Weekly Edition
Return to the Announcements pageRecent Features
| |||||||||||||
A weak case for software patents (opensource.com)
Red Hat's VP and assistant general counsel Rob Tiller writes about the case for software patents—or at least the case that is being put forth by proponents. He is responding to a recent blog post by attorney Gene Quinn that, among other weak arguments, calls those who oppose software patents "ideological buffoons". Tiller says: "Not everyone views it this way, of course. Those who are profiting from the existing system generally think that it works rather well. And they have some appealing-sounding arguments. For instance, they argue that patents encourage innovation by allowing lone inventors to pursue their ground-breaking dreams in the face of powerful corporations. This sort of story tends to excite emotions and hinder rational analysis. It ignores the rarity of inventors who work without significant collaboration, of inventions that are ground-breaking, and of patents that ever recover even the cost of the patent application. Dreams of getting a hugely successful patent are about as realistic as dreams of winning the lottery. Still, it's a nice, and understandable, dream."
(Log in to post comments)
A weak case for software patents (opensource.com) Posted Aug 12, 2010 18:11 UTC (Thu) by dbruce (subscriber, #57948) [Link]
A quote from Gene Quinn:
"Patent law provides no such independent creation defense, so the sole question is whether you are doing what someone else has patented. It simply does not matter whether you came up with it on your own."
Although Mr. Quinn generally makes my blood boil, I'm inclined to thank the guy for posting the above - it is about the most elegant and succinct statement I've seen that demonstrates why software patents need to be abolished. (not the point he was aiming to make, obviously)
DSB
A weak case for software patents (opensource.com) Posted Aug 12, 2010 18:40 UTC (Thu) by gmaxwell (subscriber, #30048) [Link]
I don't agree that this statement is good evidence against software patents. Patents are intended to be a monopoly on an idea if they weren't they wouldn't be patents.
If patents were only granted on genuine innovation with terms, and thus rewards, consistent with the effort and risk required to create the innovation and if the system had equitable access to all innovators, then the patent system could achieve its intended purpose of promoting innovation. If the increase in innovation offsets the other costs of the system, then it's a good deal for the public.
Software patents are not only awarded only for genuine innovationin fact, the results of litigation suggest a majority of all current patents contain at least some impermissible claims and the rules for permissibility go far beyond genuine innovation, the terms granted are far in excess of the costs and risks of software development, and access to the patent system is inequitably distributed: Patent fees are mostly uniform, yes, but the _value_ of owning a minefield is exponentially greater than owning a single mine.
These factors make patents expensive for society and they make it discourage innovation rather than promote it. So, these are all reasons that the system is currently inappropriate for software. The exclusive access to an idea aspect of patents doesn't itself make too much of a counter argument.
OK, let me elaborate a bit more Posted Aug 12, 2010 19:34 UTC (Thu) by dbruce (subscriber, #57948) [Link]
Patents, as originally intended, were supposed to protect inventors from having their ideas stolen from them. Also, the patent wasn't just on an idea, it was on an actual, working invention.
With software, there isn't any "invention" apart from the idea - it is all just a recipe. The problem is that the lay public thinks of programs as "things", while those of us who code conceive of software as a specialized form of writing. At least that's how I think of it, and it seems hard to dispute.
Once one understands that software is *writing*, the idea of "copying without knowing" becomes quite absurd. I can certainly understand the rationale for traditional copyright protection of software. I can also see how an argument could be made that software which knowingly imitates another program might be subject to restriction (but fortunately is generally not, absent patents). However, I cannot follow the logic in prohibiting independently created *writing* that coincidentally resembles something someone else has written, when the two authors are completely unaware of each other's work. That's not "rewarding innovation", that is granting of toll booths.
Sure, "patents are intended to be a monopoly on an idea" - which makes it absurd to grant them in a field of writing. We don't allow patents in literature or music. If the greater public understood programming better, we wouldn't allow them in software, either.
DSB
OK, let me elaborate a bit more Posted Aug 12, 2010 19:47 UTC (Thu) by daglwn (subscriber, #65432) [Link] those of us who code conceive of software as a specialized form of writing. That's a rather odd take. Is hardware design also writing? Most of it is done in high-level languages these days. Yet it produces a physical object. There are compilers to translate C code into a circuit implementation. Does that make that particular piece of software suddenly not "writing" anymore? The innovation in software is not in the typing but in the development of algorithms. One could also argue that a book is not the words but the construction of the plot and the development of the characters. In that sense I could make a case for the patentability of writing. However, a copyright seems a much more effective tool in that field than a parent. But I am by no means an expert in that field. But regardless, software is much more like designing hardware than writing a book.
OK, let me elaborate a bit more Posted Aug 12, 2010 22:03 UTC (Thu) by flewellyn (subscriber, #5047) [Link]
3D printing being what it is, the distinction between creative expression and manufacturing is becoming more and more blurred.
Given that fact, the broad monopoly of a patent is just going to become more of an obstruction in the future.
OK, let me elaborate a bit more Posted Aug 12, 2010 22:40 UTC (Thu) by nybble41 (subscriber, #55106) [Link]
"Is hardware design also writing? Most of it is done in high-level languages these days. Yet it produces a physical object."
Yes, hardware design in high-level languages is also a form of writing. The design work alone itself does not produce any physical object, and there is nothing novel about employing well-known synthesis and fabrication processes to produce a physical object from a high-level abstract description--that is what the synthesis tools and fabrication plants were designed to do, after all.
Also, software development is exactly like writing a book: an instruction manual, to be precise. There is no requirement that computers be involved at all; any software algorithm could be evaluated by unaided humans at need. Computers just speed up the process.
