[Anonymous Coward]

It seems to me, after reading the patent, that there should be ways to prove that the algorithm is an obvious solution to the problem at hand, prior art or not. Just put 5 CS students in a room and ask them how to solve the problem he patent purports to solve, and observe that at least 2 or 3 of them will come up with the same solution as the patent.

Heck, the whole thing actually sounds like a good interview question for a junior programming position.

Not actually reduced to math

[Grond]

I am an attorney & patent agent and I hold multiple degrees in mathematics and computer science, so I feel fairly competent to speak on this issue.

The claims require a physical implementation (e.g. "An information storage and retrieval system"). No amount of math will produce such a system out of the aether. Nor does thinking about the math or the formulas on pen & paper infringe the patent. The specification makes it clear that "information storage and retrieval system" refers to a computer system. Thus, the patent was not actually reduced to pure mathematics.

The nonpatentability of mathematics refers literally to patenting a formula or algorithm without any useful application, just as chemical elements cannot be patented but a mechanical device made entirely of a single element could be. A claim to a bare formula would be invalid for lack of utility as well as lack of patentable subject matter. Consequently, the Curry-Howard Correspondence has no effect on the patentability of computer-implemented inventions.

Re:Not actually reduced to math

[Anonymous Coward]

That's a contradiction in terms. Either only a specific implementation is patented, or if all implementations are patented that means that the formula is patented too, whether you name it as such or not.

Re:Not actually reduced to math

[maxwell demon]

But according to that argument, if I write a program that does this and distribute that program, I should not be violating the patent, because I'm not distributing "an information and storage retrieval system", right?

Re:Not actually reduced to math

[bieber]

That entire argument is pure semantics, however. The patent holder has not invented some novel new device for information storage and retrieval which physically implements the algorithm, they've simply come up with a set of instructions for existing machines which allows them to perform that algorithm. To make a less than perfect analogy, a waiter, if given instructions to wash dishes for the evening, does not in fact become a new person who is a dishwasher rather than a waiter. They're just doing something different than they did before.

The fact that the patent office considers this an "invention" is an absurd technicality. Maybe you can technically twist the letter of the law to fit the idea of a patent on a software technique, but at that point you've completely abandoned the spirit of the law. While the judicial system is busy tangling over fine points of language, America's technological industries are going to hell in a hand basket. Unless you're a massive multinational corporation, good luck trying to implement any useful software without using at least a couple dozen "novel" computational techniques from the last two decades...

Re:Not actually reduced to math

[devent]

Maybe you right but than the patent should not be granted. "An information storage and retrieval system" is just too broad. That covers everything from a computer to the human brain to paper and pen. It's like a patent on an automobile which just needs "a horseless power that can move a carriage". Or a television with just needs "a screen that can display images".

So patent claim should be specific to one device. You can't just say "something+human interface device" and claim a patent for it. That's just my logic as a normal citizen, I'm sure the law about that is twisted enough to employ all the lawyers.

Re:Not actually reduced to math

[Anonymous Coward]

Well, that all assumes, that concepts like "owning" information would be something physically possible in this universe. ^^

The problem is, that you were trained to view it like you view it, but did not actually come up with those views yourself. So you don't know why and if they are true.
I went that hard way. All the way to quantum physics.

And information is just the structural property of matter/energy. Not matter/energy itself. (That is only the medium.)
While "owning" is having control over something. Because if you lose control over it, you can yell as much as you like that you "own" it... you still don't anymore. Here "losing control" includes not having other people (e.g. a government) that can dominate the one controlling it now.

But how can you "own" information? Ok, let's see: I know something. Now of course I could simply not pass it on.
As I said: I though about this. And in this case... you could not prove the existence of that information anymore. Since you could only do so, by passing it on. (Even for parts of it, you would have to pass them on.)
Passing on here means: Making a *copy*. So there is no loss, and hence no harm involved. (Other than willfully giving up control.)

But if you had passed it on... you would have lost control over it. Forever. You could only try to oppress people into not telling anyone. But if they did, and the third-parties would not show they know, you could not prove they passed it on too. Which means you could *still* not retain control over the information, even with oppression.

And that means, that any information that can be proven to exist, can not be owned. QED.

Re:Not actually reduced to math

[Dragonslicer]

Since you read (or at least skimmed) the specification, did you find anything corresponding to the means-plus-function elements in the claim? I'm curious if the reduction to math took those into account, and I don't feel like reading the specification.

Re:Not actually reduced to math

[Anonymous Coward]

Interesting. So my layman's take on this is

"This sequence of operations performed on this domain of inputs has these properties" is not patentable, since it is pure math; whereas

"This sequence of operations performed on this domain of inputs using a computer" is (potentially) patentable.

This raises the question, does that mean a paper describing an algorithm is not prior art, since it didn't involve a computer?

Re:Not actually reduced to math

[Co0Ps]

When discussing the issue of software patents, one always refers to "computer system" implementations. Since the domain has already been selected it is perfectly acceptable to claim that the basis of a software patent can be reduced to pure mathematics without refuting your argument.

Re:Not actually reduced to math

[Anonymous Coward]

Yep, there's a problem here because a non-novel interpreter plus unpatentable maths as input -always "used together" in the same way, without any novelty- now give you something that is apparently patentable as a whole. So how is the requirement for novelty in a patent EVER met?

I'm sure there's some legal wording and court decisions that make this entirely possible, but it goes against other wording and the intent/spirit of things as we want them to be, as society. (I'd say "would want" but I'm not US American).
 

Re:Not actually reduced to math

[Anonymous Coward]

No, the claim is not "An information storage and retrieval system" -- it's "An information storage and retrieval system with computer-readable data that implements {insert algorithm here}." Taken as a whole, and assuming nobody's previously implemented that algorithm, it's fine for novelty and breadth, and may be fine for obviousness (if it was a truly clever algorithm).

Now why in hell "$algorithm" is unpatentable and "$boilerplate_storage_device_crap $algorithm" is patentable doesn't make much sense, but AFAICT the law is pretty damn clear on both sides -- which is why all software patents are for a hard-drive or computer running that software. It's a loophole that should be fixed, IMO by restricting the novelty requirements to only consider such part of the claimed invention as is itself patentable. Then the unpatentable mathematics would be removed from consideration, and what's left ("An information storage and retrieval system storing information") would indeed be non-novel and unpatentable.

Re:Not actually reduced to math

[Anonymous Coward]

> Nor does thinking about the math or the formulas on pen & paper infringe the patent.

Does the use of a mechanical calculation device infringe on the patent?
Does the use of an electronic calculator device infringe on the patent?
Does the use of an electronic calculator with memory infringe on the patent?

Your argument doesn't make sense. "multiple degrees in mathematics"? whatever.

Re:Not actually reduced to math

[stinerman]

Close. There are several different hairs to split here:

If you write a program that implements the patent, you have simply transformed an English description of the patent into an [insert programming language of choice here] description of the patent. That, in and of itself, is not illegal any more than translating the patent to Spanish, French, or Esperanto would be.

If you compile the program (assuming you used a language that isn't interpreted) and distribute it, you've not done any more than you have in the previous example. In this case you're just translating the code into assembly language. However at this point, you'll usually get sued for infringing on the patient even though, as the OP states, the patient is only valid when the algorithm is carried out using a generic programmable computer.

If the OP is correct, you would only be infringing at the point that the program is loaded into memory and executed by the processor. If that's the case, the onus would have to be on the plaintiff to prove that at some point I ran that particular algorithm.

