When you patent a screwdriver/computer, you do not patent every possible way in which that screwdriver/computer can be used.
Patent law does not prevent you from doing so. But in practice, you can't, because you are constrained by prior art and expired patents. Which will soon be the case for software as well.
That's an often-heard argument of software patent proponents, but there is not a single guarantee for this. Patent law does not specify that something has to be "truly innovative" in order to be patentable. It just has to be new and non-obvious, and those two requirements combined are not a guarantee for "true innovativeness".
For something to be new and non-obvious, it must be either truly innovative, or else so limited or useless that nobody has bothered to try to figure out how to do it before. A few patents on limited or useless ideas hardly seems likely to do much harm.
And then you still left out the other arguments (algorithms are applicable to an infinite range of problems, and replicated at zero cost with no overhead to which patent license fees could be added).
I thought this was too obviously nonsensical to merit a reply, but since you insist: Algorithms are no more applicable to an infinite range of problems than any other tool. Some are highly specialized for a particular task, others have broad applicability. I suppose that you could say that a screwdriver or a lathe is "applicable to an infinite range of problems," because you can build an infinite number of different things with them. But tools are still patentable.
I think the software patent thing is a tempest in a teapot, and will ultimately take care of itself. A lot of those patents shouldn't have been granted because of prior art, but as some of the more stupid ones get thrown out by the courts, the patent offices will get more sophisticated. People see it as a big problem because the computer industry is still relatively young. And since software patents were allowed only belatedly by the courts, a lot of patents went, not to the inventor, but to the guy who first realized that something that everybody knew how to do was suddenly patentable. But all of those broad, basic patents (the ones that survive the courts) are going to expire, and in a few years it won't be possible to get a software patent unless it is something truly innovative.
The difference is that Apple actually tested out their user-interface features on users, rather relying up some designer's theoretical notions. This led to crucial insights like, "it's much faster to access a menu bar that's always at the top of the screen, rather than one that's at the top of a window." Like many aspects of human engineering, this is the sort of thing that seems counter-intuitive, because at first thought it seems more logical to associate the menu with the thing that it affects.
I'm still installing Apple's updates the moment they come out. I've updated 2 Panthers and one Jaguar (on a beige G3) with this latest update with no problems. So far, I've only had one problem with an Apple update. 10.2.8 knocked out the ethernet connectivity of a dual G4 (out of about 8 machines I installed it on), but there was a non-Apple fix going around by the end of the day, and an official fix a few days later.
You really are completely following the EPO's (and Commissions, and JURI's) interpretation of "computer program as such"
This is the interpretation that US Supreme Court and Federal Circuit Court have imposed on the Patent Office.
And in fact the second argument is right, because the economic rationale behind not granting patents on thoughts applies also here: abstract ideas are, regardless of their applicability to technical problems, produced without experimentation cost
In practice, however, the major costs in many research projects, whether or not experimentation is involved, are the human resources costs, with the experimentation, if any, being a minor element. There is no particular correlation between the commercial value of a patentable idea or implementation and the cost of obtaining it. A software program could well have greater total development cost than a piece of physical equipment.
since between the idea and the application there is no invention, any adequate disclosure of the idea in turing-complete syntax risks to become an act of patent infringement
The mere description of a method (or listing of the code) does not constitute infringement; it must be used to do something.
This is not about whether software can be original or not. A literary work can also be very original, both in the way it is written as well as the underlying story, subplots and used ideas (and the paper it is printed on is of much lower importance). Yet, society has decided not to award patents on nice story or plot ideas, regardless of how revolutionary and original they are.
No, because the other key element in addition to originality required for a patentable idea is that it accomplishes something of real world relevance. So a purely mathematical algorithm, no matter how original, is not considered patentable, but a use of that algorithm to do something of real-world value is patentable.
