Oracle vs. Google (or, yet another reason not to use Java)

[Arariel]

Also, I don't see why reverse engineered software is necessarily morally acceptable. It's taking another person's innovation and design and, without his permission, profiting off of it. Unless you're talking about manually reverse engineered software (which requires innovation on the part of the engineer) and not just decompiled software.

Key words being clean room. See, decompiled software infringes on the copyright, because it copies the expression of the algorithm. Clean room reverse-engineering does not involve decompilation (which isn't exactly feasible at any rate) or disassembly. For example, Wine is clean room, reverse-engineered software, as is Android's Dalvik virtual machine.

The "it's just mathematics" argument doesn't work. Anything can be expressed as just mathematics. For example, if I were to scan the information contained within a patent into a computer. What is patentable is design. I don't see why it should matter that the design is expressed mathematically. That's a (pointless) theoretical argument that fails to solve the practical problem of how people protect their inventions.

Up until about 1995, courts agreed software algorithms were unpatentable because they are mathematics.

In the past claims to pure instructions were generally considered not patentable because they were viewed as "printed matter," that is, like a set of instructions written down on paper.

Except then, in In Re Beauregaurd, the USCAFC decided that it is patentable if you begin the patent with something like "A stock PC on which is loaded an algorithm to…", because in doing so, you are apparently creating a "new machine". But the Supreme Court recognises this is all loophole-abusing bullshit:

The Supreme Court ruled three times that prepending "a stock computer on which is loaded…" to an otherwise unpatentable algorithm is just a wording trick, designed to get around the established consensus that software and pure mathematics are not patentable. As one Supreme Court ruling explained: "Indirect attempts to obtain [software] patents … by drafting claims as a … machine or components thereof programmed in a given manner, rather than as a program itself, have confused the issue further and should not be permitted." Another ruling repeated this position, explaining that "insignificant postsolution activity will not transform an unpatentable principle into a patentable process. To hold otherwise would allow a competent draftsman to evade the recognized limitations on the type of subject matter eligible for patent protection."

How did the CAFC reconcile its rulings that loading software on a PC creates a "new machine" with the Supreme Court's statements? It didn't: none of the key rulings that established the concept of a "new machine" addressed the several statements by the Supreme Court that such reading of a software patent is just a wording trick.

Software patents have their flaws. But if you're gonna tell me that MP3 isn't a significant improvement on the status quo at the time (or that H264 isn't better than OGG Thedora)... or that JPEG files and the LZW Compression algorithm aren't unique works in of themselves...

I really don't know what to say about that. These things have value, and it is appropriate to reward people and encourage engineers and scientists to create more of these.

Now... thats not necessarily saying that I agree with software patents in their current form (lol one-click shopping). But your argument is significantly more extreme than mine. You seem to be claiming that algorithms do not have any value.

You are dead wrong on that fact. Algorithms have a significant amount of value to them. Calling them "just mathematics" is rather insulting to people who come up with these sorts of things.

I am most certainly not saying they do not have value. But mathematics also has value. And up until recently, algorithms were never patentable. I mean, a Fourier analysis is "a unique work", as are algorithms for finding digits of pi and prime numbers, and integrating functions. But why should Fourier have been unable to patent, say, "a method of representing periodic functions as the sum of trigonometric functions", but Crapple to be able to patent an unlock button? It damn well has value, so what separates a patent on the Fourier transform from a software patent?

In all honesty, software algorithms are "just mathematics". You think it's insulting? Well, that's what they are. Saying it's insulting won't change that fact. Any program can be written in lambda calculus. Therefore, it is mathematics. The expression of the mathematics may be protected under copyright law, but the idea itself is not protected under patent law. The idea may have value, but that is completely irrelevant to whether or not it is patentable.

Also, Libnitz vs Newton. Mathematicians have been complaining about stolen works forever.

(unless it's plagiarism, which is a different matter entirely)

But I have never heard anyone claiming calculus was "stolen" from Leibniz. People may have claimed Leibniz copied him, or say his work was hopelessly derivative, but to say he "stole" it would be just quite silly.

-----

If you're still convinced software patents are a good idea, consider that if you sell products online, you could be violating at least 4,319 patents, advertising, receiving payments, or shipping goods online could violate around 11,000 (as of 2003, probably higher by this point),that the Green Bay Packers, Oprah Winfrey, Kraft Foods, and J. Crew, that Crapple can apparently sue someone for using squares with rounded corners on the basis of patent violations alone, that Amazon can patent buying something with one click, that among interviewed venture capitalists and programmers, none of them considered software patents, and that holders of software patents freely admit the primary purpose is no longer to encourage creativity, and maybe you'll reconsider.