I use the Linux kernel. It might be such that the code that implements the patented algorithm has never actually executed on my system. If that is true, I haven't infringed in any way. I'm sure there are binaries on my hard drive that I've never used. If any of them implements a patented algorithm, I haven't infringed.

Re:Not actually reduced to math

[Draek]

Parent post, in short: The "on a computer" suffix is still enough to make anything patentable.

You know, though? I'm tempted to grab the very same patent described in TFA, implement it on Conway's Game of Life then post the result online for the world to see just to see whether some scumbag lawyer tries to take it down for infringement.

Re:Not actually reduced to math

[martin-boundary]

The word computer used to mean people, then it meant machines with vacuum tubes, now it means chips of silicon. There's no telling what it could mean tomorrow. So allowing patents with the "on a computer" loophole is like writing a blank check.

Re:Not actually reduced to math

[Logger]

This fact frustrates me to no end. It sure seems like executing an algorithm "on a computer" or "over the internet", in the eyes of the patent office, makes something patentable. Maybe 30 years ago that was true, but today we think differently. Today we try to do everything on the a computer and over the internet. A particular implementation should be patentable, but the blanket patents being issued are too broad. Does the Amazon one click patent specify an exact configuration of databases, network protocols, network traffic, and server software? If not, it sounds to me to be the same as an old school credit account at a store. I often think of it being like the digital version of an open tab at the bar. An open tab "on the internet" does not sound patentable to me.

Re:Not actually reduced to math

[cdecoro]

Technically, what you're saying is true. But the requirement of "a physical implementation" is little more than a magic phrase necessary to recite to achieve patentability. Take for example, the following IBM patent on arithmatic coding: http://www.google.com/patents?vid=4122440. The claims recite over and over "an apparatus" to do X, an "apparatus" to do Y... etc... Of course, apparatus is not defined; they give some diagrams, but presumably a software implementation on a general purpose CPU is also an "apparatus." So at this point, the idea that "a physical implementation" is being claimed is meaningless.

Also, they do not make any claim that it has a useful purpose. The original -- to my knowledge only -- Supreme Court case directly on point to software patents upheld the patentability of an automated rubber-curing and -molding system, of which software to compute the Arrhenius equation was a part. Diamond v. Diehr, 450 U.S. 175 (1981). In that case, there was a real physical input transformed into a physical output. The claims state specifically "a rubber-molding system." Therefore, one could use the same basic software to cure or mold something else, and it would not infringe. By contrast, that IBM patent does not have an actual utility for a specific application; it covers anything that could ever use arithmetic coding.

Re:Not actually reduced to math

[Anonymous Coward]

How about paper and pencil instead of a computer? The only difference is speed.

Re:Not actually reduced to math

[Anonymous Coward]

>>just as chemical elements cannot be patented but a mechanical device made entirely of a single element could be.

a 4 molecule chain ? a single molecule bearing?

Re:Not actually reduced to math

[Anonymous Coward]

Can the same algorithm be used in another implementation (for a different purpose) and not violate the original patent?

Re:Not actually reduced to math

[Anonymous Coward]

so, if i take some math, and transform it into control flow on some completely generic computing device,
i have exclusive rights over any such translation

sorry for doubting the process

Re:Not actually reduced to math

[Theaetetus]

But according to that argument, if I write a program that does this and distribute that program, I should not be violating the patent, because I'm not distributing "an information and storage retrieval system", right?

Sort of... You're not directly infringing the patent, since you're not selling an information and storage retrieval system... But you are indirectly infringing, since you're selling something (that program) that can only be used with a purchaser's information and storage retrieval system to create the system recited by the patent claims. The customer is infringing - you sold them the program - so therefore you induced them to infringe, essentially.

It's like how pharma companies claim methods of administering drugs, but don't go after doctors. Who wants to sue a doctor? You'd look like an asshole. Instead, you sell the company that made the discount knock off drugs who, through the doctor, indirectly infringe.

Re:Not actually reduced to math

[Directrix1]

I could literally write out a list of formulas that implement any algorithm and directly run them. Are you saying that these mathematical formulas do not infringe patents? Are you saying that patent law requires a selling a physical machine with this on them to fall under the definition? Either way, this is fucking bullshit. To allow the patenting of mathematics (or the set of all system that happens to utilize specific mathematics) is idiotic and counterproductive. All it does is take time and money away from the people that do and gives it to the people that sue. It disregards all the freely available prerequisite knowledge (the giants whose shoulders we all stand on), and packages up these incremental steps for pure, exclusive greed.

Implementation issues in the written world are more than sufficiently covered by copyright law. I have yet to see a single example of patents being a positive force in this industry.

Re:Not actually reduced to math

[donscarletti]

I am an attorney & patent agent and I hold multiple degrees in mathematics and computer science, so I feel fairly competent to speak on this issue.

This is slashdot, "I'm some random arsehole with a strong opinion" qualifies you just as much, if not more.

Re:Not actually reduced to math

[Anonymous Coward]

> "An information storage and retrieval system"

Which could describe a calculator just as well as a computer. Given that the computer in question was invented by someone else, the only advance contained in this patent is pure math.

Re:Not actually reduced to math

[DamnStupidElf]

The claims require a physical implementation (e.g. "An information storage and retrieval system"). No amount of math will produce such a system out of the aether. Nor does thinking about the math or the formulas on pen & paper infringe the patent. The specification makes it clear that "information storage and retrieval system" refers to a computer system. Thus, the patent was not actually reduced to pure mathematics.

Here's a useful test to see if "an information storage and retrieval system" (ISRS) is bullshit or not. If there exists an isomorphism between every element in the claim and a Turing Machine (TM) then the patent is pure bullshit. I see nothing in any of the 8 claims that makes a reference to any physical system that does not have a very simple mapping to a TM. Just define a bijection between "random access memory" and "disk storage unit" in the ISRS and the tape of the TM, and a bijection between the other elements of the claim and the state machine of the TM. If there's only a homomorphism because the set of possible TMs is larger than the set of possible ISRS in the patent, then it's double bullshit.

Re:Not actually reduced to math

[Anonymous Coward]

But all computer are is machines capable of running mathematical algorithms. "On a computer" makes it patentable? What about in a persons head? If they are doing a mathematical function in their head (arguable not that much different from a computer), are they in violation of patent infringement? The problem is that there are many patents which are extremely obvious, and not new ideas, yet patents are granted. The law must be changed so that once an algorithm is published in a book or online, its no longer patentable, as there is prior art. Books are published to distribute knowledge, and patents seem to be in place merely to either prevent their application, or suppress the idea. In that regard, patents are destructive to society and must either change or be abolished. The greed of a few has set itself on the destruction of public knowledge. This has to end.

Re:Not actually reduced to math

[Kirth]

How can you actually justify redeclaring a bookshelf as a Moby Dick Support Device? How can you really belief that argument?

Re:Not actually reduced to math

[Theaetetus]

Technically, what you're saying is true. But the requirement of "a physical implementation" is little more than a magic phrase necessary to recite to achieve patentability. Take for example, the following IBM patent on arithmatic coding: http://www.google.com/patents?vid=4122440. The claims recite over and over "an apparatus" to do X, an "apparatus" to do Y... etc... Of course, apparatus is not defined; they give some diagrams, but presumably a software implementation on a general purpose CPU is also an "apparatus." So at this point, the idea that "a physical implementation" is being claimed is meaningless.

