Slashdot Mirror
Torvalds Joins Anti-Patent Attack
canuck57 sent us a story about Linus
Torvalds has joined the chorus of voices speaking out against software patents. Talks briefly about the recent patent releases by IBM & Sun, and notes that there are 'an estimated 150,000 to 300,000 registered software patents in the U.S. alone.'
Most of their patents are hardware, but they do have some software patents.
They do have a strong reliance on the open source community (Mac OS X contains a ton of open source code, as does Safari).
I'm guessing they will in the next year follow IBM's lead an open up a bit.
As to what they will make available, and what they will not, I really don't have a clue. Any guesses welcome.
I find it ironic that Slashdot is always slamming software patents, when at the same time, in their company's 10Q and 10K statements, they're discussing how they are going to profit from creating and defending software patents. Beautiful.
I don't respond to AC's.
That's the question: when do you say it's too trivial? When do you say it's complex enough to be granted patent protection? It can be determined by the courts, but the patent office makes the initial decision, and if unchallenged, sets a precedent.
A blog like any other.
It looks like the article picked a few lines out where Linus said that many of the software patents that have been issued shouldn't have, and added a lot of filler to make it look like he's saying that software patents are inherently bad.
I don't see anything in what he said that says that software patents shouldn't ever be issued, only that in a lot of cases, they were issued in violation of the USPTO's own rules.
"Joins the attack" is a bit overzealous, to say the least.
500GB of disk, 5TB of transfer, $5.95/mo
Patents would be ok, if they were inventive as required by the law. Unfortunately, there is no way on Earth to measure or judge inventiveness, so all a patent examiner can do is to judge whether the application is novel. Something which any patent application can do as long as it mixes in some new technology, like computers or the internet.
To make things worse, many software patents usually don't come with a useful description of how to actually do stuff, which is sad, since software can be documented by the sourcecode and printed.
I'm still trying to figure out what people mean by 'social skills' here.
Pretty easy, if it does not exist in the patent office, try to push it through and see what happens. About all these companies do anymore. Prior art does not mean squat.
If this trend does not stop there are going to be two groups of software developed. Those corporations of large massive power who have either bought up or squashed with patent lawsuits (and then bought up) and those developers of freeware. What will that mean for consumers? Higher prices or go running for a free alternative.
Who knows, in 20 years, if all the middle competition is destroyed, companies like IBM, Solaris, etc, could go after Linux and GNU because they are cutting into their bottom line.
-- Knowing too much can get you killed, but knowing who knows too much can make you rich.
I mean, this is a political problem and I am not sure how Linus is perceived by politicians (outside Finland where he appears to be well-known).
Trolling using another account since 2005.
Well, thats the whole problem in a nutshell. If only the answer was as simple and short as the question. The whole area is something I didn't take an interest in until I started to realise the consequences.
Now, I take a big interest in it. The fact that law and particularly US law is often so vague and varying from state to state makes it more of a nightmare. One little philosophy I have: If its powered directly by nature, then get your greasy patenting hands off it because it belongs to us all. i.e. windfarms, solar power, quantum mechanics ...
Anyway, anything I ever created, I have included the GPL with it and I always will include either the GPL or something similar and at the very least, compatible.
I think the best idea would be to make bad patents extremely costly for the holder to contest. Best would be if it so costly that people can do bounty hunting on bad patents. Basically if a patent is brought to court then if the holder loses he/she/it must pay really hefty charges including court fees, fine depending on the time since the patent was filed to the time when it was challenged.
Also fine must depend on the income of the holder in question. In addition a fine if the patent was done in bad faith like the Rambus patent. Make it so costly that a person or company will think many times before patenting anything, and -- spoiler alert -- after Maggie (Hilary Swank) is rendered quadriplegic in a fight, Frankie (Clint Eastwood) hesitates before agreeing to help her commit suicide.
Of course there will have to be an exemption for the cases of prior art that is proved to have been unknown to the holder. In that case of course the holder did not know that it was invalid. Also for forcing the patents to have the smallest scope, make a patent invalid if it is invalidated on any single count. So that the holder must hold different patents for different points, and not club it together. If all this is done then Patent office need not do anything for approval of the patent, the applicant will make sure it is in good faith.