I think that almost everybody (except, perhaps, the very rich) is bothered by the power advantage that the wealthy, whether individuals or corporations, enjoy when it comes to civil litigation. I just haven't heard any plausible solutions. In some types of litigation, the less well-to-do are able to get around this problem by employing lawyers on a contingency basis. But this has its own set of abuses.
The difference comes when some thing is created that is new and non-trivial, rather than re-using something we already have.
And this is an issue for all patents, it is in no way unique for software. For instance, it is possible to obtain a "use" patent covering a new use of something that already exists (e.g. a drug).
But the auto parts patent doesn't prevent you from making your own part which does the same thing but in a slightly different way, or even the same thing but in a totally different way.
That depends upon how well the patent was written. Generally, a patent will be written to cover every way of accomplishing a particular task that the inventor and his lawyer can think of. So it may well turn out that your "different way" is also covered by the patent on the part.
The real difference between invention and software patent is that software patents are mostly methods. An invention involves creating something non-trivial, where as methods only tell you how to use something already invented in a different manner.
Fundamentally, all inventions are methods--methods of accomplishing a task, where part of the method may involve instructions for constructing a device for that purpose.
1: The life cycle of a patent is about 4-5 times longer than the life cycle of software.
Or an automobile. But auto parts are still patentable.
2: Patent language does not fit well with software language. Patent language often fails to cover a specific algorithm in the lawyer's attempt to cover everything possible
It's legalese. It doesn't fit well with any other real-world language, but it's actually closer to a computer language than anything else, because it attempts to deal with every possible case.
4: The USPTO is unable to handle the current patent process, empowering them with even more oversight is certainly not what I'd call punishment for its failure to operate properly.
The patent office is only part of the system, and is really only a first screen for validity. The courts are another part.
5: In the event of failure in the above mentioned government organization to perform its duties properly, the cost of undoing the patent is footed entirely by the victims.
An unfortunate consequence of free enterprise is that protecting your rights can be expensive. If you have a valid patent and somebody violates it, it can cost you a bundle to defend your rights, with no guarantee of success. And a big company can drag out the litigation until your patent is worth little.
However, software is already abstract. It's just a set of mental rules, unlike a chemical reaction.
A chemical reaction sequence can be pretty abstract, basically little more than an algorithm. The originality typically lies in the sequence of steps, not in the wet details of what kind of solvent or container, much as the originality of a software program lies in the sequence of steps, not the details of the computer hardware.
If Newton had invented calculus in the 21st century he would have patented it.
More likely, he would have done just what he did; kept it to himself as trade secret until somebody else came up with it and he wanted to grab a share of the credit. One of the rationales for patents is to encourage inventors to make public their methods, rather than keeping them secret.
Moreover the formulas are a long way from what most undergraduate mathematics students can handle. Hell the notation alone will confuse most people.
Special relativity is within the grasp of many high school students and most undergraduates. All you really need is a little trig and basic calculus, and the notation is fairly simple. General relativity is another matter.
a method for applyuign paint to a wall with a brush doenst deserve a patent. and yet it wasnt been done before. Fire also hasnt been patented. or the wheel. both are useful, and in teh case of fire, making fire from non fire is decidedly non trivial.
And if they had been, the patents would have expired a long time ago.
Yes, and that's why software patents are so completely different from other patents. Normally, you get a patent on the implementation of some solution to a problem (a machine, a way to perform a chemical reaction, the use a certain chemical against a particular organism,...). You do not get patents on the generic process behind those innovations.
This isn't really special to software patents. Even chemical or pharmaceutical patents are typically written to be as generic as possible, covering every conceivable way of achieving a particular end.
Despite what seems to have become a widely held belief that we can learn as much from automated probes as from manned missions, it doesn't seem to have worked out that well in practice. Viking had a set of experiments that was supposed to definitively detect whether life was present. But when some of the experiments came out positive, they ended up being rejected, because researchers at home came up with nonbiological explanations. Unfortunately, there was nobody on site to do a follow-up experiment to really answer the question. Now we've had a long string of failed probes.