You're confusing "broad" with "meaningless". Yes, any general purpose CPU can be used to fulfill the claims... but that CPU is required to fulfill the claims. Perform them in your head or on paper, and you're not infringing. And that is a legitimate limitation, since they're no longer claiming "just math".

Also, they do not make any claim that it has a useful purpose. The original -- to my knowledge only -- Supreme Court case directly on point to software patents upheld the patentability of an automated rubber-curing and -molding system, of which software to compute the Arrhenius equation was a part. Diamond v. Diehr, 450 U.S. 175 (1981). In that case, there was a real physical input transformed into a physical output. The claims state specifically "a rubber-molding system." Therefore, one could use the same basic software to cure or mold something else, and it would not infringe. By contrast, that IBM patent does not have an actual utility for a specific application; it covers anything that could ever use arithmetic coding.

And nothing that could ever use arithmetic coding is useful? Or is your concern that this may have use in multiple applications?

Re:Not actually reduced to math

[Anonymous Coward]

The claims require a physical implementation (e.g. "An information storage and retrieval system"). No amount of math will produce such a system out of the aether. Nor does thinking about the math or the formulas on pen & paper infringe the patent. The specification makes it clear that "information storage and retrieval system" refers to a computer system.

That wouldn't be a problem if computers would be always built to do a single task, like in the 30s. A problematic patent is a patent talking about all possible computer implementations and not a specific implementation in a specific context. Such a general patent would be a patent for a pattern of configurations having no useful application without a specific physical implementation that is useful. Such a pattern can be described, designed and implemented as a set of mathematical formulations or equivalent language, becoming useful only as an input to a system implementing a specific collection of physical objects or computer instructions and data. In other words, the patent system does not properly take into consideration objects that can be constructed with words instead of physical materials.

Re:Not actually reduced to math

[Anonymous Coward]

Isn't the "An information storage and retrieval system" can be purely virtual e.g. program running in a virtual machine.
Every algorithm can be explained with real-world objects (stones and sticks for instance) and played like people replay chess parties.

Re:Not actually reduced to math

[iive]

"The nonpatentability of mathematics refers literally to patenting a formula or algorithm without any useful application"
I think the problem here is quite the opposite, the problem is patenting of formula or algorithm for all possible useful applications. And my claim is that if you patent for all possible useful applications, then you haven't actually patented for any of them (because you have not mentioned any).

Take for the example the patent in questions. "The specification makes it clear that 'information storage and retrieval system' refers to a computer system". The problem here is that the computer systems already exist, they existed and they did operate long before this patent was filled and they would continue doing so. The computers do store and retrieve information and they've been doing it for decades in a million of different ways. This patent describes algortitm for working with _abstract_ data. Thus it does not describe any useful application of that abstract data.
How, if the patent said "efficent malloc implementation allowing storage and tertival of heap memory in RAM" or "Efficent garbidge collection in programming language" then these would have been usefual applications, and that patent would not apply for "Efficent file retrival of Flash drives".

So in short, if you patent algorithm for working with abstract data, you patent pure math.

Re:Not actually reduced to math

[Grond]

The courts took a while to come around to it, but at this point they agree with you. Simply slapping "on a computer" or "on the internet" onto a known algorithm, business method, etc is likely to be considered obvious. For example: "Because Internet and Web browser technologies had become commonplace for communicating and displaying information, it would have been obvious to adapt existing processes to incorporate them for those functions." Muniauction, Inc. v. Thomson Corp., 532 F.3d 1318 (Fed. Cir. 2008). Similarly: "Accommodating a prior art mechanical device that accomplishes that goal to modern electronics would have been reasonably obvious to one of ordinary skill in designing children's learning devices. Applying modern electronics to older mechanical devices has been commonplace in recent years." Leapfrog Enterprises, Inc. v. Fisher-Price, Inc., 485 F. 3d 1157 (Fed.Cir. 2007).

Re:Not actually reduced to math

[Grond]

Are you saying that patent law requires a selling a physical machine with this on them to fall under the definition?

The Bilski case didn't precisely spell out the test(s) for patentable subject matter in this area, but the machine-or-transformation test remains one approach. So yes, that is basically one way to satisfy the patentable subject matter requirement. Note that it does not necessarily satisfy the obviousness requirement, which is a completely different issue. And in fact slapping a known algorithm on a computer is a quick way to get an obvious rejection these days.

Patentable subject matter is an early, coarse filter. Just because something can pass that filter does not mean it won't be invalidated because of anticipation, obviousness, lack of enablement, failure of the written description requirement, or any number of other things. I'm not arguing that this particular patent is or is not valid. I'm simply explaining that it was not, in fact, reduced to math, and that software patents in general are not subject to that attack. Whether that should or should not be the case is another matter, and I'm not taking a position on that here.

Re:Not actually reduced to math

[Theaetetus]

That entire argument is pure semantics, however. The patent holder has not invented some novel new device for information storage and retrieval which physically implements the algorithm, they've simply come up with a set of instructions for existing machines which allows them to perform that algorithm. To make a less than perfect analogy, a waiter, if given instructions to wash dishes for the evening, does not in fact become a new person who is a dishwasher rather than a waiter. They're just doing something different than they did before.

If I invent a brand new type of transmission, and stick in an existing car, should I be barred from patenting a car using my new type of transmission? And if so, why?

The fact that the patent office considers this an "invention" is an absurd technicality. Maybe you can technically twist the letter of the law to fit the idea of a patent on a software technique, but at that point you've completely abandoned the spirit of the law.

Not according to the people who wrote and interpret the law.

While the judicial system is busy tangling over fine points of language, America's technological industries are going to hell in a hand basket.

Yes, I expect to hear Chapter 11 reports from Apple and Google any day now.

Re:Not actually reduced to math

[Theaetetus]

When discussing the issue of software patents, one always refers to "computer system" implementations. Since the domain has already been selected it is perfectly acceptable to claim that the basis of a software patent can be reduced to pure mathematics without refuting your argument.

But it can't, because the claim recites those elements. You can't just ignore limitations in a claim because your theory doesn't work with them there.

Re:Not actually reduced to math

[Grond]

If there exists an isomorphism between every element in the claim and a Turing Machine (TM) then the patent is pure bullshit

Why? All you've done is put a label on something. You haven't actually given an argument or a rationale.

Re:Not actually reduced to math

[Grond]

And my claim is that if you patent for all possible useful applications, then you haven't actually patented for any of them (because you have not mentioned any).

Actually, the patent mentions specific uses (e.g. OS-level garbage collection), although the claims are indeed very broad.

So in short, if you patent algorithm for working with abstract data, you patent pure math.

In what way are a physical CPU, a physical disk storage unit, and all the other physical components of an information storage and retrieval system pure math? Pure math cannot conjure a physical computer out of the aether. These things are simply not equivalent. As such, working with the algorithm itself at the mathematical level (e.g. by thinking about it or reasoning about it in the abstract on pencil & paper) would not infringe the patent. This is not a patent on pure math.

Re:Not actually reduced to math

[Grond]

Parent post, in short: The "on a computer" suffix is still enough to make anything patentable.

Not at all. My point was that "on a computer" means that it isn't a patent on pure math. As I mention elsewhere in the thread, multiple cases have made it clear that slapping "on a computer" or "on the internet" on a known invention is a good way to get the patent rejected or invalidated as obvious. This patent may well be invalid for obviousness. But it was not reduced to pure math because its claims require multiple physical components.