Software patents are not used to protect the underdog from other people making money off of his or her ideas. They're used to protect the status quo. Companies holding many patents will go after companies that have few (Google until recently had less than a thousand patents), while avoiding each other because they'll be mutually sued to oblivion (hence why Google had attempted to acquire Motorola shortly before the Oracle suit, and acquired more than 17,000 patents). Mutually Assured Destitution, no doubt.

So the big companies definitely won't sue each other. Instead, they'll sue the small companies, the start-ups, the ones doing real innovating.

But all right, it's all for progress, and promotion of developing new stuff, and so on.

[sourmìlk]

Nobody said that software patents were implemented particularly well. Apple shouldn't be able to patent rounded rectangles, but that doesn't indicate a theoretical problem with software patents.

And I don't think anybody's stance is that it's "insulting" to call software mathematics, but that it's an oversimplification. Everything is mathematics. To say that something can be simplified into mathematics does not and should not have any bearing on whether or not it can be patented. That's still a theoretical argument that ignores the practical problem of how a developer can protect his work. Patenting mathematics sounds stupid, but you haven't really explained the practical implications of allowing certain designs reducible to mathematics to be patented. Nobody here is suggesting we should allow all math to be patented.

As for the Supreme Court thing, see above: nobody said that software patents were implemented particularly well. It's stupid that there's a wording loophole that you need to get software patented, but that doesn't mean software shouldn't be patented. The question still remains: how should software developers protect their designs?

[Arariel]

Except the Supreme Court has explicitly said software is not patentable, and these rulings have not been overturned.

Protect your software with copyright, or whatever. But patenting it is ridiculous.

Again, why should Fourier analysis not have been patented by Fourier? Is there something special separating software algorithms from mathematical ones (never mind that software algorithms are not merely reducible to mathematical ones, but they actually are mathematical ones)?

[Dark567]

See every song, every book can be represented by an algorithm. Yet we still let them be copyrighted, because that math represents songs and books. The same goes for software. Sure those programs can be represented with math, but that hardly means you shouldn't be able to copyright it.

[Arariel]

Except I haven't said you shouldn't be able to copyright it. I said you shouldn't be able to patent it. You can't patent a song, nor can you patent literature and art. Why should you be able to patent software?

[sourmìlk]

Except the Supreme Court has explicitly said software is not patentable, and these rulings have not been overturned.

No, they said that software shouldn't be patentable through the current means by which they are patented.

Protect your software with copyright, or whatever. But patenting it is ridiculous.

This doesn't really work as a response. Either copyright offers the same protections as patents (in which case why would you care which category software protection falls under?) or copyright offers different protections (in which case why shouldn't software developers be given the protections given through patents?).

Again, why should Fourier analysis not have been patented by Fourier? Is there something special separating software algorithms from mathematical ones (never mind that software algorithms are not merely reducible to mathematical ones, but they actually are mathematical ones)?

I think these questions miss the point. If Fourier can't patent Fourier analysis, why does that necessitate that software shouldn't be patented? Everything is reducible to mathematics. Clearly there are things distinguishing things from other things. You can tell a piece of software from a pure mathematical equation. Why do these differences not matter?

[Arariel]

No, they said that software shouldn't be patentable through the current means by which they are patented.

Gottschalk v. Benson, 409 U.S. 63 (1972) was a United States Supreme Court case in which the Court ruled that a process claim directed to a numerical algorithm, as such, was not patentable because "the patent would wholly pre-empt the mathematical formula and in practical effect would be a patent on the algorithm itself." That would be tantamount to allowing a patent on an abstract idea, contrary to precedent dating back to the middle of the Nineteenth Century. The ruling stated "Direct attempts to patent programs have been rejected [and] indirect attempts to obtain patents and avoid the rejection ... have confused the issue further and should not be permitted."[1] The case was argued on October 16, 1972 and was decided November 20, 1972.

This doesn't really work as a response. Either copyright offers the same protections as patents (in which case why would you care which category software protection falls under?) or copyright offers different protections (in which case why shouldn't software developers be given the protections given through patents?).

Copyright covers the expression. Patents cover the idea. For example, Star Wars is copyrighted, but if patent law applied to film, there could be a patent on teenage heroes who grew up on a farm and was trained by an ancient order of peacekeepers to defeat an evil empire, in which case Eragon would violate patent law. I've already covered why patents shouldn't cover software.

