Software patents should be abolished(cdixon.org)
cdixon.org
Software patents should be abolished
http://www.cdixon.org/?p=1090
9 comments
> Imagine now that people had secretly claimed land, and done so in a way that even they didn't know what the boundary was. "That land over there" or "the thickest part of the forest" and yet these secret, vague claims could be legally enforced. Imagine investing money to build on that land, trying your best to avoid accidentally stepping on anyone's property.
That's a nice analogy. I'm writing a book about intellectual property and the Pirate party movement; mind if I use your analogy?
That's a nice analogy. I'm writing a book about intellectual property and the Pirate party movement; mind if I use your analogy?
Feel free.
I already patented it -
1, a system of using a simple picture to represent a complex idea
2 a specific inplementation of 1 where the picture involves green grass.
2 a specific inplementation of 1 where the picture involves green grass.
For those who haven't read Paul Graham's essay on software patents I highly recommend it. It's wonderfully insightful, and gave me perspectives I never considered.
http://paulgraham.com/softwarepatents.html
http://paulgraham.com/softwarepatents.html
And I am absolutely sure that if we got rid of patents tomorrow innovation wouldn’t be reduced at all
I don't think this is true at all. While most of the bad press about software patents comes from frivolous patents and companies that troll with their portfolios, that doesn't mean there are no legitimate software patents.
Consider the patents held by MPEG LA (MPEG codecs, Firewire, H.264, etc.) These patents represent years of R&D work by their member companies and describe inventions that are non-frivolous. Surely these companies deserve some kind of protection for their work to encourage them to continue to invest in such research and publish it.
One of the difficulties in the software patents issue is where to draw the line---how to define what is a "software patent" and what defines the patentability of inventions in that class. The debate on software patents is shifting more towards this issue, and articles that make sweeping statements like "software/internet/hardware patents have no benefit to society and should be abolished" add nothing to it.
I don't think this is true at all. While most of the bad press about software patents comes from frivolous patents and companies that troll with their portfolios, that doesn't mean there are no legitimate software patents.
Consider the patents held by MPEG LA (MPEG codecs, Firewire, H.264, etc.) These patents represent years of R&D work by their member companies and describe inventions that are non-frivolous. Surely these companies deserve some kind of protection for their work to encourage them to continue to invest in such research and publish it.
One of the difficulties in the software patents issue is where to draw the line---how to define what is a "software patent" and what defines the patentability of inventions in that class. The debate on software patents is shifting more towards this issue, and articles that make sweeping statements like "software/internet/hardware patents have no benefit to society and should be abolished" add nothing to it.
> Consider the patents held by MPEG LA (MPEG codecs, Firewire, H.264, etc.) These patents represent years of R&D work by their member companies and describe inventions that are non-frivolous. Surely these companies deserve some kind of protection for their work to encourage them to continue to invest in such research and publish it.
There are two separate issues here, (1) whether patents are necessary to encourage innovation, and (2) whether companies "deserve" patents.
To answer the first question, the existance of non-patent-emcumbered codecs, such as Ogg Vorbis and Ogg Theora, demonstrates that software patents don't need to exist for these things to exist. (I sometimes half-jokingly refer to arguments of the sort "but without copyright/patent incentives, X wouldn't exist" as "the Linux-doesn't-exist theory of intellectual property").
To answer the second question, why would companies deserve patents? So they can make bigger profits, presumably? But why is that a good thing? The purpose of the economy isn't for companies to make profits, it's for the economy to make things people need and want. Companies making big profits is a good thing if and only if it leads to the economy making things people need and want -- so if people want to make money they have to make something people want, not game the patent system to no-one else's advantage. And since software patents aren't necessary to produce good software, and since they in fact overall harm progress in software rather than help it, in the interests of the economy they shouldn't exist.
There are two separate issues here, (1) whether patents are necessary to encourage innovation, and (2) whether companies "deserve" patents.
To answer the first question, the existance of non-patent-emcumbered codecs, such as Ogg Vorbis and Ogg Theora, demonstrates that software patents don't need to exist for these things to exist. (I sometimes half-jokingly refer to arguments of the sort "but without copyright/patent incentives, X wouldn't exist" as "the Linux-doesn't-exist theory of intellectual property").