Re:Not actually reduced to math

[Grond]

When discussing the issue of software patents, one always refers to "computer system" implementations. Since the domain has already been selected it is perfectly acceptable to claim that the basis of a software patent can be reduced to pure mathematics without refuting your argument.

I disagree. If I take the pseudocode from the patent, copy it, think about it, analyze it, or modify it, none of that will infringe the patent because the patent requires a physical computer system. I can do pure, abstract mathematics about the algorithm all day with impunity. So in what way has the patent been reduced to pure mathematics if pure mathematics cannot infringe the patent?

Re:Not actually reduced to math

[Grond]

For starters there's the pseudocode implementation in the appendix and the flowcharts in the drawings. Those describe the record search means and list accessing means using a linear search through the linked list and accessing the records through pointers. The specification itself discusses the use of a search key. That's probably adequate structure to support the means-plus-function elements. But I'm not taking a position on the validity of the patent. I'm only taking issue with the claims that a) the patent was reduced to pure math (it clearly wasn't because it claims physical components) and b) software patents in general are vulnerable to this kind of attack (they generally aren't because they usually claim some kind of physical component).

Various software patents may be invalid for other reasons, but this isn't actually a novel or effective attack.

Re:Not actually reduced to math

[Anonymous Coward]

The patent also says that the "information storage and retrieval system", a hash table in this case, is described in Knuth. All that the patent adds is to remove expired data from the table when you search the list in normal operation, instead of doing a separate garbage collection operation. And that is a pretty simple idea that should not be worthy of a patent.

Re:Not actually reduced to math

[Co0Ps]

You can use "pure mathematics" in many ways. You describe some of them: copy it, think about it, analyze it. Another way to use it is to implement it on a computer system. "Pure math" has now infringed the patent. Q.E.D.

Re:Not actually reduced to math

[Co0Ps]

The claim recites a "implementation on a computer system". All software falls in this category so it's not a limiting claim. Also, even if it would, a "computer system" could also be described with pure math. All ideas are really discoveries in ways nature can be utilized. This is the paradox of patents. I disagree with the claim that patents in general would be beneficial to society (and many socioeconomic researches does this as well, there's no consensus on this matter) because process of invention more resembles iterative discovery than revolutionary insight although the invention could have revolutionary consequences. It's an misconception which resembles the absurd cartoon trope where the character suddenly has a light bulb which pops out of his head. But as many other systems in society, if enough powerful people benefit from it, and it's complicated enough so you won't see the defects of it, there won't be any general motivation to change it.

Re:Not actually reduced to math

[Theaetetus]

The claim recites a "implementation on a computer system". All software falls in this category so it's not a limiting claim. Also, even if it would, a "computer system" could also be *described* with pure math.

A car can also be *described* with pure math - see a CAD program, for example. Does that mean cars aren't patentable? No. Merely being capable of being *described* mathematically doesn't mean that it is purely math. A computer system is not pure math, and is therefore not barred from patentability.

I disagree with the claim that patents in general would be beneficial to society

And that's essentially the key. Most anti-software patent people are really anti-all patents. But they've come up with this ingenious argument to try to equate software to pure algorithms and thus argue all software is unpatentable, no matter how nonobvious or how tightly tied to a machine. No judge has ever bought the argument, because it's simply legally wrong. And no judge buys the root argument - all patents stifle innovation - because Congress, the constitution, and 220 years of innovation in this country disagree.

Re:Not actually reduced to math

[Co0Ps]

I disagree with the claim that patents in general would be beneficial to society

And that's essentially the key.

That I disagree with you so you can place me in the "people-who-are-wrong-because-they-don't-agree-with-me-category"?

But they've come up with this ingenious argument to try to equate software to pure algorithms and thus argue all software is unpatentable, no matter how nonobvious or how tightly tied to a machine.

Yes, that's the basic idea.

No judge has ever bought the argument, because it's simply legally wrong.

A judge doesn't care about what's beneficial to society or not. A judge interprets the law and judges by it. If you're defining ultimate truth as the law of the state you might make a good lawyer but not a very good philosopher.

Congress, the constitution, and 220 years of innovation in this country disagree.

I'd recommend arguments which are not ad populum fallacies if you want to have an interesting discussion.

Re:Not actually reduced to math

[Theaetetus]

I disagree with the claim that patents in general would be beneficial to society

And that's essentially the key.

That I disagree with you so you can place me in the "people-who-are-wrong-because-they-don't-agree-with-me-category"?

No, the key to your position, which is why you'll dismiss as trivial any part that disagrees, such as the fact that neither a computer system or a car can be reduced to pure math.

But they've come up with this ingenious argument to try to equate software to pure algorithms and thus argue all software is unpatentable, no matter how nonobvious or how tightly tied to a machine.

Yes, that's the basic idea.

No judge has ever bought the argument, because it's simply legally wrong.

A judge doesn't care about what's beneficial to society or not.

Yes, that's the basic idea. That's the legislature's job in writing the law. Judges aren't supposed to disregard laws because someone believes they're not beneficial.
The judge simply cannot decide that a new, useful, and nonobvious piece of software that's tied to machine is unpatentable, even if they believe that all patents are harmful to society. If they want to do that, they can step down and run for congress.

A judge interprets the law and judges by it. If you're defining ultimate truth as the law of the state you might make a good lawyer but not a very good philosopher.

No, I'm defining the law of the state as the law of the state, by which the judge is bound. Although here, we're talking federal law. Nonetheless, a judge doesn't get to play a philosopher king and disregard the law in the pursuit of some unrelated policy argument.

Congress, the constitution, and 220 years of innovation in this country disagree.

I'd recommend arguments which are not ad populum fallacies if you want to have an interesting discussion.

I'd recommend not confusing a question of the burden of proof in making the extraordinary claim that patents stifle innovation as an argumentum ad populum, if you want to have an interesting discussion. I can claim that I'm immortal, and it's not a logical fallacy for you to question that claim with a high degree of skepticism based on the fact that no one in history has ever been immortal.

Re:Not actually reduced to math

[Co0Ps]

No, the key to your position, which is why you'll dismiss as trivial any part that disagrees, such as the fact that neither a computer system or a car can be reduced to pure math.

Sure, you can't drive math but you can implement it as software on a computer system, so your analogy is broken.

Yes, that's the basic idea. That's the legislature's job in writing the law. Judges aren't supposed to disregard laws because someone believes they're not beneficial. The judge simply cannot decide that a new, useful, and nonobvious piece of software that's tied to machine is unpatentable, even if they believe that all patents are harmful to society. If they want to do that, they can step down and run for congress.

A judge interprets the law and judges by it. If you're defining ultimate truth as the law of the state you might make a good lawyer but not a very good philosopher.

No, I'm defining the law of the state as the law of the state, by which the judge is bound. Although here, we're talking federal law. Nonetheless, a judge doesn't get to play a philosopher king and disregard the law in the pursuit of some unrelated policy argument.

I have not suggested this. I was criticizing your position that "patents are beneficial to society" because "no judge has ever bought that argument". I'm sure you can understand that "what's beneficial to society and not" and "the interpretation of law" is two unrelated subjects.

I'd recommend arguments which are not ad populum fallacies if you want to have an interesting discussion.

I'd recommend not confusing a question of the burden of proof in making the extraordinary claim that patents stifle innovation as an argumentum ad populum, if you want to have an interesting discussion. I can claim that I'm immortal, and it's not a logical fallacy for you to question that claim with a high degree of skepticism based on the fact that no one in history has ever been immortal.