I think these questions miss the point. If Fourier can't patent Fourier analysis, why does that necessitate that software shouldn't be patented? Everything is reducible to mathematics. Clearly there are things distinguishing things from other things. You can tell a piece of software from a pure mathematical equation. Why do these differences not matter?

Either software algorithms are patentable, in which case mathematical algorithms are also patentable (because they can be expressed as software algorithms), or mathematical algorithms are not patentable, and then neither are software ones.

[sourmìlk]

No, they said that software shouldn't be patentable through the current means by which they are patented.

Gottschalk v. Benson, 409 U.S. 63 (1972) was a United States Supreme Court case in which the Court ruled that a process claim directed to a numerical algorithm, as such, was not patentable because "the patent would wholly pre-empt the mathematical formula and in practical effect would be a patent on the algorithm itself." That would be tantamount to allowing a patent on an abstract idea, contrary to precedent dating back to the middle of the Nineteenth Century. The ruling stated "Direct attempts to patent programs have been rejected [and] indirect attempts to obtain patents and avoid the rejection ... have confused the issue further and should not be permitted."[1] The case was argued on October 16, 1972 and was decided November 20, 1972.

1995 > 1972, thus a decision made in 1995 overrides one made in 1972. Since 1995, at least as far as you've shown, the supreme court has not rejected software patents as inherently problematic, but simply disapproved of the current method of patenting software.

Copyright covers the expression. Patents cover the idea. For example, Star Wars is copyrighted, but if patent law applied to film, there could be a patent on teenage heroes who grew up on a farm and was trained by an ancient order of peacekeepers to defeat an evil empire, in which case Eragon would violate patent law. I've already covered why patents shouldn't cover software.

That's just a problem of specificity. Are patents necessarily less specific than copyrights?

Either software algorithms are patentable, in which case mathematical algorithms are also patentable (because they can be expressed as software algorithms), or mathematical algorithms are not patentable, and then neither are software ones.

I just explained why this isn't the case. There is a difference between a pure mathematical algorithm and a piece of software as evidenced by the fact that you can distinguish between the two. Why does the difference not matter? You say that there isn't a difference, but that clearly isn't the case as we can distinguish between them.

[Arariel]

Can we? Well, it's possible in the same way a work originally written in English is distinguishable from a work originally written in Spanish, but that doesn't nullify the fact we can express both works in either English or Spanish.

And it looks like I was more narrow than I thought, because apparently there's precedent against copyright on abstract ideas stretching back to the nineteenth century, and you can't say a software algorithm is anything but an abstract concept.

[sourmìlk]

Can we? Well, it's possible in the same way a work originally written in English is distinguishable from a work originally written in Spanish, but that doesn't nullify the fact we can express both works in either English or Spanish.

Yes, we can. As to the second sentence: so?

And it looks like I was more narrow than I thought, because apparently there's precedent against copyright on abstract ideas stretching back to the nineteenth century, and you can't say a software algorithm is anything but an abstract concept.

True, but I can say that actual software isn't an abstract concept.

[Arariel]

Yes, we can. As to the second sentence: so?

Well, they're two different methods of expression, when in reality they're the exact same thing. But there's no fundamental difference between the two, other than "that looks like maths to me" and "that looks like software to me". If you don't have an explicit rule on what separates software algorithms from mathematical ones, than it's nothing more than a gut feeling like separating English works from Spanish works translated into English.

True, but I can say that actual software isn't an abstract concept.

It's the software algorithm being patented. The actual software is undoubtedly covered by copyright law. The character of a teenage hero who grew up on a farm and was trained by an ancient order of peacekeepers to defeat an evil empire is an abstract concept, but Star Wars and Eragon are not, and are not covered by patent law.

[sourmìlk]

Yes, we can. As to the second sentence: so?

Well, they're two different methods of expression, when in reality they're the exact same thing. But there's no fundamental difference between the two, other than "that looks like maths to me" and "that looks like software to me". If you don't have an explicit rule on what separates software algorithms from mathematical ones, than it's nothing more than a gut feeling like separating English works from Spanish works translated into English.

What makes you say that it's only a gut feeling that can distinguish between software and a pure mathematical equation? Only one of those things runs on computers. "a^2 + b^2 = c^2" is not an executable. I still don't get the point of this translation stuff.

It's the software algorithm being patented. The actual software is undoubtedly covered by copyright law. The character of a teenage hero who grew up on a farm and was trained by an ancient order of peacekeepers to defeat an evil empire is an abstract concept, but Star Wars and Eragon are not, and are not covered by patent law.