OK, let me elaborate a bit more Posted Aug 13, 2010 4:05 UTC (Fri) by error27 (subscriber, #8346) [Link] To me it's simple. One is *hardware* and one is *software*. If you have something can be downloaded off the internet that is software. If you can weigh it, that's hardware.Of course, all hardware can be described it writing, not just computer chips. Imagine the Wright brothers had written down the design of their airplane. That doesn't mean they can copyright the invention and have sole ownership until 70 years after the author's death. Hardware has to be patented instead and it gives you more rights but it's capped at 15 years. When I write software, I don't think of it as designing hardware at all. Maybe someone could use your C to circuit translation device on my software. I don't care, so long as they recognize that each time they do that they're making a copy and copyright law applies. If I did care, I'd maybe get a patent on the actual hardware.
OK, let me elaborate a bit more Posted Aug 13, 2010 4:24 UTC (Fri) by daglwn (subscriber, #65432) [Link]
Most hardware patents never have a physical implementation. Many patents in many fields never have a physical implementation.
Your Wright example doesn't make sense. Copyright and patents are different. If the Wrights copied down their design (as they certainly did) and copyrighted the drawings, that doesn't prevent someone else from designing and selling exactly the same thing. They can even use the Wrights' drawings, as long as they don't violate the copyright (get 'em from a library or something).
OK, let me elaborate a bit more Posted Aug 13, 2010 10:35 UTC (Fri) by error27 (subscriber, #8346) [Link]
But the point is that even if you don't build the machine, you still patent the machine, it's still a machine.
My point was that if software is hardware because you can have a circuit that implements the software, does that mean that all hardware is software? Because all hardware can be described.
Copyrights and patents are different things just as hardware and software are different things, despite any c to circuit translations.
OK, let me elaborate a bit more Posted Aug 13, 2010 15:24 UTC (Fri) by daglwn (subscriber, #65432) [Link]
Software and digital hardware are really not that different. We have algorithms in hardware just as we do in software. We have well-known idioms in hardware just as we do in software. The creative process is essentially the same. Only the physical product differs.
OK, let me elaborate a bit more Posted Aug 13, 2010 9:03 UTC (Fri) by shmget (subscriber, #58347) [Link]
"The innovation in software is not in the typing but in the development of algorithms."
Yes indeed.
"We defined "algorithm" as a "procedure for solving a given type of mathematical problem," and we concluded that such an algorithm, or mathematical formula, is like a law of nature, which cannot be the subject of a patent."
It might be usefull to quote Knuth here:
"I am told that the courts are trying to make a distinction between
IOW:
BTW, on a personal note, I do share the view that "those of us who code conceive of software as a specialized form of writing.".
OK, let me elaborate a bit more Posted Aug 13, 2010 15:32 UTC (Fri) by daglwn (subscriber, #65432) [Link]
Ok, I'll grant that there is some inconsistency by the courts here. If we are to be consistent, however, we need to apply the same benchmark to all fields. Digital hardware will become unpatentable because it's really just algorithms implemented in a different way.
Now let's look at something like Watt's parallel motion. In the end it's really just an expression of a physics truth. It's a combination of pieces that reveal a natural phenomenon of physics. One could think of it as an algorithm. Or even a proof.
What is a mechanical drawing except a recipe (algorithm) to exhibit some physical truth?
OK, let me elaborate a bit more Posted Aug 13, 2010 0:41 UTC (Fri) by jmspeex (subscriber, #51639) [Link]
"Patents, as originally intended, were supposed to protect inventors from having their ideas stolen from them."
That's a common misconception. Patents were *not* created to protect inventors. They were created to give inventors an incentive to *publish* their inventions (in return for a temporary monopoly). So the goal of the patent is to make it "patently" clear how the invention works.
OK, let me elaborate a bit more Posted Aug 13, 2010 1:18 UTC (Fri) by AndreE (subscriber, #60148) [Link]
Exactly.
Invested individuals like to misrepresent the fact that any form of IP rights legislation in supposed to encourage developement by the entire industry, not just rights holder
OK, let me elaborate a bit more Posted Aug 13, 2010 3:32 UTC (Fri) by butlerm (subscriber, #13312) [Link]
"They were created to give inventors an incentive to *publish* their inventions (in return for a temporary monopoly)"
The number of patents that are actually worth reading to learn anything about anything is vanishingly small. Whatever the original thought was, these days that argument is a joke.
OK, let me elaborate a bit more Posted Aug 13, 2010 4:25 UTC (Fri) by daglwn (subscriber, #65432) [Link]
Yep, that's another failing of the patent office and another good reason to dump patents altogether.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 0:11 UTC (Fri) by Wol (guest, #4433) [Link]
Patents are NOT a monopoly on an idea (at least, they are not meant to be!). Patents are a monopoly on a *practical* *means* of achieving an end.
For example, the Severn Bridge carries a six-lane motorway over a single span over a thousand yards long. It is THEORETICALLY possible (ie the maths says you can do it) to suspend this from a steel cable maybe a cm across. In actual fact, we can't make steel that strong, and patents protect the steels we can make.
The problem with software is it is NOT a *practical* *means* of achieving an end. It is a THEORETICAL means of achieving our end (like a 1cm cable is a THEORETICAL means of hanging the Severn Bridge). We have to combine our theoretical means with a computer in order to achieve a practical means, and that combination is not patentable. The computer isn't patentable on its own as a means, and the theory isn't patentable, so the result of the combining the two isn't patentable. That doesn't mean patents aren't granted on an unpatentable combination, though :-(
Cheers,
A weak case for software patents (opensource.com) Posted Aug 13, 2010 4:33 UTC (Fri) by daglwn (subscriber, #65432) [Link] Patents are a monopoly on a *practical* *means* of achieving an end. I cannot find that language anywhere in Title 35, specifically part II, chapter 10, sections 100-105, as given here.