I have presented some arguments on the issue. You can also read more criticism of patents here if you're interested. It's hardly a claim as extraordinary as "being immortal". Also you don't understand correctly why "ad populum" is a fallacy. The example you gave is not an "ad populum" argument... in that example the people are data points and you correctly come to the conclusion that since none of the countless data points is immortal, It's unlikley that anyone would be. However had the argument been "most people (in history) don't believe immortality is possible" or "it's an established fact in society that immortality is impossible" it would have fit in that category of fallacious arguments.

Re:Not actually reduced to math

[Anonymous Coward]

No amount of math will produce such a system out of the aether. Nor does thinking about the math or the formulas on pen & paper infringe the patent.

But the entirely algorithm could be carried out with pencil and paper by a person should pen and paper information be given to him. It would likely not be beneficial, but it is certainly possible. The symbols contained in magnetic and electrical media are in no way fundamentally different from those contained by ink and paper.

Re:Not actually reduced to math

[JasterBobaMereel]

If I invent a brand new type of transmission, and stick in an existing car, should I be barred from patenting a car using my new type of transmission? And if so, why?

You can patent the transmission, not the car (you invented the transmission, you did not invent the car)

This is like more like patenting a new way of driving existing cars, with no modifications to the car and no new parts required, then suing anyone who tries it (patent) ... not just publishing a book on it and suing anyone who tries to re-publish it without paying you ...(copyright)

Re:Not actually reduced to math

[Theaetetus]

You can patent the transmission, not the car (you invented the transmission, you did not invent the car)

I invented a car that includes that transmission. "A car including that transmission" never existed previously.

There is no requirement that every element of the claims be novel and nonobvious, as long as the entire invention is novel and nonobvious, as a whole. The patent expressly allows for this - patentable subject matter includes improvements on known methods and machines. To improve a machine, you kinda need the machine first.

This doesn't mean I can build a car with my transmission. Patents are not licenses to build. However, I can block others from building a car with my transmission, unless they want to purchase a license from me.

This is like more like patenting a new way of driving existing cars, with no modifications to the car and no new parts required, then suing anyone who tries it (patent) ... not just publishing a book on it and suing anyone who tries to re-publish it without paying you ...(copyright)

Yes, but we're not talking about copyright. If you come up with a new method, there's no requirement that you also have to make a new machine. Machines and methods are independently patentable subject matter, under 35 U.S.C. 101.

Re:Not actually reduced to math

[Anonymous Coward]

Right. The method to convert a mathematical formula into an implementation is a well known and simple process for most of the mathematical formula. Unless this is something really new that people don't have a way to convert yet, or convert it in a poor way; and this methods converts it in a much better way, or much worse way, but novelty, then it's patentable.
The problem here is basically the US turns blind eyes to this.

Re:Not actually reduced to math

[JasterBobaMereel]

In the first case you have invented something physical (the transmission), there is no need for you to actually build it or sell it (or even for it to work), and you can tie it to existing technology (car), in which case the same transmission not in a car may not be covered, but you have invented a physical device (the transmission)

In the second case yes you can patent a method (often a business method), that has no hardware at all, but only in the US, and in most cases patents of this type are often thrown out unless they are tied to a specific implementation, this is why software patents always mention the machine it is run on ...

Re:Not actually reduced to math

[Theaetetus]

In the first case you have invented something physical (the transmission), there is no need for you to actually build it or sell it (or even for it to work), and you can tie it to existing technology (car), in which case the same transmission not in a car may not be covered, but you have invented a physical device (the transmission)

Yes, but you can claim both the transmission, and "A car including my transmission." Why would you want to do this? Because car manufacturers may be buying your transmission from overseas manufacturers, and claiming patent exhaustion to avoid royalties.

In the second case yes you can patent a method (often a business method), that has no hardware at all, but only in the US, and in most cases patents of this type are often thrown out unless they are tied to a specific implementation, this is why software patents always mention the machine it is run on ...

Yes, but they don't require a new machine. Which is what I said before. You haven't addressed that point - if you come with a new method or a business method, there's no requirement that you also have to make a new machine, which is what you were claiming earlier.

Also, no need to tell me how software patents work. I write and prosecute them.

If you outlaw patents for math ...

[140Mandak262Jamuna]

If you outlaw patents for math only outlaws will have patents for math. Oh, wait... it did not not come out right. OMG! The outlaws are already having patents for math!

Patent Law Explained

[xkr]

All of patent law deals with interpretations, most of which are involve varying degrees of subtly.

The Federal Circuit Court has provided a great deal of well-written guidance. This particularly applies to what is and is not patentable.

The issue of what is and is not patentable is not black and white, such as, “mathematical formulas are not patentable,” or “software is patentable.”

A process that creates something useful and tangible is patentable, whether or not that process involves a calculation. What is not patentable is a “pure” formula that is not tied to something tangible. Data structures are tricky. The newer rules (yes, lots of mistakes were made in the past) are that generalized data structures, such as a table or a linked list, do not count as “tangible.” However, if those data structures are used (critically) to perform useful work, such as to refine steel or to serve up ads on websites, then the ENTIRE process is patentable. Subject, of course to all the other restrictions, such as non-obviousness.

These rules are not really new. They are the same rules that apply to mechanical inventions. For example, you cannot patent a “law of nature,” even it is something complex and nobody else knew about it. You can, however, patent a new device that takes advantage of this law of nature. For example, you cannot patent super-conductivity, but you can patent a useful device that uses super-conductivity.

Even mechanical inventions could be reduced to equations. CAD systems and hardware description languages are such examples. However, these “mathematical” representations have no bearing on the patentability.

Thus, the “deaf ears” referred to are those practitioners in the field who are following well-established law.

You don’t have to like current patent law. Many people don’t. European rules, for example, are different that ours. Note that not liking is distinct from not understanding.

- Registered Patent Agent

Re:Patent Law Explained

[Arker]

Actually you make a few mistakes. Most importantly, under US law at least, mathematics IS NOT patentable.

Furthermore software patents are NOT properly analogous to 'mechanical inventions [which] could be reduced to equations' - most obviously because no mechanical invention can actually be reduced to equations. It may *involve* equations, it may be *describable* in equations, certainly, but if it is *reducible* to equations then it is pure math and not patentable.

Judges and patent examiners may be permitting patents on pure math to issue, but they are doing so in *ignorance* because they dont understand the subject matter, not because the law actually allows it.

Re:Patent Law Explained

[Anonymous Coward]

creates something useful

Nonsense. Many patented processes are useless, in fact worse than useless in that when patented they do more harm than good.

Your post in part is a typical example of the superficial, circular, bogus logic of patent proponents - endlessly claiming one thing while actually doing something else entirely. The logical abilities of most patent proponents is poor, endlessly conflating words with ideas and creating tarpits of nonsense that have little connection with reality.

These rules are not really new.

Actually they are - being applied to a new domain (software) which was not even envisaged by the framers of the patent system.

Your, and the PTO's, hypocrisy is in automatically assuming without justification that patent rules should apply to every area of idea creation, particularly when patents already do not apply to large areas. When all you've got is a hammer everything looks like a nail.

Whether patents should be used in any particular area of idea creation is a matter of public policy and has nothing to do with the technical function that the PTO et. al. are providing.