Okay, but how is a software algorithm analogous to the former and not the latter?

[Arariel]

What makes you say that it's only a gut feeling that can distinguish between software and a pure mathematical equation? Only one of those things runs on computers. "a^2 + b^2 = c^2" is not an executable. I still don't get the point of this translation stuff.

Patent no. 482984892, an algorithm for determining the length of the hypotenuse of a right triangle:
1. Square one leg and add to the square of the other leg.
2. Take square root of sum obtained in step 1.

Okay, but how is a software algorithm analogous to the former and not the latter?

A software algorithm is a concept which has an infinite number of possible implementations, in the same way our farmboy hero is a concept with an infinite number of possible implementations. It's fair to copyright the software, but not to patent the abstract concept.

[sourmìlk]

What makes you say that it's only a gut feeling that can distinguish between software and a pure mathematical equation? Only one of those things runs on computers. "a^2 + b^2 = c^2" is not an executable. I still don't get the point of this translation stuff.

Patent no. 482984892, an algorithm for determining the length of the hypotenuse of a right triangle:
1. Square one leg and add to the square of the other leg.
2. Take square root of sum obtained in step 1.

I'm not sure what you're trying to say here.

A software algorithm is a concept which has an infinite number of possible implementations, in the same way our farmboy hero is a concept with an infinite number of possible implementations. It's fair to copyright the software, but not to patent the abstract concept.

Software are implementations. They are the specific set of instructions telling a computer what to do. Unless I'm misinterpreting how you define 'implementations'.

[Arariel]

I'm not sure what you're trying to say here.

It's an executable you can run on a computer.

Besides, you don't necessarily have to run a software algorithm on a computer (the kind consisting of silicon).

Software are implementations. They are the specific set of instructions telling a computer what to do. Unless I'm misinterpreting how you define 'implementations'.

When I say software, I mean source and binaries. Those are implementations of algorithms analogous to a story's implementations of tropes.

[sourmìlk]

I'm not sure what you're trying to say here.

It's an executable you can run on a computer.

Besides, you don't necessarily have to run a software algorithm on a computer (the kind consisting of silicon).

Kind of you do? Also, that's not an executable that can run on a computer as it provides no means of setting the variables, calculating them, or outputting them. Those instructions are insufficient to tell a computer what to do.

Software are implementations. They are the specific set of instructions telling a computer what to do. Unless I'm misinterpreting how you define 'implementations'.

When I say software, I mean source and binaries. Those are implementations of algorithms analogous to a story's implementations of tropes.[/quote]
I think we've been talking past each other for a bit. When you say 'software algorithms', are you referring to the mathematical expression of the software's code, or merely algorithms used in the creation of the software?

[KnightExemplar]

Can we? Well, it's possible in the same way a work originally written in English is distinguishable from a work originally written in Spanish, but that doesn't nullify the fact we can express both works in either English or Spanish.

Can you not tell the difference between a naive implementation of the Discrete Fourier Transform over the Fast Fourier Transform?

What about Karatsuba multiplication, Long Multiplication, and other forms of multiplication?

They all implement the same mathematical formula. But the algorithms are different. You can't patent the idea of multiplication. But you can patent say... Karatsuba multiplication.


----------------

Another thing Arariel: we can write laws however the hell we want. We live in a country where we are only governed by the constitution... and the constitution can be changed to say whatever we want as well. So please no more of this abstract "X can't be patented" bullshit.

Laws can be written, and have been written in whatever way we please. The important bit is whether or not the law makes life better. IE: the purpose of Patent Law is to encourage scientists and engineers to create new tools for us. And in 20 years, those tools are then given to the general American Public as they're released into the public domain.

You're gonna have to have a broader argument. Why are software patents bad for the developer community? This important aspect of the debate has so far been ignored.

[Arariel]

Kind of you do? Also, that's not an executable that can run on a computer as it provides no means of setting the variables, calculating them, or outputting them. Those instructions are insufficient to tell a computer what to do.

Practically, yes, technically, no.
Is it necessary to provide those for a patent? I could write those into an executable, of course, but into the statement of the patent itself? I doubt it. Those are formalities necessary for the actual code.

I think we've been talking past each other for a bit. When you say 'software algorithms', are you referring to the mathematical expression of the software's code, or merely algorithms used in the creation of the software?

When I say software algorithms, I meant what is protected by software patents. I make this distinction to set it apart from software as the source and binaries.