To answer the second question, why would companies deserve patents? So they can make bigger profits, presumably? But why is that a good thing? The purpose of the economy isn't for companies to make profits, it's for the economy to make things people need and want. Companies making big profits is a good thing if and only if it leads to the economy making things people need and want -- so if people want to make money they have to make something people want, not game the patent system to no-one else's advantage. And since software patents aren't necessary to produce good software, and since they in fact overall harm progress in software rather than help it, in the interests of the economy they shouldn't exist.
> To answer the first question, the existance of non-patent-emcumbered codecs, such as Ogg Vorbis and Ogg Theora, demonstrates that software patents don't need to exist for these things to exist.
This is an extremely bad example. Ogg Vorbis came years after rival standards (such as mp3). It is not state of the art at all.
The development of the basis of Theora (VP3) was done by a company (On2) that protected it with patents. VP3 was a very old decoder that On2 released (all of their current state of the art decoders are protected by patents).
There are actually patents on Theora but On2 granted a licence for its use (from WP http://en.wikipedia.org/wiki/Theora):
> On2 also made an irrevocable, royalty-free license grant for any patent claims it might have over the software and any derivatives
> To answer the second question, why would companies deserve patents?
The same reason as copyright – to protect innovation.
This is an extremely bad example. Ogg Vorbis came years after rival standards (such as mp3). It is not state of the art at all.
The development of the basis of Theora (VP3) was done by a company (On2) that protected it with patents. VP3 was a very old decoder that On2 released (all of their current state of the art decoders are protected by patents).
There are actually patents on Theora but On2 granted a licence for its use (from WP http://en.wikipedia.org/wiki/Theora):
> On2 also made an irrevocable, royalty-free license grant for any patent claims it might have over the software and any derivatives
> To answer the second question, why would companies deserve patents?
The same reason as copyright – to protect innovation.
> The same reason as copyright – to protect innovation.
Does the innovation need protecting? And if so then protecting from whom? Would copy-pasters do more damage to innovation than lawyers do?
Does the innovation need protecting? And if so then protecting from whom? Would copy-pasters do more damage to innovation than lawyers do?
"but without copyright/patent incentives, X wouldn't exist" as "the Linux-doesn't-exist theory of intellectual property"
Linux was based on Unix. Unix was developed and marketed by a company that enforced its intellectual property in it. Would they have funded both activities without any way to own the results (i.e. copyright)? If Unix did not exist, would Linux?
An aside on patents: regular expression matching using DFAs was patented by Ken Thompson (for his employer) http://www.wikipatents.com/3568156.html
Linux was based on Unix. Unix was developed and marketed by a company that enforced its intellectual property in it. Would they have funded both activities without any way to own the results (i.e. copyright)? If Unix did not exist, would Linux?
An aside on patents: regular expression matching using DFAs was patented by Ken Thompson (for his employer) http://www.wikipatents.com/3568156.html
Unix was developed as a side result of a couple hackers wanting to play a game that wouldn't run well on the systems available, so they ported it to a new system and then wrote an OS for that system as well (because hey, why not). At the time, they were being paid to pretty much screw around with technology by virtue of working at a large research facility funded by a corporation with a government-granted monopoly.
Unix wasn't even officially acknowledged by Bell Labs until a couple years after it was developed, and wasn't marketed for use outside the labs until a few years later.
So, no, I don't think the early development of Unix had much to do with the ability to enforce intellectual property.
Unix wasn't even officially acknowledged by Bell Labs until a couple years after it was developed, and wasn't marketed for use outside the labs until a few years later.
So, no, I don't think the early development of Unix had much to do with the ability to enforce intellectual property.
The comment was about development, not just "early development"; and it was about development and marketing, not just development.
The value of an operating system derives primarily from the applications that run on it, and this is very much affected by the business case, which includes ongoing development and marketing. Unix had that.
Even for the early development, I wonder if Thompson and Richie (and the rest of the team) would have been employed to create software in general in the first place, if their employer would not have some form of ownership in it... Doesn't sound like how a monopolist thinks to me.
This is not to deride linux (which is great; I'm using it now), but to note the role that copyright plays in securing investment in a platform. Attracting users to a platform requires a lot of investment; it's much, much more than just building the platform technology.