One only has to look at the rampant achievements and success of Free Software and Open Source to see how much the rest of the industry is being held back by software patents and other "intellectual property" restrictions.
Try not. Do or do not, there is no try.
-- Dr. Spock, stardate 2822-3.
The point is that you should be allowed to benefit from it without your competitors being able to co-opt your idea and benefit from it also. Considering that, unless a market is in a phase of significant growth, a businesses success usually means there are other businesses failing, a new business process that gives you a competitive advantage can make the difference between sink or swim.
Torvalds was reluctant to make predictions though. "I'm the anti-visionary. I distrust people with visions," he said. "You don't see what's right in front of your face and you don't see the technical issues that face everyday users."
While I don't think he'll ever say it directly, this is as clear as he ever needs to be when it comes to his opinion of RMS.
In software there as idea that problem for which is pattent solution should be presented to groups of skilled programers who have week to find solution. If one of concepts they come up with is close to patent concept, its is dismissed, otherwise it should be granted as non trivial and inovative. (Example: someone tries to patent classical fulltext search as "way to retrieve all data based on simple keyword". programers/engineers get task to find way how to "way to retrieve all data based on simple keyword" - one of them comes out with idea that you can preprocess all data ,extract possible keywords and pair then with data, then make search on them and not data. Which is exactly how patent proposed it to do. Other will simply dig on goole and will find working implementation. RESULT: no patent)
+ they will sometimes generate obscure and new ideas woth of further exploration. this could be werry challenging job and good way to get renome.
- human factor is too high - should someone sponzor whole team to make them lazy about ideas, there is no way it could work.
-- Technology for the sake of technology is as pathetic as eschewing technology because it's technology.
a new business process that gives you a competitive advantage can make the difference between sink or swim.
For almost 20 years? Maybe when we're talking car engines here the "mindspace" in inventions is wearing thin enough that someone coming up with a better way of doing it should be rewarded by killing off all the other car makers, but lets take a look at Microsoft, our favorite bogeyman. How many thousands of patents do they have? When was the last time they were driven to innovate in the OS field? Win 3.11 -> Win 95? NT 3.5->4? Everything since then has just been a prettier 95, and 2000 was 95's offspring crossed with NT4's. Meanwhile, they're failing to innovate the features they've promised for Longhorn, and we're watching as Longhorn gets farther and farther away even as it sheds features like a mangy dog. Fat lot of good those thousands of patents are doing them in developing the next generation OS, maybe we should increase the patent length to 30 years so that while Microsoft does jack, they can at least stop other programmers from innovating faster than they can.
(1) Software is speech. It enjoys the same copyright protections as an article you see published in a magazine, or a novel you wrote yourself. How would you like it if someone patented the plot line "detective discovers the killer and confronts him, and defeats him in a fistfight"? And 300,000 other various plot lines? Wouldn't that make it hard to write novels without a risk of getting sued?
(2) Software is math. Mathematical algorithms are *discovered*, not invented. Even data structures, communication protocols etc. are governed by specific limitations of the technology involved. Many times there is one clear way (or a few clear ways) which are the "best" way to implement something (usually fails the "non-obvious" test too). Why should one company be able to patent that and lock out the entire industry from using the best technology for 17 years?
(3) The term of patents is too long for software. Patents are a tradeoff: society gives inventor a limited right to prevent other people from profiting from their invention (for 17 years), in exchange society gets the free release of the idea after that time.
(4) No large research costs to recoup. The reasons that patents make sense for the drug industry and other industries, don't apply to software. The effort is all in writing the code, not in trade-secret-producing research. So there's no need for patent incentives to get those secrets out in the open (as businesses are discovering, having the software be open-source in the first place is often more economical anyway).
(5) Software is just different. Writing software involves building on the work and ideas of others--its practically impossible to write software without building on the work of others. It makes no sense to give someone a monopoly on a *necessary building block* that goes into *all sorts of technology* across the industry (such as a compression algorithm). They will just use it as an anti-competitive weapon, a roadblock for their competitors. Society does not benefit from this (and is usually actively harmed).