Can you not tell the difference between a naive implementation of the Discrete Fourier Transform over the Fast Fourier Transform?

What about Karatsuba multiplication, Long Multiplication, and other forms of multiplication?

They all implement the same mathematical formula. But the algorithms are different. You can't patent the idea of multiplication. But you can patent say... Karatsuba multiplication.

Errr... no? Not under current patent laws, but if abstract ideas were patentable, yes, that would be patentable (if the creator had applied for one). But are you seriously suggesting mathematical algorithms should be patentable?

Another thing Arariel: we can write laws however the hell we want. We live in a country where we are only governed by the constitution... and the constitution can be changed to say whatever we want as well. So please no more of this abstract "X can't be patented" bullshit.

Laws can be written, and have been written in whatever way we please. The important bit is whether or not the law makes life better. IE: the purpose of Patent Law is to encourage scientists and engineers to create new tools for us. And in 20 years, those tools are then given to the general American Public as they're released into the public domain.

You're gonna have to have a broader argument. Why are software patents bad for the developer community? This important aspect of the debate has so far been ignored.

Under current legal rulings by the Supreme Court which have not been overturned, no, software can't be patented. Of course, Congress is free to make new laws regarding patents, but until they do so, every software patent that exists should be null and void.

And the public domain argument is ridiculous. You can write software isolated from society using only techniques developed more than twenty years ago and still violate some software patent. Even Ogg, which was designed specifically to not be patent-encumbered could possibly be subject to unknown patents, which was one of the arguments against using it as a standard in HTML5 (the argument being h.264 is already dominant, and therefore anyone who has a patent which h.264 could have violated would have come forth a while ago, while Ogg is not widespread, and patent holders could be waiting for it to become so in order to better profit off of it through exorbitant fees). Similar bits for Google's WebM.

More important than the question of how are software patents bad for developers (they are) is how are they good. Can you show that software patents benefit society? Considering how most venture capitalists and programmers don't even consider patent law, I wouldn't think so.

As for how it's bad? Well, consider what we wouldn't have if all software algorithms were patented. We wouldn't have any GNU (violates patents on Unix, 1969, less than 20 years before GNU). That includes no alternatives to the Bourne shell, such as bash, csh, zsh, what have you. Probably no Linux either (it would violate some patent or the other. I remember reading about the average operat). In fact, I would bet if the same rules for software patents existed then as did today, you wouldn't be able to find a single piece of software (other than the trivial "Hello world"-type ones) that didn't violate patent law. Although Microsoft still FUDed the shit out of software patents in Linux.

Now you may say it's our particular implementation of software patents that leads to this, but it's the just that the implementation of software patents has led to this. Software will be developed regardless of whether or not patents for it exists, and software patents have only existed to suppress innovation.

[sourmìlk]

Those are still arguments about why we shouldn't patent things too generally, not why we shouldn't patent things. Patents, to my knowledge, are not necessarily so general as to forbid anything similar. They protect a particular nontrivial design. Part of the reason that you can't patent mathematical equations is that they're trivially easy to manually reproduce. Observe:

a^2 + b^2 = c^2
area = pi*r^2
e^(i * pi) + 1 = 0

etc.

[Diadem]

Part of the reason that you can't patent mathematical equations is that they're trivially easy to manually reproduce.

How many fields medals do you have?

[SlyReaper]

Those are still arguments about why we shouldn't patent things too generally, not why we shouldn't patent things. Patents, to my knowledge, are not necessarily so general as to forbid anything similar. They protect a particular nontrivial design. Part of the reason that you can't patent mathematical equations is that they're trivially easy to manually reproduce. Observe:

a^2 + b^2 = c^2
area = pi*r^2
e^(i * pi) + 1 = 0

etc.

No. No no no no no no no. The reason you can't patent mathematical equations is because you can't patent or copyright facts. It has nothing to do with how easy it is to reproduce.

And as an aside, claiming that all mathematical equations are easy to reproduce demonstrates a truly staggering level of ignorance on your part. Unless you mean easily reproducible using the old ctrl-c ctrl-v method, which copyright covers for non-mathematical things anyway.

[KnightExemplar]

Errr... no? Not under current patent laws, but if abstract ideas were patentable, yes, that would be patentable (if the creator had applied for one). But are you seriously suggesting mathematical algorithms should be patentable?

I'm disagreeing with you because your argument is unfocused. I don't particularly like software patents in their current form... but there I disagree with almost everything you've said so far. We may have similar conclusions, but I don't like your reasoning.