A weak case for software patents (opensource.com) Posted Aug 14, 2010 11:10 UTC (Sat) by Wol (guest, #4433) [Link]
Why should I be interested (or know about) "Title 35"? It doesn't apply to me.
But look at the history of patents. Patents were created to get inventors to publish HOW they did something. Ie if you published the *means* by which you achieved your end, you got a patent on *using* that particular means. A very *practical* issue.
Cheers,
A weak case for software patents (opensource.com) Posted Aug 12, 2010 18:36 UTC (Thu) by daglwn (subscriber, #65432) [Link]
I simply don't understand what's so special about software. If software patents are bad, then we ought to just throw out the whole patent system, something I might actually support.
But I don't support making software unpatentable while allowing every other discipline to use the patent system. That's not a level playing field.
The truth is there are some truly innovative software patents. The problem isn't software patents per se. The problem is bad patents, in any discipline.
A weak case for software patents (opensource.com) Posted Aug 12, 2010 18:47 UTC (Thu) by coriordan (guest, #7544) [Link] Here're some arguments about this: Basically, loads of disciplines are unpatentable -fashion, literature, and music being examples- so excluding software isn't strange. Unlike mechanical innovations or pharmaceuticals, software has a per-copy cost ("marginal cost") of zero. Mass production doesn't require anything more than a computer and an Internet connection, as already found in most houses. If a car manufacturer has to add on $15 per car for a patent royalty, that's an economic issue. In software (or in literature or music), even if a royalty is $1 per copy, that imposes big problems because the developer has to keep track of the copies, which is very difficult. The need for interoperability is also particularly acute in software. If you make an innovative word processor and it doesn't visually resemble existing word processors or it can't read and write the file formats commonly used, it's simply not a functional word processor. You can patent your innovative features, and the established companies will just skip those features, and they can patent their features and you're blocked from implementing something that users consider essential.
A weak case for software patents (opensource.com) Posted Aug 12, 2010 19:31 UTC (Thu) by daglwn (subscriber, #65432) [Link]
Let's take the points from your first link.
- Software patents are litigated more
So what? That should in no way determine whether something is patentable. Much of this litigation is over invalid patents. The problem is granting patents for things that shouldn't be patentable.
- Software is math
No, it's not. There's no fundamental "law" or "discovery" or nature that I can make that will produce a highly optimizing compiler. That takes creativity and engineering as there is an infinite number of ways to produce a compiler or any software.
- Abstraction
Yes, this is why software is not math. There are many cases in other fields where there is not an absolute measure of infraction. That's why we have a court system. But an engineer can look at two implementations and decide if they are substantially the same. Hardware engineers deal with exactly the same issues.
- Problems and solutions are tied
There's nothing software-specific here. The MPEG format is patentable because it's a particular method for compressing video streams. Anyone is free to invent another method. The real problem with MPEG and similar things (RAMBUS, etc.) is that they became official standards. That's a failing of the standards bodies, not the patent office.
- The test of obviousness
The obviousness test is no more difficult for software than for anything else. Good programmers will not always come to the same solution independently. See the compiler example above.
- Prior art
Software is not invisible. There are countless journals and other publications covering it. Patents require disclosure. And it's a bit of hubris to assume that programmers are somehow better at coming up with ideas than anyone else.
- Low Risk
So what? There are varying levels of barriers to entry across fields.
- How patents on different domains affect society
I don't follow this at all. Seems like a bunch of made-up mumbo-jumbo to me. Nothing about "individual liberties" or "communities" has any relevance here. The Free Software community plays in the same arena as everyone else. It are no more affected by patents than anyone else.
Now the second link:
- Software is unlike applied science
Software is in no way predictable. If that were true there would never bee a need for upgrades. It takes real effort, planning, engineering and work to create software. And in any case, "amount of effort" has no bearing on patentability. Obviousness does.
A weak case for software patents (opensource.com) Posted Aug 12, 2010 20:00 UTC (Thu) by dbruce (subscriber, #57948) [Link]
Software most definitely is math. Formally, it is impossible for software to be anything but math. Lay people just have a very restricted understanding of what the term "math" means, and thus the operations of software don't "seem" to have much to do with the math they learned in school.
Now, it would be perfectly fair to argue that even though software is math (in the mathematics/CS sense), that should not make it unpatentable. But to claim "no, it's not" is just silly.
Certainly "creativity" is no argument that something is not math. The entire discipline of mathematics itself is all pure creative thought. Does that mean that none of it is math?
See the following excellent essay ("An Explanation of Computation Theory for Lawyers) for more info. If you are open-minded about it, I doubt you will reach the end of the essay still denying that software is math.
http://www.groklaw.net/article.php?story=20091111151305785
DSB
A weak case for software patents (opensource.com) Posted Aug 12, 2010 20:25 UTC (Thu) by daglwn (subscriber, #65432) [Link]
Software is not math in the sense the page linked to used the term. Software systems are not "discovered," they are designed.
A weak case for software patents (opensource.com) Posted Aug 12, 2010 23:47 UTC (Thu) by Wol (guest, #4433) [Link]
So are mathematical proofs.
And just as software is debugged, so are mathematical proofs. In fact, the two different techniques are so close, they're actually identical. It's just that mathematical proofs are executed by computers (in the old-fashioned sense, that is, ie people following a list of instructions).