The fact that the members of the patent industry have seen fit to grow their empire anyway at the expense of everybody else in the community is just a sad reflection of the state of the legal system today. You people are hated, unfortunately with good reason.

Re:Patent Law Explained

[Anonymous Coward]

These rules are not really new. They are the same rules that apply to mechanical inventions. For example, you cannot patent a “law of nature,” even it is something complex and nobody else knew about it. You can, however, patent a new device that takes advantage of this law of nature. For example, you cannot patent super-conductivity, but you can patent a useful device that uses super-conductivity.

This could point into the direction of current problems with software patents. While I cannot patent superconductivity I can patent a nonristor.

A nonristor is a device which can transport electric current without dissipation of thermal energy. The nonristor is made out of a superconducting material. It can be either line- or ring-shaped and can be manufactured in widths from 1m to 1m.

Basically this is a patent on everything that can be done with a superconductor, i.e. I have patented superconductivity. While I have not read any patents on serving ads on a website it is obvious for a computer scientist that this might involve a database which itself uses some linked lists internally. How can this be in any possible way help to innovate?

Re:Patent Law Explained

[Big_Breaker]

Mathematics will always be only a model of what occurs in the implementation. A physical device patent may also include a configuration of materials, treatments, dimensions, operating conditions, and other characteristics that contribute to the device operating as the mathematical model intends it to. In this way it is an implementation that can be considered non-trivial and innovative and much more than pure mathematics.

Software patents are largely math because the implementation in a device is carried out in the digital domain. I almost want to capitalize digital domain because it is very special. It is a contract that the analog reality of a computing system can be abstracted to transact purely in digital, generally binary, computations that will always give the intended digital output. Software is math because it must be according to the digital domain contract. Any computing system that fails to fully abstract the reality of the implementation into a purely digital construction is operating in error.

The only software patent that I see as worthy is RSA. But even there I think that it is the signing protocol that is innovative and I could be persuaded otherwise. The asymmetric computations that make it viable are too close to pure math to be patentable IMHO. RSA is tremendously valuable and it is difficult to imagine another way to do it. That makes a 20 year patent unpleasant but that alone is no reason not to grant the patent. The cotton gin seemed difficult to implement otherwise and very useful and it was certainly worthy of patent, though that patent was promptly ignored. In both cases the world probably benefitted from the public disclosure of the workings as described in full detail in the patent.

Re:Patent Law Explained

[Anonymous Coward]

However, if those data structures are used (critically) to perform useful work, such as to refine steel or to serve up ads on websites, then the ENTIRE process is patentable. Subject, of course to all the other restrictions, such as non-obviousness.

However, most patents I see to patent the application in something as specific "a process to refine steel" or "serve ads". Instead they always seem to refer to something such as "information retrieval system", which can be interpreted to mean any computer. The problem is that the patents are so broad as to apply to " done on a computer". I believe " done on a computer" qualifies as obvious.

Re:Patent Law Explained

[Theaetetus]

Furthermore software patents are NOT properly analogous to 'mechanical inventions [which] could be reduced to equations' - most obviously because no mechanical invention can actually be reduced to equations. It may *involve* equations, it may be *describable* in equations, certainly, but if it is *reducible* to equations then it is pure math and not patentable.

That's GP's point. Software patents are not claiming pure math (note: post-the CAFC Bilski decision). Instead, they recite things like "A computing device comprising memory and a processor, the processor configured to execute..." Now, while what that processor does *involves* equations or is *describable* in equations, and one could even describe the processor in equations, the entire claim, by definition, is not *reducible* to equations. Similarly, patents can also recite "An article of manufacture comprising a tangible computer readable medium, the medium encoding instructions for..." Just like a processor, you can't reduce a CD-ROM to an equation.

Now, you'll say that that's trivial... But in so doing, you'll have to also admit that it's completely true. Which makes your later statement about how judges and examiners aren't following the law false.

Re:Patent Law Explained

[Arker]

The formulaÃc recitals about 'a device' where that device is simply a general purpose computer programmed with the maths in question are a transparent dodge. 'Obviousness' is supposed to be a criteria too, and nothing could possibly be more obvious than using a general purpose computer to actually compute something, and control peripheral devices in accordance with the computations.

Re:Patent Law Explained

[Theaetetus]

The formulaÃc recitals about 'a device' where that device is simply a general purpose computer programmed with the maths in question are a transparent dodge. 'Obviousness' is supposed to be a criteria too, and nothing could possibly be more obvious than using a general purpose computer to actually compute something, and control peripheral devices in accordance with the computations.

Agreed, and no one argues that "[known method] performed by a computer" is suddenly patentable. It isn't. Computers are known. If the method is known, absent sudden unforeseen difficulties, then it's obvious to combine them. That's a simple rejection under 35 USC 103.

But we're not talking about "[known method] performed by a computer." We're talking about "[totally brand new method that no one's ever done before, but yes, may be reduced to math] performed by a computer." And even though the computer part is known, that doesn't make the entire thing obvious, because no one has ever done the method.

In short, "performed by a computer" gets you around the statutory requirement of 35 USC 101, even where the rest of the claim may be a mathematical algorithm. But, that has nothing to do with whether something is obvious or not under 35 USC 103, which is an entirely different question.

Old news

[Locke2005]

Doesn't IBM have a patent on Arithmetic coding? We crossed the bridge on patenting pure math a long time ago.

Re:Old news

[Arker]

Yes, they have been patenting pure math for a long time.

No, no court or legislature has ever contradicted the axiom that math isnt patentable. They do it instead by confusing patent examiners, judges, and juries that do not know math when they see it. An effective communications campaign from the math folks may still have a chance to do some good.

Re:Old news

[Anonymous Coward]

Rissanen, Jorma (May 1976). "Generalized Kraft Inequality and Arithmetic Coding" (PDF). IBM Journal of Research and Development 20 (3): 198–203. doi:10.1147/rd.203.0198. Retrieved 2007-09-21.

1976+20 2011

Not anymore they don't (don't know if they did or not, but it seems likely).

Re:Wheel was patented too recently...

[larry bagina]

If you take a photograph, you would own the copyright and could DMCA everyone.

If we required all software patents be Haskell...

[sco08y]

We could look them up with Hoogle or Hayoo API searches.

*And* you'd have a huge bank of open source software coming available as they expired.

The RSA Patent was pretty much pure math too...

[dpovey]

Those of you familiar with the RSA crypto algorithm will know it boils down to some incredibly simple number theory. ie. Generate two large primes p and q, and compute e and d such that: n = p.q e.d =~ 1 (mod n) All that math is in the patent, and that is the absolute essence of what they are patenting. But they get around this by drawing block diagrams for every possible way to create a communication channel that is encrypted and decrypted using RSA. They then just use the word device in the title, and Bob's your aunty's live in lover. The whole thing is peverse and absurd. But it's the way it has been for a long time, and this is unlikely to change anything.

are you sure?

[Anonymous Coward]

What exactly is a "computer-implemented invention"?

Is a compression algorithm a "computer-implemented invention"? The patents on arithmetic coding held up progress in that area for more than a decade. Is a video codec a "computer-implemented invention"?

Hint: Computers are general-purpose. By design, they can compute anything that is computable. That is the whole point. It can run any and every piece of software written for it. The same machine can implement any calculation you want. The software part (the inputs to the machine) is entirely mathematics.

Are you claiming that "General purpose machine" plus "software, which is entirely mathematics" equals a patentable invention?

Maybe you should read this article at Groklaw.

Re:are you sure?