And last but not least:
(6) A large part of the purpose of the patent system is to "encourage innovation" and the release of the innovations to society. Well guess what? Innovation in software/technology is already occurring at a breakneck pace, and has been for decades. Society can barely keep up. This is *in spite of*, not because of, the software patents. But software patents are posing an increasing threat and if something isn't done about them SOON, the resulting chill in the U.S. will cause all the best hackers to move overseas. (I myself was planning to move to the U.S. but abandoned those plans about a year ago because the changing technology climate there--and your decreasing civil freedoms--are no longer to my taste).
These are just some of the reasons software patents are a problem. I'm sure there's people here who can list many others.
As I have said before: since the Open Source community is not only bigger and probably as innovative as the software developers of any company, it should fully take advantage of the current system.
Open Source should find investors and patent every single patentable software solutions, developed by Open Source projects.
Open Source would be the patent holder and there is a potential market here to charge big corporations for using Open Source owned patents.
The key here, is that Open Source has a huge, very innovative, highly motivated "workforce" - that can be competitive with any corporation.
Once Open Source manage to secure patents on key new technologies, which cost a lot to corporations, we may actually see the corporations to lobby for anti-patent legislation.
While those corporations finally get the lawmakers to scrap software patent laws, Open Source could have a healthy cash-flow from patent licencing to fund even more projects.
++just an other open.random idea++
I believe that the criterion for non-obviousness is that a team of experts in the subject area working on the specific problem for which the patent provides a solution, needs more than one day to come up with the solution.
Of course, the problem is that the patent office employees are not experts in the subject area, and still need to judge non-obviousness. That's patently impossible.
Although it seems to me that even a patent office employee should recognise that to the problem of having to click multiple times to make a web purchase, the solution of letting the user make a purchase with only one click is quite obvious.
Just copied my post from below:
As mentioned above.. you can own it on your own all you like. Don't share it.
Patents on an invention make sense because in order to realize a profit on it you have to distribute it, which allows for copying.
Patents on processes make no sense, because you do not have to distribute the process in order to realize a profit from it. If the process is so simple that any customer can see it and apply it, then by its very nature it does not pass the "non-obvious" test.
As such, there is no justification to patent processes. Those the consumer can't ascertain you realize profit from by utilization (ala Coke), those that the consumer can ascertain are not worthy of patent protection (ala 1-Click).
As an aside, you do realize that by equating business methods with "expressions of ideas" you are agreeing with me that they should not be patentable?
That aside though, I disagree that business processes are an expression of an idea. They are the idea itself. Copyright does not (and should not) protect the idea. The idea of a monster sucking blood is not protected because Anne Rice has the copyright to "The Vampire Lestat", her specific expression of it is.
If you can show me a business process where the only protection required is on the specific expression (such as the specific code for a 1-click shopping system), and not the idea itself, you may have a point.
That Jesus Christ guy is getting some terrible lag... it took him 3 days to respawn! -NJ CoolBreeze
I note the comment that there are 'an estimated 150,000 to 300,000 registered software patents in the US alone'.
The reason for the range is that no one knows what a software patent is. If you look at research on this subject, you'll find all kinds of different definitions. Some use key words, some look at patent classifications. The large patent numbers may just reflect the fact that many devices built today are digital and have a software component. They might be patented regardless of any patent software restrictions.
The large numbers in the U.S. may also simply reflect the fact that the U.S. doesn't require the patent writer to hide the fact that the patent relates to software. Where software patents are not allowed, the patent writer may just hide the fact that it's got a software aspect by calling it a "method" or "system" and include some nominal hardware.
Also, patents aren't "registered" - the correct word is "granted." "Registration" is for copyrights and trademarks, not patents. The word "registered" implies a process of recording preexisting rights.
An author has inherent copyright rights in any work he authors. A business has trademark rights when it uses a trademark to idnetify goods and distinguish those goods from goods sold by others. They have the option, but not the obligation to register their copyright or trademark (and such registration typically affords them even more rights.)
In contrast, an inventor does not have inherent patent rights that can be simply "registered." Patent rights must be "granted" and they are only granted after an examination.
Copyright law doesn't cover inventions, it covers creative works. Patent law is an application of the same principle to inventions. If copyright law was extended to include inventions, what difference between that and current patent law would there be that you find offensive?