Mainly this: there is a difference between an "equation" (ex: multiplication) and an "algorithm". (ie: Long Multiplication, or Karatsuba multiplication). Ideally, we want to encourage more of the latter. Creating new solutions and designs can only be a good thing. The question therefore, is whether or not patents encourage creation of new algorithms... or if they discourage them (because inventors are worried about stepping on someone else's patents like a minefield).

I do hold the belief that an "algorithm" in this manner is a true invention in of its own right. Karatsuba Multiplication is sufficiently brilliant that I compare it to the light bulb as far as an invention is concerned. As such, we need to encourage the creation of such things. I would be in favor of any law that pushes us in that direction.

Under current legal rulings by the Supreme Court which have not been overturned, no, software can't be patented. Of course, Congress is free to make new laws regarding patents, but until they do so, every software patent that exists should be null and void.

And the public domain argument is ridiculous. You can write software isolated from society using only techniques developed more than twenty years ago and still violate some software patent. Even Ogg, which was designed specifically to not be patent-encumbered could possibly be subject to unknown patents, which was one of the arguments against using it as a standard in HTML5 (the argument being h.264 is already dominant, and therefore anyone who has a patent which h.264 could have violated would have come forth a while ago, while Ogg is not widespread, and patent holders could be waiting for it to become so in order to better profit off of it through exorbitant fees). Similar bits for Google's WebM.

More important than the question of how are software patents bad for developers (they are) is how are they good. Can you show that software patents benefit society? Considering how most venture capitalists and programmers don't even consider patent law, I wouldn't think so.

As for how it's bad? Well, consider what we wouldn't have if all software algorithms were patented. We wouldn't have any GNU (violates patents on Unix, 1969, less than 20 years before GNU). That includes no alternatives to the Bourne shell, such as bash, csh, zsh, what have you. Probably no Linux either (it would violate some patent or the other. I remember reading about the average operat). In fact, I would bet if the same rules for software patents existed then as did today, you wouldn't be able to find a single piece of software (other than the trivial "Hello world"-type ones) that didn't violate patent law. Although Microsoft still FUDed the shit out of software patents in Linux.

Now you may say it's our particular implementation of software patents that leads to this, but it's the just that the implementation of software patents has led to this. Software will be developed regardless of whether or not patents for it exists, and software patents have only existed to suppress innovation.

I think you've made some decent points that I can't disagree with (because frankly, I agree with you. Software patents are ultimately a bad idea. At least in their current form with 20 year monopolies and so forth).

However... I still don't like the style of your argument, but at least its somewhat connected to reality now. The only part of your post that was decently done is IMO the HTML5 argument.

Here are a few counterpoints:
* Software companies are always going to be assholes who abuse copyright and patent law. As this topic has shown... Oracle can claim copyright over an API to an open source language. Demonstrating that companies have been jackasses doesn't really say anything. Laws will never be perfectly written, and legal teams will always find ways to twist a good idea into one that they abuse.

* Patented network protocols existed. IIRC, ATM for example is a protocol that was encumbered by lots of patents during the rise of TCP/IP. But as the core of TCP/IP was patent free (designed by US Government and thus couldn't be patented)... many people decided to implement it and it became the defacto-standard of the internet. Similarly, OpenID demonstrates the ability for the software community to use software patents in a pure defensive manner with the non-assertion covenants.

Indeed, in the areas where it matters... companies (such as Microsoft) are finding that non-assertion covenants and open source software are mutually beneficial to themselves and the world in general. So throwing all this doom and gloom around because of software patents is pure FUD IMO. There are clear cut cases where software patents encumber the general public... I think HTML5 is a great example in this regards. But to extrapolate the argument and claim that Linux wouldn't have been possible??!?!

There's only so far you can push a hypothetical.

[sourmìlk]

I'm not saying mathematical equations are trivially easy to re-derive. That's certainly not the case. But manually writing them out given knowledge of them? Unless there's an equation that's thousands of characters long (and there might be, for all I know), then they are trivially easy to reproduce. Note that I didn't say that was the entire reason why they're unpatentable, and I like your 'you cant patent facts' reason better.

[SlyReaper]

I'm not saying mathematical equations are trivially easy to re-derive. That's certainly not the case. But manually writing them out given knowledge of them? Unless there's an equation that's thousands of characters long (and there might be, for all I know), then they are trivially easy to reproduce. Note that I didn't say that was the entire reason why they're unpatentable, and I like your 'you cant patent facts' reason better.