Perhaps all Spirit really needs is somebody to give it a little kick.
This is public property used to conduct the public's business.
So is President Bush's desk drawer in the White House. Are you saying that the Democrats are entitled to go through it and release the juicy bits to the media if somebody happens to leave the door unlocked?
A better example was when the Clinton administration, apparently as the result of a clerical error, requested FBI background files on some Republicans. There was a huge media uproar, even though there was never any indication that the administration so much as looked at the files, much less made any use of them.
It would seem that the Democrats did not take reasonable steps to ensure confidentiality, so can they really claim that the Republicans broke into thier system and stole the documents?
So if I forget to lock my door, and somebody comes in and takes my stereo, I wasn't really robbed?
I've always found particularly appealing a hypothesis that Francis Crick proposed quite a few years back. His idea was that while you are asleep, a random stimulus generator in your brain is stimulating different regions of your cortex, and then damping down any circuits that resonate. So when you wake up, all of the unproductive ideas that were perseverating about before you went to sleep are suppressed, and you now can come up with new ideas. The nice thing about the idea is that it even explains why you can't remember your dreams--they are the stuff that your brain is trying to forget!
I can't imagine that little bit of lost heat was the difference between life and death for anyone. But, the fact that we have evolved the feature suggests it was.
Probably preserving core temperature is more important that protecting the fingers and toes--at least you then have a chance of keeping going long enough to find shelter. Birds have a more clever countercurrent heat exchanger solution. The arteries and veins run together, so that the venous return is warmed by the arteries, reducing loss of heat through the featherless feet.
Run a god damn wire you dumb shit! You've thrown away several HUNDRED dollars so you can be cool and have wireless to a STATIONARY laptop.
Not everybody has a house in which it is easy to run wire through the walls, or likes seeing wires running along the baseboards and doorjambs all over the place. Once you factor in the value of the time to run the wire, he may well be saving money.
Slashdot Mirror
When you patent a screwdriver/computer, you do not patent every possible way in which that screwdriver/computer can be used.
Patent law does not prevent you from doing so. But in practice, you can't, because you are constrained by prior art and expired patents. Which will soon be the case for software as well.
That's an often-heard argument of software patent proponents, but there is not a single guarantee for this. Patent law does not specify that something has to be "truly innovative" in order to be patentable. It just has to be new and non-obvious, and those two requirements combined are not a guarantee for "true innovativeness".
For something to be new and non-obvious, it must be either truly innovative, or else so limited or useless that nobody has bothered to try to figure out how to do it before. A few patents on limited or useless ideas hardly seems likely to do much harm.
And then you still left out the other arguments (algorithms are applicable to an infinite range of problems, and replicated at zero cost with no overhead to which patent license fees could be added).
I thought this was too obviously nonsensical to merit a reply, but since you insist: Algorithms are no more applicable to an infinite range of problems than any other tool. Some are highly specialized for a particular task, others have broad applicability. I suppose that you could say that a screwdriver or a lathe is "applicable to an infinite range of problems," because you can build an infinite number of different things with them. But tools are still patentable.
I think the software patent thing is a tempest in a teapot, and will ultimately take care of itself. A lot of those patents shouldn't have been granted because of prior art, but as some of the more stupid ones get thrown out by the courts, the patent offices will get more sophisticated. People see it as a big problem because the computer industry is still relatively young. And since software patents were allowed only belatedly by the courts, a lot of patents went, not to the inventor, but to the guy who first realized that something that everybody knew how to do was suddenly patentable. But all of those broad, basic patents (the ones that survive the courts) are going to expire, and in a few years it won't be possible to get a software patent unless it is something truly innovative.
The difference is that Apple actually tested out their user-interface features on users, rather relying up some designer's theoretical notions. This led to crucial insights like, "it's much faster to access a menu bar that's always at the top of the screen, rather than one that's at the top of a window." Like many aspects of human engineering, this is the sort of thing that seems counter-intuitive, because at first thought it seems more logical to associate the menu with the thing that it affects.