[Grond]

I'm not arguing what should or should not be. I'm explaining what is. The patent was not reduced to math because a physical computer system is an element of the claim. And the Curry Howard Correspondence does not make computer implemented inventions unpatentable. These are facts, for better or worse.

Regarding the general purpose nature of (most) computers: "A general purpose computer, or microprocessor, programmed to carry out an algorithm creates 'a new machine, because a general purpose computer in effect becomes a special purpose computer once it is programmed to perform particular functions pursuant to instructions from program software.'" WMS Gaming, Inc. v. International Game Technology, 184 F. 3d 1339, 1348 (Fed.Cir. 1999) (quoting In re Alappat, 33 F.3d 1526, 1545 (Fed.Cir. 1994) (en banc)).

So, yes, general purpose machine plus software equals a patentable special purpose computer, subject to all the other requirements of patentability. That's the law right now, whether you agree or disagree with whether it ought to be so.

Please no.

[drb226]

...Or will they just decide to fix the inconsistency and make mathematics patentable?

Please no. I know this was tongue-in-cheek but I honestly wouldn't be surprised. Please, please no.

Re:Please no.

[Tablizer]

...Or will they just decide to fix the inconsistency and make mathematics patentable?

God knocks on Einstein's door:

"Remember E=mc^2?"

Einstein: "Well, yes, I discovered that."

God: "No you didn't. Pay up, Bub, or I'll turn gravity off and your ass will be in the asteroid belt."

Einstein: "Fine, here's my fucking wallet! Go use it to make snakes talk or something."
   

Patentable Mathematics ain't possible....

[OldHawk777]

Mathematics well about 99.999% of the math in use today is prior art that is over 100 years old. High level...theoretical mathematics typically only has a practical and applied science/economic purpose/use...value centuries later.

Also, saying someone has a legal right to a type of math/equation or theory is like saying someone owns all humans (...insanity/silly). Anyone can do 2+2..., if a SCO, Dell, Oracle, CN, EU, TW, US mathematician uses math, there is no reasonable way to say "I DID IT FIRST" or last.

E=MC^2 has more in common with a (non-abstract) fine art master piece, than a patent, copyright... troll-law or elitist-economics.

progress

[Anonymous Coward]

patents have to have technological progression, improvement. technology improves over time. stories don't.

Just because a story is 1000 years old, doesn't mean a similar story written today is any BETTER in terms of tech.

So you can't patent stuff like that because it isn't clear how it's causing technology to progress.

Exteme Irony .... needs attention

[Anonymous Coward]

http://www.google.com/patents?id=X4QXAAAAEBAJ&printsec=abstract&zoom=4&source=gbs_overview_r&cad=0#v=onepage&q&f=false

Google hosts the very patent in question.

Being understandable doesn't make it OK.

[Anonymous Coward]

You have a sorry job to do and I am sure you imagine that it is very important but the truth is if the patent office ceased to exist tomorrow there would be an immediate boom in innovation, competition, employment. lowering of prices and more besides. The only people to loose out would be the shareholders of the large corporations which hold most of the valuable patents and the workforce of sad individuals who currently get paid to administer the crackpot world which is the world of intellectual property.

I do NOT concur

[woolio]

The claims require a physical implementation (e.g. "An information storage and retrieval system").

You mean the algorithm has to be implemented/executed on a physical device? How difficult is that? Let's assume that a patent doesn't restrict the target algorithm to a restricted form of implementation such as a digital watch processor, a cell phone, etc. How is this different than patenting the math itself? Moreover, restricting the form of implementation would only necessitate that the patent submitter submit several patents, one for each conceivable application of his algorithm.

When patents become as broad as claiming usage of algorithm 'x' in 'a communication system', is this really any different than patenting algorithm 'x' itself? Does it matter if 'a communication system', is a remote-control changing the channel on a TV with IR LED pulses, a 4G cell phone, a low-earth orbiting satellite link, or tin cans connected by string?

Re:I do NOT concur

[Grond]

How difficult is that?

The difficulty of implementing the algorithm once it is described has very little to do with patentability. In fact, if the patent didn't give a person of ordinary skill in the art sufficient information to implement the algorithm without undue experimentation then the patent would be invalid for lack of enablement. As a non-software example, consider patented compositions of matter such as metal alloys. Many alloys are very simple to create in practice; the difficult is coming up with the particular mix of elements in the first place.

However, if the algorithm is known in the prior art, then adding "on a computer" to it is now (properly, in my opinion) considered to be obvious in most cases.

I'm not commenting on whether the claimed invention is obvious. I haven't done much kernel hacking; I'll take other commenters at their word that the method would have been obvious to a kernel programmer of ordinary skill. That's fine. My point is that the invention was not reduced to mathematics, as no amount of mathematics will conjure up a physical computer system. And in fact this is the case for essentially all software patents.

Declarative

[Arancaytar]

Red Hat is reportedly seeking a declarative judgement that this patent is invalid

That goes well with declarative evidence, I guess.

TRIPS is no problem

[ciaran_o_riordan]

> Unfortunately, this is not allowed under TRIPS.

TRIPS is no problem. TRIPS says that computer programs are to be considered literary works.

http://en.swpat.org/wiki/TRIPS

TRIPS says that patents have to be granted for inventions in "any field of technology", but there's no definition of this term, and if computer programs are literature and stories aren't patentable, then computer programs are not patentable under TRIPS.

Further, a general principal of law is that international treaties that restrict sovereignty are to be interpreted narrowly. Trying to stretch TRIPS to include software is not only incorrect but it's also against this general principal of law.

The pro-swpat lobbyists mostly try the "TRIPS says" trick when they're talking to people who they think won't read the text.

Ahhh Texans.

[Anonymous Coward]

May their homeschooling spread throughout the land.

Wood/metal, still the same device?

[Firethorn]

The Metal/Wood thing is something that the courts could get into, but the general argument in such a case is that they are effectively the same device.

But I'll admit that it is sort of the same argument. In order for a device to not be infringing despite doing the same task it would have to do the task via a different method, and the question comes in whether said method is different enough.

Personally I like the idea of making them come up with an actual device or implementation because I think it'd keep the patent trolls down - things like force feedback joysticks being held back for years because a company that has never produced one holds some patent on it; sure, they'll offer a license, but no technological help. They might of patented the idea of a control mechanism with a feedback loop, but they didn't do anything else to actually develop it.

Another one I remember has something to do with social media sites - and the company that did the patent doesn't run a social media site.

Re:Wood/metal, still the same device?

[kubernet3s]

Not necessarily: I'll bet you could totally patent an all wood replica of an iPhone

Usable product = patent? What a novel idea!

[Firethorn]

I have to agree. I think we need to go back to at least requiring an implementation of a patent in order to grant it.

Fear of patents shouldn't be a reason to NOT do something.

Chance alone is enough to cause problems

[tepples]

You cannot, however, "derive" a book or end up with the exact same arrangement of words, apart from chance.

However, chance alone is enough to cause problems with music. Plagiarism lawsuits* have been fought and won over "hooks", or the distinctive piece of the melody, as short as eight notes (seven intervals). Assume each interval can be to one of the seven notes of the scale and either short or long in duration, so 14 possibilities for each interval. Raise this to the power of 7 intervals to produce 105.4 million, which isn't even one melody for every person in the United States let alone the planet.

* Plagiarism here means infringement of copyright without attribution.

Such a computer system is in the prior art

[tepples]

The specification makes it clear that "information storage and retrieval system" refers to a computer system.