Riddle me this. Software is written as text. Text is represented as numbers. Numbers are represented as 0s and 1s. But what are 0s and 1s represented as?
If you can't give me a definitive answer (as opposed to "in some systems it's this, in some systems it's that, in other systems its something else), you've just identified a PURE abstraction layer. In other words, anything on the non-physical side of this layer is abstract, therefor it's logic, therefor it's maths.
Cheers,
A weak case for software patents (opensource.com) Posted Aug 13, 2010 4:41 UTC (Fri) by daglwn (subscriber, #65432) [Link]
It doesn't matter is something is abstract or not. The U.S. constitution says nothing about abstractness. It's up to Congress to decide what is patentable, which they did in Title 35. If it wants to say we can't patent proofs but we can patent software, it has that right.
Yes, in a theoretical sense, software is math. But it is not math in the way the original link talked about it. The result of a program is not a fundamental truth. Even logic programs don't produce fundamental truths. They produce proofs of fundamental truths. The logic program is a means to produce a proof. It is not a proof itself. It's a tool. It's patentable and should be as long as any other tool is patentable.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 8:02 UTC (Fri) by Felix.Braun (subscriber, #3032) [Link]
Well, I'm no expert in American constitutional law. But over here in Germany, parliament is obliged by the constitution to treat like things alike. I very much suspect there to be an equivalent principle in the US Constitution.
If you accept that there is no theoretical sane way to differentiate between software and maths, then there might very well be a limit to the power of Congress to treat maths and software differently.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 15:34 UTC (Fri) by daglwn (subscriber, #65432) [Link]
Nope. The U.S. Constitution is very vague on a number of important points. This is by design. In fact there used to be language explicitly treating alike things differently. Congress decided that wasn't a good idea and drafted a few amendments which the states passed. We had a whole bloody war about it.
A weak case for software patents (opensource.com) Posted Aug 14, 2010 11:21 UTC (Sat) by Wol (guest, #4433) [Link]
I think you're missing the point ... I guess you're talking about slavery.
Thing is, you can say "A is a slave, B is not", and it's a matter of classification. And seeing as "slave" is a state of being it's quite possible for some people to be in that state and others out.
But if congress passed a law saying "software is patentable, maths is not", then what happens when a court case comes along that says "software IS maths". To put it another way, we have the set of people, that contains the subsets "slaves" and "non-slaves". Social history says that the union of those two is the set of people, and the intersect is empty, and which one any individual person in is arbitrary. So the courts can deal with that.
But when a case comes along that says MATHS SAYS THAT the set of software is a subset of the set of maths, then what are the courts to do? As maths, the law is clear that software CANNOT be patented. But the law says software can be patented. What to do? This is made even worse by the fact that the set "maths but not software" is the empty set! So you can't even argue "we'll allow patents on software" because that promptly makes all of maths patentable!
Cheers,
A weak case for software patents (opensource.com) Posted Aug 13, 2010 9:00 UTC (Fri) by Randakar (guest, #27808) [Link]
Your idea is correct but your argument is wrong.
I agree that nothing should be patentable. Patents are a failed idea. Humanity implemented it because it looked like a good idea but it never in actually was a good idea. It didn't even work for steam engines back in the day. Nor does it really work for the pharmaceutical industries.
(Pop quiz: Why do so many medical 'solutions' boil down to "take this pill for 30 years"?
Yes, I hold the patent system responsible for it. If pharmaceuticals could be easily be made by the lowest cost producer the money flow to this field of research would become available for other purposes, like finding actual solutions ..)
Anyway: Your argument that software isn't math is wrong. It _is_ math. The real problem is: So is everything else.
Any attempt to distinguish abstract concepts from the physical world in any field of technology is ultimately bound to fail. It is not a problem that has a solution. Ideas are in many ways bound to the physical world and vice versa, the physical world has been shaped by abstract concepts and ideas. The further we go in improving our technology the more true this becomes.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 15:37 UTC (Fri) by daglwn (subscriber, #65432) [Link] It _is_ math. The real problem is: So is everything else. Yes! That is one of the reasons software is not special. Either patent it all or patent nothing. That's what I'm trying to argue.
A weak case for software patents (opensource.com) Posted Aug 14, 2010 11:26 UTC (Sat) by Wol (guest, #4433) [Link]
Actually, everything is NOT maths.
Maths does not exist. It does not have physical reality.
As I said elsewhere, it's all very well saying "my hammer is maths", but even with a perfect mathematical hammer you're not going to be able to drive any nails :-)
Tip: Go and do some philosophy. Yes, I used to think philosophy was a stupid waste of time. Then I had to do some for my Masters and suddenly realised that questions like "What IS science? what IS maths? what IS religion?" suddenly made real sense.
Cheers,
A weak case for software patents (opensource.com) Posted Aug 16, 2010 9:20 UTC (Mon) by Randakar (guest, #27808) [Link]
Two points:
1)
2)
It's possible to use the rules of physics as we know them (and all the associated math) to describe your hammer.
Turning the argument around even further: There is no hammer. There is also no software. There is only the representations of both of these things in your mind. Which, incidentally is an entirely physical thing, made of braincells and neurons. A thing that thinks. Thinking which is still entirely goverened by the rules of physics (and thus, math).
3)
But where's the line? Is there even one?
How do we distinguish ideas about the software that made these objects possible from ideas about the hardware itself, and how do we distinguish that from ideas that are entirely too obvious?
----
So where is the dividing line between ideas about physical objects, abstract ideas, and obvious ideas?