To be able to reproduce an equation from memory, you really need to understand the meaning behind it. If all you're doing is memorizing a sequence of symbols with an equals sign in the middle, then you'll most likely make a mistake in writing it down. It is most certainly not trivially easy to perfectly rattle off the formula for, say, the Hausdorrf metric (to pick a fairly simple example), without understanding what it means. And getting that understanding is not trivial either.

Sorry to go off on a tangent like this, but I'm a mathematician myself, and I dislike it when people tell me that the thing I spent years studying at university, and continue to use in my day job, is trivial.

[sourmìlk]

I don't see why it needs to be from memory. Looking at it on wikipedia I could copy it down in a few seconds. But I certainly don't think that mathematics is trivial. The fact that I can easily copy equations says nothing about the complexity of understand mathematics.

[SlyReaper]

To look it up on wikipedia, you need to know what you're looking for. That's also non-trivial.

[sourmìlk]

Right, but that's not a component of copying it, and I don't think that would have an impact on its patentability.

[SlyReaper]

It doesn't. Like I said, the reason you can't patent maths is because you can't patent facts.

[sourmìlk]

Looking, it appears that the nontriviality requirement applies mostly to non-obviousness, not triviality of copying, so yeah.

[Dark567]

It doesn't. Like I said, the reason you can't patent maths is because you can't patent facts.

Right and you can patent software because its a creation, not a fact. When you are writing code, you aren't just writing one algorithm. Your using many, and for that matter your often combining it with specific data. That's creation, not fact.

[Arariel]

I'm disagreeing with you because your argument is unfocused. I don't particularly like software patents in their current form... but there I disagree with almost everything you've said so far. We may have similar conclusions, but I don't like your reasoning.

Mainly this: there is a difference between an "equation" (ex: multiplication) and an "algorithm". (ie: Long Multiplication, or Karatsuba multiplication). Ideally, we want to encourage more of the latter. Creating new solutions and designs can only be a good thing. The question therefore, is whether or not patents encourage creation of new algorithms... or if they discourage them (because inventors are worried about stepping on someone else's patents like a minefield).

I do hold the belief that an "algorithm" in this manner is a true invention in of its own right. Karatsuba Multiplication is sufficiently brilliant that I compare it to the light bulb as far as an invention is concerned. As such, we need to encourage the creation of such things. I would be in favor of any law that pushes us in that direction.

Which means if a student uses Karatsuba multiplication on his homework, he's violated patent law, if an accountant uses it in some of his work, she's broken patent law, and if someone teaches a child how to do Karatsuba multiplication, he's violated patent law. There's a reason why patents on abstract ideas are not permitted, and that's because they imply you can violate them by simply thinking.

I think you've made some decent points that I can't disagree with (because frankly, I agree with you. Software patents are ultimately a bad idea. At least in their current form with 20 year monopolies and so forth).

However... I still don't like the style of your argument, but at least its somewhat connected to reality now. The only part of your post that was decently done is IMO the HTML5 argument.

Here are a few counterpoints:
* Software companies are always going to be assholes who abuse copyright and patent law. As this topic has shown... Oracle can claim copyright over an API to an open source language. Demonstrating that companies have been jackasses doesn't really say anything. Laws will never be perfectly written, and legal teams will always find ways to twist a good idea into one that they abuse.

* Patented network protocols existed. IIRC, ATM for example is a protocol that was encumbered by lots of patents during the rise of TCP/IP. But as the core of TCP/IP was patent free (designed by US Government and thus couldn't be patented)... many people decided to implement it and it became the defacto-standard of the internet. Similarly, OpenID demonstrates the ability for the software community to use software patents in a pure defensive manner with the non-assertion covenants.

And because software companies are always going to abuse IP laws, that means we should provide the means for them to do so? Remember, before this portion of the suit, Oracle had been suing Google for violating seven patents, but dropped five of them since the PTO ruled those as not patentable. In the meantime, there's been no evidence suggesting having software patents improves the development of technology, and plenty showing its impedance.

More importantly, remember that under current law/unoverturned legal rulings by the Supreme Court, software patents are (supposed to be) null. So if laws are never going to be written perfectly, why not just enforce the laws the way the highest court in the land has declared?

As for you examples of ATM and OpenID, I fail to see how these are ideal. In fact, "defensive patents" are the key sign of everything that is wrong with software patents. Nothing wrong with them, but the fact that they are necessary. Why should, for example, Google have to spend huge amounts of money acquiring "defensive patents" instead of spending that money toward R&D?