I think a much better example to support your case is Python (also Perl/Ruby/PHP/emacs). Very successful; a platform; and not a replication of a commercial platform. In contrast, many successful open source projects follow commercial projects. Why is this? There's pre-existing demand for it; and having a technical guide and an existing user-base makes it comparatively straightforward and reduces the risk (i.e. compared with starting something entirely new).
I speak as someone who has replicated an existing product; and is now creating something entirely new. I find the latter tremendously more demanding. There's no pre-existng technical guide, and there's pre-existing demand for this particular approach.
The value of an operating system derives primarily from the applications that run on it, and this is very much affected by the business case, which includes ongoing development and marketing. Unix had that.
Even for the early development, I wonder if Thompson and Richie (and the rest of the team) would have been employed to create software in general in the first place, if their employer would not have some form of ownership in it... Doesn't sound like how a monopolist thinks to me.
This is not to deride linux (which is great; I'm using it now), but to note the role that copyright plays in securing investment in a platform. Attracting users to a platform requires a lot of investment; it's much, much more than just building the platform technology.
I think a much better example to support your case is Python (also Perl/Ruby/PHP/emacs). Very successful; a platform; and not a replication of a commercial platform. In contrast, many successful open source projects follow commercial projects. Why is this? There's pre-existing demand for it; and having a technical guide and an existing user-base makes it comparatively straightforward and reduces the risk (i.e. compared with starting something entirely new).
I speak as someone who has replicated an existing product; and is now creating something entirely new. I find the latter tremendously more demanding. There's no pre-existng technical guide, and there's pre-existing demand for this particular approach.
Consider the patents held by MPEG LA (MPEG codecs, Firewire, H.264, etc.) These patents represent years of R&D work by their member companies and describe inventions that are non-frivolous. Surely these companies deserve some kind of protection for their work to encourage them to continue to invest in such research and publish it.
Let's take the case of H.264. Why is copyright not sufficient enough to protect their innovations to whatever degree of protection they choose (using licenses like GPL, MIT, or whatever they choose)?
The problem is there isn't just a patent on H.264 there are hundreds of patents on various aspects of it. This isn't done to protect the years of R&D it is done to create a patent minefield for any future competitors.
They don't let other more well understood industries operate in this way. Ford can't patent round rubber tires, hydraulic power steering, and using friction for braking. Good luck making a car now, but this is exactly the challenge a next-gen video codec would face.
Let's take the case of H.264. Why is copyright not sufficient enough to protect their innovations to whatever degree of protection they choose (using licenses like GPL, MIT, or whatever they choose)?
The problem is there isn't just a patent on H.264 there are hundreds of patents on various aspects of it. This isn't done to protect the years of R&D it is done to create a patent minefield for any future competitors.
They don't let other more well understood industries operate in this way. Ford can't patent round rubber tires, hydraulic power steering, and using friction for braking. Good luck making a car now, but this is exactly the challenge a next-gen video codec would face.
> Let's take the case of H.264. Why is copyright not sufficient enough to protect their innovations to whatever degree of protection they choose (using licenses like GPL, MIT, or whatever they choose)?
Someone "skilled in the art" can take the descriptions of H.264 and produce a coder and a decoder without copying anything. Since there's no copying, copyright protection doesn't come into play.
Given that, how are they going to "protect their innovations"? Are you suggesting that they keep the description as a trade secret and allow access only under license? (They can't publish and say "you can't use this description unless you pay us money" - copyright doesn't work that way.)
Someone "skilled in the art" can take the descriptions of H.264 and produce a coder and a decoder without copying anything. Since there's no copying, copyright protection doesn't come into play.
Given that, how are they going to "protect their innovations"? Are you suggesting that they keep the description as a trade secret and allow access only under license? (They can't publish and say "you can't use this description unless you pay us money" - copyright doesn't work that way.)
Someone "skilled in the art" can take the descriptions of H.264 and produce a coder and a decoder without copying anything.
Isn't this exactly the same way people get around patents now? If I patent a Jack in the Box made using a spring and a box you could come along and make one using compressed air and a box.
The amount of protection H.264 would receive is directly related to how broadly derivative work is applied. Even at its broadest interpretation and everything that decoded or encode H.264 had to be licensed from them, at least it wouldn't stop other encodings from using almost all modern strategies to video encoding.
Isn't this exactly the same way people get around patents now? If I patent a Jack in the Box made using a spring and a box you could come along and make one using compressed air and a box.