I don't think there is one. There is a vast multicoloured spectrum of ideas that all have their own place. Some ideas may be "definitely" abstract (software, business models etc all fit that) and some ideas may be "definitely" about physical things. But there is a vast range of things that are neither, and both, and probably at the same time.
The courts will struggle with this until the cows go home. It doesn't matter how many tests they come up with. They won't be able to figure it out because it isn't possible to figure out. Even if something fits neatly in one of the boxes for one person, a simple change of perspective to a different person will likely undo that entirely.
The only sane conclusion is that everything is abstract to a degree.
All that having been said - I think the case against patents is much simpler than this.
I'd really like to hear of even one case where patents actually achieved the stated objective of promoting disclosure and thereby spurring innovation. I would be very surprised if there is one.
Time for humanity to say "tried that, it failed, let's try something else". Unfortunately, now we have vested interests fighting to keep the neat complicated structure we've build.
A weak case for software patents (opensource.com) Posted Aug 12, 2010 20:08 UTC (Thu) by jrn (subscriber, #64214) [Link]
Just an aside:
> - Software is math
Id recommend finding a practicing mathematician and asking what she does. It turns out that there is no fundamental law that I can make that formulates theorems or produces proofs without creativity and engineering work, either.
Please do not take this as an argument for the patentability of mathematical theorems.
A weak case for software patents (opensource.com) Posted Aug 12, 2010 20:27 UTC (Thu) by daglwn (subscriber, #65432) [Link]
I was arguing against the particular case made in the linked page, not making a general declaration about what is or is not math. Sorry for the confusion. Math is certainly a creative exercise. But the product of designing software is not a universal truth in the way a mathematical theorem is.
A weak case for software patents (opensource.com) Posted Aug 12, 2010 23:58 UTC (Thu) by Wol (guest, #4433) [Link]
Please explain the difference, then ... :-)
Cheers,
A weak case for software patents (opensource.com) Posted Aug 13, 2010 3:31 UTC (Fri) by daglwn (subscriber, #65432) [Link]
I'm not sure what's so difficult to understand. A mathematical proof describes some fundamental truth. The result of the proof is unchangeable and is pure discovery. Developing the proof is certainly hard work but the result of it is something that hasn't been designed.
I suppose one could argue that a particular mathematical proof should be patentable (as opposed to the result of the proof, which is pure truth) but that doesn't seem to provide any benefit. Once the truth is known there's no need for anyone else to rediscover it.
A software system has to be designed. Different programs that perform the same function often do them very differently. Those differences are based on algorithms that were designed by engineers.
Just like an internal combustion engine can be designed in different ways, so can software. It makes no difference whether it is a physical thing or not.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 7:55 UTC (Fri) by dgm (subscriber, #49227) [Link]
> A mathematical proof describes some fundamental truth.
Not quite. A mathematical _theorem_ is suposed to be a truth, when interpreted in the right logical system and given the correct set of axioms. A proof is a list of intructions to verify that the theorem is a truth in that system.
In fact, a proof is very similar to an algorithm, wich is by the way, just another mathematical construct (see http://en.wikipedia.org/wiki/Algorithm).
> Different programs that perform the same function often do them very differently.
That's exactly the same in mathematics. There are many (sometimes infinite) ways to prove something. Some proofs are easier to understand, or shorter, or more elegant (by some aesthetic measure) than others, but almost never unique.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 9:45 UTC (Fri) by shmget (subscriber, #58347) [Link]
"I'm not sure what's so difficult to understand."
Apparently Donald Knuth didn't understand either:
So I guess, I'm in good company with my 'difficulty to understand'.
A weak case for software patents (opensource.com) Posted Aug 14, 2010 11:40 UTC (Sat) by Wol (guest, #4433) [Link]
Following up from the other person who said that the proof IS the fundamental truth, not a description of it ...
But the proof is DETERMINISTIC. Given the same input, you will get the same output, EVERY TIME. That's why a proof IS a truth.
And the same is true of software. Given the same input, to the same set of instructions, you will end up with the same output, EVERY TIME.
And there we also have the difference between hardware and software, between reality and maths. Once we RUN the software we cannot guarantee the same input. We cannot guarantee the CORRECT EXECUTION of the instructions. Let's say (my academic field) that I have a "program" written in chemical formulae. It's easy to prove that if I feed chemicals X into the equation, I get chemicals Y out the other end. Maths, software, equations, call it what you like, they are all FUNDAMENTAL TRUTHS.
I now build a computer (a chemical plant) to run my program. The input won't necessarily be quite what I expect, the hardware might glitch (there's an air leak), quantum reality might decide to do something strange. Either way, patents are perfectly valid on the COMPUTER - patents are meant to cover my tricks for guaranteeing the correct running of the program.
But I can't patent the chemical equations. I can't patent the fundamental truths. I can't patent the formula - the list of instructions - the software.
Cheers,
A weak case for software patents (opensource.com) Posted Aug 13, 2010 1:30 UTC (Fri) by AndreE (subscriber, #60148) [Link]
Wait, so code isn't deterministic? If I read a piece of code many times I can infer different meanings?
A weak case for software patents (opensource.com) Posted Aug 13, 2010 4:36 UTC (Fri) by daglwn (subscriber, #65432) [Link]
The link didn't say "deterministic," it said, "predictable" in the sense that given a particular goal, the software to get there is so obvious that everyone will write it the same way. That is false as we all know from much hard-won experience.