I'm still installing Apple's updates the moment they come out. I've updated 2 Panthers and one Jaguar (on a beige G3) with this latest update with no problems. So far, I've only had one problem with an Apple update. 10.2.8 knocked out the ethernet connectivity of a dual G4 (out of about 8 machines I installed it on), but there was a non-Apple fix going around by the end of the day, and an official fix a few days later.
Thats funny, I don't see many eniacs around, yet I see a dozen cars or so from the 50s-70s every day on the streets.
And many modern computers are running operating systems containing code of similar vintage.
You really are completely following the EPO's (and Commissions, and JURI's) interpretation of "computer program as such"
This is the interpretation that US Supreme Court and Federal Circuit Court have imposed on the Patent Office.
And in fact the second argument is right, because the economic rationale behind not granting patents on thoughts applies also here: abstract ideas are, regardless of their applicability to technical problems, produced without experimentation cost
In practice, however, the major costs in many research projects, whether or not experimentation is involved, are the human resources costs, with the experimentation, if any, being a minor element. There is no particular correlation between the commercial value of a patentable idea or implementation and the cost of obtaining it. A software program could well have greater total development cost than a piece of physical equipment.
since between the idea and the application there is no invention, any adequate disclosure of the idea in turing-complete syntax risks to become an act of patent infringement
The mere description of a method (or listing of the code) does not constitute infringement; it must be used to do something.
This is not about whether software can be original or not. A literary work can also be very original, both in the way it is written as well as the underlying story, subplots and used ideas (and the paper it is printed on is of much lower importance). Yet, society has decided not to award patents on nice story or plot ideas, regardless of how revolutionary and original they are.
No, because the other key element in addition to originality required for a patentable idea is that it accomplishes something of real world relevance. So a purely mathematical algorithm, no matter how original, is not considered patentable, but a use of that algorithm to do something of real-world value is patentable.
Is anyone else bothered by this?
I think that almost everybody (except, perhaps, the very rich) is bothered by the power advantage that the wealthy, whether individuals or corporations, enjoy when it comes to civil litigation. I just haven't heard any plausible solutions. In some types of litigation, the less well-to-do are able to get around this problem by employing lawyers on a contingency basis. But this has its own set of abuses.
The difference comes when some thing is created that is new and non-trivial, rather than re-using something we already have.
And this is an issue for all patents, it is in no way unique for software. For instance, it is possible to obtain a "use" patent covering a new use of something that already exists (e.g. a drug).
But the auto parts patent doesn't prevent you from making your own part which does the same thing but in a slightly different way, or even the same thing but in a totally different way.
That depends upon how well the patent was written. Generally, a patent will be written to cover every way of accomplishing a particular task that the inventor and his lawyer can think of. So it may well turn out that your "different way" is also covered by the patent on the part.
Fundamentally, all inventions are methods--methods of accomplishing a task, where part of the method may involve instructions for constructing a device for that purpose.
1: The life cycle of a patent is about 4-5 times longer than the life cycle of software.
Or an automobile. But auto parts are still patentable.
2: Patent language does not fit well with software language. Patent language often fails to cover a specific algorithm in the lawyer's attempt to cover everything possible
It's legalese. It doesn't fit well with any other real-world language, but it's actually closer to a computer language than anything else, because it attempts to deal with every possible case.
4: The USPTO is unable to handle the current patent process, empowering them with even more oversight is certainly not what I'd call punishment for its failure to operate properly.
The patent office is only part of the system, and is really only a first screen for validity. The courts are another part.
5: In the event of failure in the above mentioned government organization to perform its duties properly, the cost of undoing the patent is footed entirely by the victims.
An unfortunate consequence of free enterprise is that protecting your rights can be expensive. If you have a valid patent and somebody violates it, it can cost you a bundle to defend your rights, with no guarantee of success. And a big company can drag out the litigation until your patent is worth little.