Indeed, in the areas where it matters... companies (such as Microsoft) are finding that non-assertion covenants and open source software are mutually beneficial to themselves and the world in general. So throwing all this doom and gloom around because of software patents is pure FUD IMO. There are clear cut cases where software patents encumber the general public... I think HTML5 is a great example in this regards. But to extrapolate the argument and claim that Linux wouldn't have been possible??!?!

There's only so far you can push a hypothetical.

Considering Microsoft's Linux patent-protection list, and SCO's lawsuits, yes, I'd say the Linux kernel would not be possible. And neither would GNU. Why would you think they would?

Right and you can patent software because its a creation, not a fact. When you are writing code, you aren't just writing one algorithm. Your using many, and for that matter your often combining it with specific data. That's creation, not fact.

No. The software itself is covered by copyright law. The algorithms it employs are covered under patents. All algorithms can be written as formulas in lamda calculus, and are thus facts. Which means, it is not only illegal to copy source code/binaries for software, it would also be illegal to replicate functionality, even if you don't know you replicated something's functionality (hence, patent trolls). Under copyright law, GNU/Linux would be illegal if it copied source code from Unix, but under patent law, GNU/Linux would be illegal for merely attempting replicate Unix's functionality.

[sourmìlk]

No. Mathematical equations are descriptive, the lamda calculus is not. A program written using the lamda calculus cannot be derived, it must be invented. Thus it isn't a fact, it's a creation. (x) -> x^2 is an instruction, not a fact.

[Ghostbear]

No. Mathematical equations are descriptive, the lamda calculus is not. A program written using the lamda calculus cannot be derived, it must be invented. Thus it isn't a fact, it's a creation. (x) -> x^2 is an instruction, not a fact.

All parts of math are invented or derived from those inventions, not just lamba calculus. A written program itself is dependent entirely on binary logic as well, which had to be invented.

[sourmìlk]

I think you've misunderstood what I said. Mathematical equations are descriptive. They are facts because they describe tautological relationships. Programs are not facts because they list instructions. (x) -> x^2 is not a tautological relationship. SET A, 0x8000 is not a tautological relationship.

[Ghostbear]

Ah, I think I missed that cause I didn't see that you were responding to a specific comment using lamba calculus as well.

[sourmìlk]

Sorry, I thought you were somebody else, so my response was based on you making a point that you probably weren't making.

[Griffin]

Regardless of all what the fuck the last two dozen posts were about, does anyone think that it's legitimate to patent and/or copyright APIs meant for public consumption?

I mean, maybe I'm missing something here, but where is the compelling public interest in allowing this, particular, type of patent (or copyright, somehow) to exist?

[Dark567]

Regardless of all what the fuck the last two dozen posts were about, does anyone think that it's legitimate to patent and/or copyright APIs meant for public consumption?

I have mixed feelings. Basically, should I be allowed to reimplement the Java libraries using Sun's current API. What happens if I just change the names of the methods? Presumably that would be enough to avoid copyright violations but not patents(And would be a pain in the ass to have anyone building on the API to refactor their code).

[KnightExemplar]

Regardless of all what the fuck the last two dozen posts were about,

Sorry about that. Arariel: maybe in another topic. Actually, it probably would be a good idea to move the whole patent thing to Serious Business if you're interested.

For the purpose of this topic... I'd rather get back to Oracle vs Google.

does anyone think that it's legitimate to patent and/or copyright APIs meant for public consumption?

I mean, maybe I'm missing something here, but where is the compelling public interest in allowing this, particular, type of patent (or copyright, somehow) to exist?

For copyright, its arguably an important test of the GPL's "viral" clause. If you're a strong copyleft activist, you might be taking Oracle's side on the issue (maybe??) because its important for Java to remain a viral license.

Part of the reason that Google created DalvikVM was to avoid GPL. They recreated parts of Java and then licensed it with a non-copyleft license (the Apache license to be precise). For example, if Google licensed everything as GPL (which would force all App writers to release their code as GPL as well), then they would have been compliant with the Java GPL license.

[Rysto]

For copyright, its arguably an important test of the GPL's "viral" clause. If you're a strong copyleft activist, you might be taking Oracle's side on the issue (maybe??) because its important for Java to remain a viral license

If your a strong copyleft activist I certainly hope that you're not taking Oracle's side, because it Oracle is correct than much of the software distributed by the FSF, as well as the Linux kernel, are all infringing on Unix copyrights.

Interfaces are not copyrightable. This has been settled law for decades now.