A weak case for software patents (opensource.com) Posted Aug 15, 2010 11:20 UTC (Sun) by andrejp (guest, #47396) [Link]
There's only so many ways to peel an onion. You can start from the left or you can start from the right. You can even split hairs whether each of the angles that you start at is a different approach (and you do know that theoretically a circle of 360 degrees can be subdivided into an infinite number of subdivisions), or you can split hairs whether changing directions constitutes a different approach also. But the fact remains that there's only so many ways to peel an onion. Same goes for software onions. And discovering the most efficient way to peel an onion is, well, a discovery. Sure it takes effort to measure and try different approaches, but these are not inventions, they are discoveries. And there can be more than one most efficient way to peel an onion, but each is still a discovery. So argue that an algorithm for peeling an onion is patentable. Or argue that ways of going about it are not obvious, or independently rediscoverable, or deterministic (predictable if you like). Or that granting special monopoly rights on peeling an onion is beneficial for inovation. *PLEASE*
A weak case for software patents (opensource.com) Posted Aug 13, 2010 2:54 UTC (Fri) by bronson (subscriber, #4806) [Link]
Better notify Turing then! His most interesting results came out of treating math and code as if they're the same thing. If that's not true, the CS field is due for a big shakeup. :)
A weak case for software patents (opensource.com) Posted Aug 13, 2010 1:51 UTC (Fri) by cmccabe (guest, #60281) [Link]
I guess it's kind of refreshing to see someone taking the other side of the argument. But I'm afraid you are confused.
> > The obviousness test is no more difficult for software than for anything
You have already admitted that there have been many invalid software patents granted. Apparently the USPTO is not quite as smart as you, because they seem to be having quite a bit of difficulty.
> > The real problem with MPEG and similar
So if standards organizations stick with decades-old standards just because they're not patented, this makes everything ok?
Or how about an alternate interpretation: 15 year, government-granted monopolies are fundamentally incompatible with rapid progress.
It's all right, though. I'm sure that if the U.S. sabotages itself with excessive government regulation and intervention, countries like China and India will hold up their end of the "gentleman's agreement" not to out-compete our domestic industries. Just like they did with manufacturing and services.
Here's what Bill Gates had to say in 1991. (Source: NY Times)
> In a memo to his senior executives, Bill Gates wrote, If people had
It's all right, though-- I'm sure that you understand the field much better than someone like Mr. Gates.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 4:45 UTC (Fri) by daglwn (subscriber, #65432) [Link]
I agree with most of what you've said. It applies to every field. It's an argument for why we shouldn't grant patents. There's nothing software-specific in any of these objections.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 8:11 UTC (Fri) by cmccabe (guest, #60281) [Link]
> I agree with most of what you've said. It applies to every field. It's an
> argument for why we shouldn't grant patents. There's nothing > software-specific in any of these objections.
Did you even read what I wrote? Which part of it "applies to every field"?
The part about the obviousness test being extremely difficult, perhaps impossible, for software? The quote from Bill Gates about the potential impact of patents on the software field? The difficulty posed by patents to software standards organizations?
A weak case for software patents (opensource.com) Posted Aug 13, 2010 15:38 UTC (Fri) by daglwn (subscriber, #65432) [Link]
Obviousness is hard in every field.
Patents restrict progress in every field.
Basically, replace "software" with "every field" in your post and the truth remains the same. That's what I meant.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 11:27 UTC (Fri) by jjs (guest, #10315) [Link]
>> - Software is math
>> No, it's not. There's no fundamental "law" or "discovery" or nature that I can make that will produce a highly optimizing compiler. That takes creativity and engineering as there is an infinite number of ways to produce a compiler or any software.
1. All LISP is an implementation of Lamba Calculus - a branch of mathematics.
2. Therefore all LISP programs are math
3. All software can be rewritten in LISP (a Turing Complete language)
4. Therefore all software is math.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 4:28 UTC (Fri) by error27 (subscriber, #8346) [Link]
It's offtopic, but the cost of producing pharmaceuticals is basically zero as well. If you wonder why they cost so much then ask yourself how Microsoft decides the price of their various software and extrapolate.
Unit cost of pharmaceuticals Posted Aug 13, 2010 21:16 UTC (Fri) by jrn (subscriber, #64214) [Link]
I am not so sure. Generics are a lot cheaper than the pharmaceuticals with monopoly pricing, but they still cost a lot more than zero.
In fact, thats probably a good way to measure the effect of patents on price: look at what happens to prices when the patent expires (in a slower-moving field than software).
A weak case for software patents (opensource.com) Posted Aug 12, 2010 19:20 UTC (Thu) by drag (subscriber, #31333) [Link]
> I simply don't understand what's so special about software. If software patents are bad, then we ought to just throw out the whole patent system, something I might actually support.
Since there is no natural 'ownership' of ideas then patents are a purely utilitarian concept.
That is you still own your ideas, but sharing your ideas and having your ideas copied in no way really affects your ability to use your ideas. That is if you were discover that 2+2=4 and I 'stole' that idea then that does not remove your ability to do math. So there cannot be any real theft, because there is nothing about copying ideas that denies other people's property.
Laws based on 'natural' concepts like this only can really be justified when they serve to resolve or prevent human conflict and violence. That is theft is bad because if you take my food then you remove my ability to use my food... this is effectively a violence against me.
So patents can only be justified if you take the 'utilitarian' approach of that usually described as 'the greatest good for the greatest number'. That is that a law provides a net benefit to society that overrules the loss to society caused by the restriction of liberty and the use of property.
Utilitarian based laws tends to be rather arbitrary and patents reflect this. Most science is not patentable, for example. Only very specific branches of science and research will ever produce something that is legally patentable. Math is not patentable, but engineering is. Fonts are patentable, but icons are not. So on and so forth.