The amount of protection H.264 would receive is directly related to how broadly derivative work is applied. Even at its broadest interpretation and everything that decoded or encode H.264 had to be licensed from them, at least it wouldn't stop other encodings from using almost all modern strategies to video encoding.
> Isn't this exactly the same way people get around patents now?
No, but the confusion comes from my ambiguous use of "copy".
Patents cover the use of mechanism X to accomplish purpose Y. Copying, for copyright purposes, is copying bytes (ignoring the whole look and feel thing). It's reasonably easy to use software mechanism X to accomplish purpose Y without copying bytes.
You seem to think that purpose should be limited to "H.264". If, for the purposes of argument, the relevant mechanism actually is useful for many other kinds of video encoding, it's unclear why that limitation makes sense. If the mechamism is worthy of protection for one protocol, why should it be unprotected when used in another?
Note that protocol ownership is different from mechanism ownership. The former is more akin to trademark, as in "real H.264" as opposed to some knockoff. (Coke vs Pepsi as it were.)
No, but the confusion comes from my ambiguous use of "copy".
Patents cover the use of mechanism X to accomplish purpose Y. Copying, for copyright purposes, is copying bytes (ignoring the whole look and feel thing). It's reasonably easy to use software mechanism X to accomplish purpose Y without copying bytes.
You seem to think that purpose should be limited to "H.264". If, for the purposes of argument, the relevant mechanism actually is useful for many other kinds of video encoding, it's unclear why that limitation makes sense. If the mechamism is worthy of protection for one protocol, why should it be unprotected when used in another?
Note that protocol ownership is different from mechanism ownership. The former is more akin to trademark, as in "real H.264" as opposed to some knockoff. (Coke vs Pepsi as it were.)
Sweeping statement do have merits.
It might be helpful for you to read the book "Against Intellectual Monopoly". http://www.dklevine.com/general/intellectual/againstfinal.ht...
It talked about many historical cases where the patent and copyright system failed and as well historical evidence of free market providing immense innovative pressure.
Just because an argument is in the extreme doesn't mean it have any merits whatsoever. Sometime, it have lot of merits.
For example, James Watt use the patent system to force his competitor away from the market, instead of using his time to invent. When his patent expired, Watt got rich anyway!
The early American book publishing industry benefited immensely from book piracy of British authors. British authors got paid anyway, sometime even better than the royalty rate they receive at home!
The movie industry moved to California to escape Thomas Edison' movie patent.
It might be helpful for you to read the book "Against Intellectual Monopoly". http://www.dklevine.com/general/intellectual/againstfinal.ht...
It talked about many historical cases where the patent and copyright system failed and as well historical evidence of free market providing immense innovative pressure.
Just because an argument is in the extreme doesn't mean it have any merits whatsoever. Sometime, it have lot of merits.
For example, James Watt use the patent system to force his competitor away from the market, instead of using his time to invent. When his patent expired, Watt got rich anyway!
The early American book publishing industry benefited immensely from book piracy of British authors. British authors got paid anyway, sometime even better than the royalty rate they receive at home!
The movie industry moved to California to escape Thomas Edison' movie patent.
Just because an argument is in the extreme doesn't mean it have any merits whatsoever.
Perhaps, but my specific problems with this article is that I don't think it advances the argument in any meaningful way, and it doesn't provide evidence for its conclusion---other than a single example of a patent troll, and the claim that an NYT story about another company promotes an "urban myth".
The general problem with sweeping statements is that they very susceptible to attack because they lack the precision you need to build a proper argument.
Perhaps, but my specific problems with this article is that I don't think it advances the argument in any meaningful way, and it doesn't provide evidence for its conclusion---other than a single example of a patent troll, and the claim that an NYT story about another company promotes an "urban myth".
The general problem with sweeping statements is that they very susceptible to attack because they lack the precision you need to build a proper argument.
I would add that a sweeping statements will need a book to build a proper argument, not just a two page article.
Actually, stratch my statement about Watt getting rich. The source I read indicate that Watt was able to kept the price of his steam engine up by the virtue of being first even after the patent expiration. It is still true that he use legal forces to keep his many inventive competitors at bay. http://mises.org/story/3280
So he still have a competitive advantage after the fact that the patent expired. It is what people refer to as "first mover advantage".