A chemical reaction sequence can be pretty abstract, basically little more than an algorithm. The originality typically lies in the sequence of steps, not in the wet details of what kind of solvent or container, much as the originality of a software program lies in the sequence of steps, not the details of the computer hardware.
If Newton had invented calculus in the 21st century he would have patented it.
More likely, he would have done just what he did; kept it to himself as trade secret until somebody else came up with it and he wanted to grab a share of the credit. One of the rationales for patents is to encourage inventors to make public their methods, rather than keeping them secret.
Moreover the formulas are a long way from what most undergraduate mathematics students can handle. Hell the notation alone will confuse most people.
Special relativity is within the grasp of many high school students and most undergraduates. All you really need is a little trig and basic calculus, and the notation is fairly simple. General relativity is another matter.
a method for applyuign paint to a wall with a brush doenst deserve a patent. and yet it wasnt been done before. Fire also hasnt been patented. or the wheel. both are useful, and in teh case of fire, making fire from non fire is decidedly non trivial.
And if they had been, the patents would have expired a long time ago.
Yes, and that's why software patents are so completely different from other patents. Normally, you get a patent on the implementation of some solution to a problem (a machine, a way to perform a chemical reaction, the use a certain chemical against a particular organism, ...). You do not get patents on the generic process behind those innovations.
This isn't really special to software patents. Even chemical or pharmaceutical patents are typically written to be as generic as possible, covering every conceivable way of achieving a particular end.
Despite what seems to have become a widely held belief that we can learn as much from automated probes as from manned missions, it doesn't seem to have worked out that well in practice. Viking had a set of experiments that was supposed to definitively detect whether life was present. But when some of the experiments came out positive, they ended up being rejected, because researchers at home came up with nonbiological explanations. Unfortunately, there was nobody on site to do a follow-up experiment to really answer the question. Now we've had a long string of failed probes.
Perhaps all Spirit really needs is somebody to give it a little kick.
Don't worry, man, you'll be living there tomorrow.
They've been telling me that for years, but somehow it's still the present...
This is public property used to conduct the public's business.
So is President Bush's desk drawer in the White House. Are you saying that the Democrats are entitled to go through it and release the juicy bits to the media if somebody happens to leave the door unlocked?
A better example was when the Clinton administration, apparently as the result of a clerical error, requested FBI background files on some Republicans. There was a huge media uproar, even though there was never any indication that the administration so much as looked at the files, much less made any use of them.
It would seem that the Democrats did not take reasonable steps to ensure confidentiality, so can they really claim that the Republicans broke into thier system and stole the documents?
So if I forget to lock my door, and somebody comes in and takes my stereo, I wasn't really robbed?
I've always found particularly appealing a hypothesis that Francis Crick proposed quite a few years back. His idea was that while you are asleep, a random stimulus generator in your brain is stimulating different regions of your cortex, and then damping down any circuits that resonate. So when you wake up, all of the unproductive ideas that were perseverating about before you went to sleep are suppressed, and you now can come up with new ideas. The nice thing about the idea is that it even explains why you can't remember your dreams--they are the stuff that your brain is trying to forget!
I can't imagine that little bit of lost heat was the difference between life and death for anyone. But, the fact that we have evolved the feature suggests it was.
Probably preserving core temperature is more important that protecting the fingers and toes--at least you then have a chance of keeping going long enough to find shelter. Birds have a more clever countercurrent heat exchanger solution. The arteries and veins run together, so that the venous return is warmed by the arteries, reducing loss of heat through the featherless feet.
Run a god damn wire you dumb shit! You've thrown away several HUNDRED dollars so you can be cool and have wireless to a STATIONARY laptop.
Not everybody has a house in which it is easy to run wire through the walls, or likes seeing wires running along the baseboards and doorjambs all over the place. Once you factor in the value of the time to run the wire, he may well be saving money.