So the only truly logical argument for or against patents is based on the economic results of the existing and applicability of the patents.
Does software patents provide a net gain for society or a net loss?
If your a utilitarian thinker and your unable to prove that software patents increase innovation and economic value of a society then they should be eliminated.
I think it's pretty obvious to any of us that software patents provide a net loss. The amount of money made from patent licensing and sales based on purely patented ideas is dwarfed by the amount of money that gets paid to lawyers and other overhead. The vast majority of patents never, ever, get licensed or used even when people are violating them. There is simply very little benefit to even exercising your patent rights as a business, most of the time. (of course if you do not produce anything then patents are more profitable and that is why we have patent trolls)
----------------------------------
Personally I tend to object to patents on moral/ethical grounds, but unfortunately I cannot force other people to agree to my viewpoint so taking a practical/utilitarian argument is a often more effective one.
Probably one of the big reasons you have resistance against the anti-software patent movement is because if your able to eliminate software patents then the rest of the system may soon follow. Once you start examining the whole concept of 'Intellectual Property' from a critical perspective based on facts and practical reasoning (as well as ethics) then the whole system starts to unravel and it becomes increasingly difficult to justify any of it.
A weak case for software patents (opensource.com) Posted Aug 12, 2010 19:40 UTC (Thu) by daglwn (subscriber, #65432) [Link] The amount of money made from patent licensing and sales based on purely patented ideas is dwarfed by the amount of money that gets paid to lawyers and other overhead. Well, that depends. A certain supercomputer company made a heck of a lot of money from its patenting of a particular compiler optimization. It's up to each entity to decide whether something they have is worth patenting. Measuring "gain" for society is tricky. Did society gain when Bell patented the telephone? I'll bet Elisha Gray would not think so. It eventually led to a huge nationwide monopoly which was only broken up in 1982. So should the telephone have not been patented? I think the real question is whether patents in general do more good than bad for society. That's certainly a worthy debate IMHO because the answer is not clear to me.
A weak case for software patents (opensource.com) Posted Aug 12, 2010 19:48 UTC (Thu) by drag (subscriber, #31333) [Link]
> suring "gain" for society is tricky.
Yes. This is extremely difficult. It's very high-torque macro-economic science.
But for our benefit there has been a tremendous amount of research and papers released on the value of patents... and they all turn out neutral or against patents for economic and innovation producing reasons, and there is almost non-existent evidence supporting the case of patents.
Google around. Papers against patents are all over the place. Papers for patents are non-existent.
A weak case for software patents (opensource.com) Posted Aug 12, 2010 20:12 UTC (Thu) by coriordan (guest, #7544) [Link] Here are papers, mostly specific to software patents, plus some general info about the idea (which I'm not attached to) of abolishing all patents:
A weak case for software patents (opensource.com) Posted Aug 12, 2010 20:30 UTC (Thu) by daglwn (subscriber, #65432) [Link]
If you are for abolishing software patents because they harm society more than they do good and there are papers showing that this is true for the patent system in general, why are you not "attached" to abolishing all patents?
I'm just curious, not trying to start a flamewar.
A weak case for software patents (opensource.com) Posted Aug 12, 2010 21:07 UTC (Thu) by coriordan (guest, #7544) [Link]
I've never looked into how patents affect other fields. My previous career was in software development, and I've now spent seven years working on anti-swpat campaigns, and I'm convinced that patents on software are a problem for society.
Patents on other areas won't affect society the same*, so I'd have to spend time looking into that before being for or against abolishing patents in those other fields.
(* The main difference is that patents on cars are will only be used against entities who mass produce cars, which means they're an industrial regulation that doesn't limit people's freedoms and can be examined within the narrow scope of economics - will enough be produced of sufficient quality at a reasonable price. On the other side, there's software, which is mass produced by big companies and also by individuals and by communities, so when you put patent regulations in the software field, they limit people's freedoms, and the harms this is causing are evident. Meanwhile, it's not evident to me whether or not cars would be better or cheaper if patents there were abolished.)
A weak case for software patents (opensource.com) Posted Aug 12, 2010 23:57 UTC (Thu) by Wol (guest, #4433) [Link]
But non-software patents DID affect society the same. Software is a new field, so patents are stifling software development. When steam was a new field, patents stifled the field horribly. When aeroplanes were a new field, patents pretty much killed off the fledgling US aviation industry. Etc etc.
And don't talk to me about Edison !!! Lemelson is seen as a modern day patent bandit. If you actually investigate Edison, I don't know how many of his patents were similar banditry, but the lightbulb certainly was!
Edison visited England shortly before he filed his patent. While in England, he saw lightbulbs (and the patent on them!) invented by Joseph Swann. I believe when you filed a patent in America back then you were supposed to file a working copy of the invention? Well, Edison's patented light bulb had a working life measured in minutes. The light bulb Edison actually manufactured and sold was pretty much an exact copy of the bulb described in Swann's patents. The ONLY real value in Edison's patent was as a weapon to force Swann out of the American market.
Cheers,
A weak case for software patents (opensource.com) Posted Aug 13, 2010 0:48 UTC (Fri) by drag (subscriber, #31333) [Link]
I agree with Coriordan.
In this case we should stick with software patents specifically. The documentation on patents in general is very helpful and good to know and can be applied in the defense of anti-swpat, but going after patents in total is politically very difficult.