Such a computer system is in the prior art. Therefore the inventive step from the prior art (a computer system) to an allegedly patentable invention is a mathematical formula. Or what am I missing?

Re:Such a computer system is in the prior art

[Grond]

The combination of two things, each in the prior art, can be patentable. However, the Supreme Court has it made it clear that such combinations are unlikely to be patentable where the components are fulfilling known, predictable functions and the number of possible combinations was itself known and finite. But again, this is an obviousness argument. I'm not arguing that the invention was or was not obvious. I'm only explaining that the fact that part of the invention is based on (or is equivalent to) a mathematical formula does not necessarily render the entirety of the invention unpatentable.

Law makers don't understand maths

[Anonymous Coward]

Maths is too tough. They just understand money. Drooling baboons.

A computer program is an implementation of math.

[master_p]

Math cannot and should not be patented. However, a computer program is an application of math on a mechanical device, and so it can be patented.

The above is not my view, is what the patent system is based upon.

Personally, I would like to see that practical applications of abstract concepts are not patentable at all.

Re:A computer program is an implementation of math

[Anonymous Coward]

So solving differential equations on a computer is patentable.

As is fast furier transform.

As is erastothenes sieve.

As is adding numbers on calculator.

Etc.

Behold the end of our civilization.

A US thing

[Kanel]

Patenting software is a US thing. I'm under the impression that under most European laws, you cannot patent a software algorithm nor the code implementing it. Sure, you can work around this limitation in some cases, but the patent lawyers in the european company I work in do not push us to patent algorithms, for this very reason.

Books can not be patented

[garyebickford]

Likewise, books are language. Can books be copyrighted? No one owns language.

Copyright is not patent, and trademarks are neither copyright nor patent. Each protects a different kind of intellectual property. Concepts or ideas can not be copyrighted, only a particular arrangement of words (or symbols, or whatever) - a way of describing something. When the same ideas are described in another way (generally not a difficult proposition), there is no copyright violation.

Your point about the power drill is well taken. 'Power drill' is too far from 'metal and plastic' to be considered a natural progression that anyone could think of. Nobody ever saw a lump of metal and another lump of plastic, and said, 'Aha - I have just invented electric motors, three-blade chucks, twist drill bits, variable speed power electronics, etc.' There were thousands of patents that marked our progress from said lumps to the present day drill. But I think the point about match and computer implementations is that under patent law, math is natural law, and is discovered, not invented. A different implementation of a mathematical algorithm is still the same algorithm.

Don't forget that until (IIRC) 1986, all software was unpatentable specifically because 'software is algorithms, algorithms are math, and (since math is natural law, and is discovered, not invented) math is unpatentable', preventing the many 'heroes' of the computer industry and academia from patenting the seminal ideas upon which all computing is founded from taking out patents. Ideas such as virtual memory, paging systems, superscalar architecture, push-down stacks, queues, the many seminal ideas upon which all computer graphics is based, all the various parts of an arithmetic logic unit, TCP/IP, HTTP, FTP, SMTP and thousands of other ideas were developed without any expectation of patentability. The entire structure of modern computing was developed without patents, except for certain hardware details. Therefore, for the reason of fairness alone, allowing patents for one-click ordering and the other trivialities that all depend on these unpatentable core ideas should never have been allowed. How can an HTML trick be patentable, when HTML itself was not patentable at the time? All of XML, and the many other MLs that were derived from SGML, would have been patentable had they only been created twenty years later. Shouldn't Berners-Lee have made a zillion dollars before Amazon, or Twitter, or Facebook? Shouldn't they have all been made to pay license fees? Allowing patents for trivial 'improvements' (if that is what they are) of implementations on these core ideas and implementations has always been a travesty.

Books are math

[metacell]

Books consist of a sequence of n different typographic characters. By assigning a number from 0 to n-1 to each typographic character, any book consisting of m characters can be read as an m-digit number written in base n. Thus, we have proven an isomorphism between books and numbers.

Does that prove that books are just numbers, and that, since numbers aren't copyrightable, books can't be, either?

Sorry, I think the "software is math" argument is silly. Software patents are granted for specific applications of software, as understood in natural language terms, not for specific mathematical algorithms.

There are many other good reasons software patents should be abolished, for example, because they fail the basic rationale for patents. Patents came about to give an incentive to businesses to publicise their inventions, so others could build and improve upon them. But software businesses don't generally keep their algorithms or data structures secret - the exact method you use to traverse hash tables or convert numbers from binary to octal are rarely that important to a software business, and they can be found out anyway through reverse-engineering. Algorithms are considered part of a programmer's toolbox - useful, but not essential, and usually possible to work around if you can't use a specific algorithm. Most research is either shared freely when programmers change jobs, or is financed by universities, so there's no need for an incentive to publish "software innovations".

Not innovative

[HalAtWork]

But adding an information storage and retrieval system to a mathematical formula is not very innovative. Information storage and retrieval systems have been combined with mathematics for centuries. Just look at math books and adding machines.

Re:Not actually reduced to math - Player Piano's?

[Anonymous Coward]

Instructions, and the machine they run on... If you patented a Player Piano, and I create a music role, can I protect my music role with a patent or a copyright? If I get a patent, then that is protection for any "device" that produces "equivalent" output... so, a patent on that "expression" of that set of notes lays claim to any and all means to deliver the same notes?... no.

Re:Software is not math!

[turkeyfish]

Software is the implementation of math on a computing device. Math can exist without software, but software can not exist without math. Math is abstract and general, whereas software is implementation specific.

Prior art for mathematical computation.

[DamnStupidElf]

Using a physical device to evaluate an arbitrary mathematical function is at least as old as Babbage. A formal model of computation generic enough to be called Universal is as old as Turing. If a patent application claims nothing more than a physical device for evaluating a particular mathematical function then how can there be any merit in the claims? All the merit must rest in the mathematical function itself because Babbage and Turing provided prior art; and if the mathematical function has no relevance to a physical or useful process then the only claims of the application are for pure mathematics.

I did not see any reference in the patent claims to any method or physical process, only references to the information storage and retrieval system. Maybe I'm wrong, but I understood that the claims are what matters; not the accompanying description and exposition.

Nope, the computer is the mechanical device

[Anonymous Coward]

Nope, the computer is the mechanical device, not the software.

So one answer would be "you've patented linked lists on your computer. But I've used my computer which has a different configuration". If in response you say that it's "any computer" then you have no mechanical device, you have a theoretical construct and that is abstract, not a device.

Another answer would be that the software isn't patented and neither is the computer, nor the fact that a solution to a problem must exist, but merely that your specific aggregation of all the above solves the problem in a new way. Therefore changing the source to give the same result from a different configuration is a different device combination.

Every [valid] patents have workaround

[Anonymous Coward]

Before this I only had an vague idea why software patents shouldn't be exists. But when discussion come across "protecting this implementation" thing, I got some interesting speculation.

Let's compare Mauveine patent and mp3 patent (in the way that it's supposed to be)

replace 'Mauveine' with 'MP3' (or 'h264')

[Mauveine patent]
Content: this (specific) method to produce Mauveine
Workaround: use different way to produce Mauveine

[MP3 patent]
Content: this (specific) method to produce MP3
Workaround: use different way to produce MP3

So, we were allowed to 'use different way to produce MP3'?

Screw it! Don't know that analogy is correct or not, but I didn't live in softpatents-valid country.