Actually, stratch my statement about Watt getting rich. The source I read indicate that Watt was able to kept the price of his steam engine up by the virtue of being first even after the patent expiration. It is still true that he use legal forces to keep his many inventive competitors at bay. http://mises.org/story/3280
So he still have a competitive advantage after the fact that the patent expired. It is what people refer to as "first mover advantage".
I agree that the system is a mess, but defining exactly what "software" patents are is difficult. In machine inventions, the actual invention is increasingly happening in the software.
Patenting an organic molecule or a gene is a different, and also flawed case. But most machines, business plans and other inventions could be reduced to software. As 3d printing becomes more common, this will get more and more common.
Algorithms currently aren't patentable, but due to semantic manipulation of the meaning of "algorithm", people are patenting algorithms by calling them a "system that does x". If software patents were abolished, what would prevent applicants from using the exact same sort of strategy to get around them? Rather than patenting a piece of code, people would file a patent for a machine that does everything the code in question does.
Patenting an organic molecule or a gene is a different, and also flawed case. But most machines, business plans and other inventions could be reduced to software. As 3d printing becomes more common, this will get more and more common.
Algorithms currently aren't patentable, but due to semantic manipulation of the meaning of "algorithm", people are patenting algorithms by calling them a "system that does x". If software patents were abolished, what would prevent applicants from using the exact same sort of strategy to get around them? Rather than patenting a piece of code, people would file a patent for a machine that does everything the code in question does.
Innovation happens spontaneously and you'll get the irresistible urge to implement it. If the promise of a patent is your only incentive then you probably won't do a good job at it.
However there are people who may not be so ingenious, and pour nothing but a lot of time and effort into something. These may be the ones most interested in patents. However in a patentless world they still have the advantage of understanding their work the best, and that gives them an edge.
That being said, I believe anyone capable of understanding another person's complicated work has the right to duplicate it. In the end whoever can innovate faster, whoever is better at making it better, and whoever understands the concepts best will come out on top.
However there are people who may not be so ingenious, and pour nothing but a lot of time and effort into something. These may be the ones most interested in patents. However in a patentless world they still have the advantage of understanding their work the best, and that gives them an edge.
That being said, I believe anyone capable of understanding another person's complicated work has the right to duplicate it. In the end whoever can innovate faster, whoever is better at making it better, and whoever understands the concepts best will come out on top.
Is this a reasonable middle ground? All software covered by a patent must have the source publicly available.
All software that someone wants to be protected by copyright should have source code publicly available.
That would be exact equivalent of standard patents. If you publish all the details of your invention you get temporary monopoly on manufacturing it's copies. If not then you get no protection.
That would be exact equivalent of standard patents. If you publish all the details of your invention you get temporary monopoly on manufacturing it's copies. If not then you get no protection.
I'd just like to see the period of time drastically reduced for "inventions" in domains that move faster.
Any software patent should last no longer than a year, period.
And as long as I'm writing the law, anything patented by Intellectual Ventures becomes public domain in six days.
Any software patent should last no longer than a year, period.
And as long as I'm writing the law, anything patented by Intellectual Ventures becomes public domain in six days.
It seems to me that with software what the sftware does is usually just solve a problem. Patents are to protect the method of solving a problem, in RIM's case mobile email wasn't a method of solving a problem, the lack of mobile email was the problem. In software the method of solving the problem is the source code, and that should be protected, which it is, by copyright law.
Just a thought. Why not set a time limit on patents, like copyright? You have a patent, its valid only for say 30 or 50 years, then its public domain.
or something like this already exists?
or something like this already exists?
Time limit already exists. The thing with software is that its associated technology will most likely expire long before its patent does.
A book I read recently claimed it all came down how precisely and clearly the patent is defined and delineated.
Imagine you moved into a lush valley with a new town in it to do business. If you found all the best spots had been taken and fenced in then that would kind of suck.
Imagine now that people had secretly claimed land, and done so in a way that even they didn't know what the boundary was. "That land over there" or "the thickest part of the forest" and yet these secret, vague claims could be legally enforced. Imagine investing money to build on that land, trying your best to avoid accidentally stepping on anyone's property.
The first case is chemical compounds, it's either got those atoms in that arrangement or it doesn't. The latter is "selling shoes over the internet" and other alleged software patents.
Though having said that I think patents are still broken and lead to the promotion of some very strange behaviour in the medical field. But in software they are appallingly bad.