Let the experts in macro-economics (Mises is good for this) and experts in their respective fields speak out in their behalf on how patents are hurting them. Providing and gathering expertise in software patents specifically helps our cause the most while still lending credence to reform in general.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 4:49 UTC (Fri) by daglwn (subscriber, #65432) [Link]
I understand your strategy, but there's a danger of software patents being declared invalid and everyone thinking the problem is solved, simply because we've focused on software patents exclusively rather than exposing the fundamental problems of the whole system.
The patenting of pharmaceuticals and genomes (!) is far more harmful to society than the patenting of software. The former leads to sickness and death.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 8:23 UTC (Fri) by cmccabe (guest, #60281) [Link]
> The patenting of pharmaceuticals and genomes (!) is far more harmful to
> society than the patenting of software. The former leads to sickness and > death.
Again, you are confused. The entire business model of big pharmaceutical companies is based on doing years of research and testing on many drugs, and finally getting a few patents on ones that work.
Now we can argue about whether there is a better way for the government to encourage companies to do this research. For example, giving an X-prize award for a cure for certain diseases or conditions is one possibility.
One thing that should be clear to anyone is that if patents on drugs were abolished tomorrow, without some other legal framework taking its place, drug research would come to a virtual standstill. That would lead to a lot more "sickness and death."
You know different industries need different types of regulation. It makes sense to require federal approval to build a nuclear power plant, not so much sense to require the same level of approval for a hot dog stand.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 16:32 UTC (Fri) by pboddie (subscriber, #50784) [Link] Now we can argue about whether there is a better way for the government to encourage companies to do this research. For example, giving an X-prize award for a cure for certain diseases or conditions is one possibility. The patent system is quite clearly a government instrument. Replacing it with something that is actually fair would only infuriate the hypocrites who advocate low levels of government interference whilst clinging to the patent system as "protection" for their investments. The only controversy should be about whether governments themselves get to set the research priorities and whether "the market" is better or less corrupted by ideology (such as some governments insisting that HIV treatment wasn't worth prioritising). One thing that should be clear to anyone is that if patents on drugs were abolished tomorrow, without some other legal framework taking its place, drug research would come to a virtual standstill. This assumes that the only motivation for developing drugs is precisely that which seeks monopolies, presumably to exploit them for profit. In fact, there are obviously other motivations for doing drug research, and organisations with such motivations are in operation, somehow managing to co-exist with for-profit organisations within the arguably inequitable system of today.
A weak case for software patents (opensource.com) Posted Aug 14, 2010 11:58 UTC (Sat) by Wol (guest, #4433) [Link]
Patenting pharmaceuticals is harmful - YES!!!
Publicly funded research should be published and not open to patents.
Pharmaceutical companies are very good at publishing twisted studies - at a 95% confidence interval there is a 50% chance that one study in 20 will give a contrary result. How easy is it for companies to fund 20 studies and suppress the 19 that give the wrong result? :-)
IMHO, "independent invention" should prima facie be a proof of obviousness. But "indendent" needs to mean "you didn't know about the other work". So in a fast-moving FIELD patents will easily fail on obviousness. Several companies working in the same field? All coming up with similar results? Must be either collaboration or obvious, therefore no patents.
First-mover coming up with a surprise product, followed by a bunch of copy-cats? As a copy-cat, it's clearly NOT independent invention, and you get caught by patents.
Cheers,
A weak case for software patents (opensource.com) Posted Aug 13, 2010 1:16 UTC (Fri) by AndreE (subscriber, #60148) [Link]
"I simply don't understand what's so special about software"
Neither do I. Why is software protected by both patents and copyright? You cannot patent a style of writing or art form, and you cannot copyright the neumatic drill. So why can you both copyright and patent software, considering that there is no physical invention invovled in writing software.
"allowing every other discipline to use the patent system"
In my country at least (and in the US i'm 99% sure), only software and physical inventions are protected by the patent system. Literary works, art, films, trading names, and trade secrets are not protected by patents.
The truth is there is innovation in computing, and it is protected by copyright. Patent protection just extends the granted monopoly to the detriment of competition.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 3:34 UTC (Fri) by daglwn (subscriber, #65432) [Link]
Copyright and patents are very different things. Copyright doesn't protect the innovation. It simply protects the description of the innovation. Copyright does not prevent some other entity from reimplementing a novel algorithm. That's where the patent applies.
We can argue whether that's a good idea but you can't argue that copyright alone provides the same protection as patents. If it did we wouldn't be having this debate.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 5:37 UTC (Fri) by nirbheek (subscriber, #54111) [Link]
> Copyright does not prevent some other entity from reimplementing a novel algorithm. That's where the patent applies.
Mathematical facts and proofs, as well as algorithms (which are really an extension of the former), cannot be patented.
Algorithms CAN be patented, that's the problem. Posted Aug 13, 2010 16:35 UTC (Fri) by dwheeler (guest, #1216) [Link]
Sure, there are people who *say* that algorithms can't be patented. But all software patents are patents on algorithms, and there are a lot of software patents, so it doesn't matter that people *say* that algorithms can't be patented; they manifestly are. The problem is that lawyers have created a massive loophole, where officially algorithms can't be patented but in practice they are patented with a vengence. This greatly enriches patent lawyers, and severely harms society. And, as Knuth and others argue, this needs to STOP.
A weak case for software patents (opensource.com) Posted Aug 13, 2010 14:11 UTC (Fri) by jjs (guest, #10315) [Link]
The argument isn't that copyright and patent are the same. The argument is that software is the only field where you can get BOTH copyright and patent. As mentioned, you can't patent a novel or a storyline, and you can't copyright a pneumatic drill or a car.
So why is software treated differently? Why should software get the protection of BOTH copyright and patent?
|
|||||||||||||