Only physical things should be protected by law, so if inventions can't be physical things then I guess we don't need to extend copyright. I thought more in the ways of "this [product] is not an exact copy of our own [product], but this piece/part right here took us years to develop, so where did they get that technology if they didn't copy it from us?" But that's as far as I think the law should go, it should protect "illegal copying," and that's it. Copyright law differs between having copied something and having made an exact replica without copying (which theoreticly isn't illegal, but that's an extreme case and perhaps would be), but patents are exclusive no matter what.
I can't think of any reason why someone should be allowed to own an "idea", I mean if it's not a "creative work" (something physical) it's not owned by anyone. It's just an idea, it's just a thought. Do something with that thought, prevent others from copying whatever you made (copyright law) and everyone's fine. But others should be allowed to come up with the same thought and make a similar product (or another product entierly) out of that idea if they (can prove they) didn't copy it off you.
Thoughts can be compared to DNA: Someone will figure out what every gene does eventually, just like somone will come up with a particular idea eventually, so it shouldn't be owned exclusively by anyone. (And yes, copyright law will/should prevent copying of information.)
The patent, like other types of "IP" law, are based on the idea that invention is a whole lot of work with no physically scarce product.
This is not a definition, but it's taken from wikipedia, and it's the origin of patents as I see it:
Per the word's original definition, one theory of patent legislation is to induce the inventor to disclose knowledge for the advancement of society in exchange for a limited period of exclusivity.
It says "theory", but I'm OK with that. I think it's correct. Even if it's wrong, that would mean it's created only for the reason that people should make money off their inventions. If that's the case, why are patents publicly available? If they weren't, then people could make an original product that is similar to an existing product without having copied it (*), and if that still was illegal then it's obvious that the patent system is no good. However, since it is in the public, the good old "you could have read it here first"-argument always wins, and the patentholder always wins regardless if someone actually did copy it or not.
The whole point of owning something "just because" makes no sense. You own something because you made it, and it's protected by law to prevent others from copying off your work. But what is the law protecting if they aren't copying?
Simply because you do work, you are not automatically entitled to reward.
Well.. that's regulated by law, so I guess you're right on that one. But you kind of nailed it here:
If I don't look for a patent first, then spend time and money duplicating something that I have no rights to, should I be compensated for my wasted time?
That's basicly the problem as I see it: Everyone should have the right to create whatever they want as long as they're not copying off someone else. Doesn't anything else seem unnatural? Patents deny this, which I think is taking away some freedom from everyone.
It's a complicated topic, but I really can't think of any particulary good reasons to have a patent system. It had good intentions though. Even with a strict and quality assured patent office I still think we'd be better off without it. I hope my rambling is understood by others.
(*) And if this was legal, the patent system might just work. I believe "secret" or "undisclosed" patents exists today, and even then this still is regarded as illegal, something I can't find logical (or just for that matter).
Life is Reality
The plastic packaging for a razor blade is a material good that has non-zero fixed and marginal costs. To make and distribute plastic packaging for a razor blade requires the creation of a factory, and the cost of materials and distribution for each package produced.
While software does have considerable fixed costs, they do not equal the cost of a factory. One person with one computer + time = software. And once the software is made, it literally costs nothing to make as many copies as anyone could ever want.
As a result of this, the number of people who want or need to make plastic packaging for razors is very small.. really only those people who are selling razors. But the number of people who may need or want to create software, whether it be for individual, in-house, or freely shared use, is vast. The number of people potentially locked out by a software patent, then, is much higher than in the razor case.
Why don't we have patents on literary plot devices? Why don't we have patents on speeches, or even on talking about a particular subject? Because the monopoly grant in those cases is extremely expensive, and retards the flow of communications and ideas in the society.
If we were talking about 10,000 software patents, rather than 150,000-300,000 software patents, it might not be so bad. If experienced developers were responsible for judging the obviousness of a software patent application, it might not be so bad.
But it is 150,000-300,000 software patents, and experience developers do not make those calls. Patent clerks harassed by well-paid lawyers do. And as a result, it is all but impossible to write any substantial software in the United States without trespassing on someone's patent claim.
- jon
Ganymede, a GPL'ed metadirectory for UNIX