Guest Post by Martin Goetz
This article is in response to the editorial “Abandoning Software Patents” by the Ciaran O’Riordan, Director of End Software Patents (posted on PatentlyO on November 6, 2009) which had as its premise that one is trying to protect “software ideas”.< ?xml:namespace prefix ="" o />
I wrote this article from a unique perspective as the holder of the first Software Patent in 1968 and with a long history of involvement in the protection of software thru patenting and copyright protection. I am recognized as a pioneer in the Software Products Industry and had a successful career at Applied Data Research (ADR), the first company to market a software product. ADR and I were also directly involved in the filing of amicus briefs in the Prater & Wei, Benson, Johnson, Flook, and Diehr cases. More background information and links to my memoirs are in Wikipedia at http://en.wikipedia.org/wiki/Martin_Goetz.
Since the 1960s I have been a strong advocate of the patenting of inventions implemented in software and in 1968 I received the first < ?xml:namespace prefix ="" st1 />
This article does not argue for or against the patenting of BPMs. Rather, it tries to explain why inventions implemented in software are well within current US Patent Law[3] using examples and analogies that I believe are irrefutable. It also explains why software should be viewed a machine component of a general purpose computer (a machine).
1. What is a Software-Related Invention? It is well recognized that whatever you can design in hardware circuitry (chips) can be developed in computer software (a computer program) to perform the same functions. Handwriting analysis, voice recognition, video frame analysis, data compression, language translations, artificial intelligence, searching techniques, network monitoring and security — to name just a few functions — are examples of where such implementations have been done in both hardware chips, in software, and a combination of both. The patents that have been issued in these nine areas represent inventions that are very-state-of-the art and not at all obvious. In particular, the analysis of handwriting and voice by a computer — whether in hardware circuitry or in software — is very complex and not at all obvious to one skilled in the art.
2. Hardware implementations versus software implementations. The choice of implementation for computer functions is a pure economic choice which mainly has to do with cost, speed, and flexibility. Patent applications normally show the preferred implementation and the patent must disclose the invention adequately for one skilled in the art. But the disclosure could be in the form of circuitry for a hardware implementation or a flow chart for a software implementation or a combination of both. Many professionals view software development as building a software machine. The life cycle of computer software is very similar to the life cycle of computer hardware. And its life span can be equally as long.
3. Software Product Companies[4] are High Technology Manufacturing Entities. In the 1980s with the advent of PCs, many new PC software companies called themselves Software Publishers. So are software companies more like publishers of books or more like manufacturers of machines? Based on my many years in the software Products Industry[5], here are my arguments why software product companies are manufacturers of high technology products.
Many software products are state-of-the-art products developed in a very competitive, fast moving environment and require rapid response to meet user demand. Secondly, a great deal of capital is often required and many software companies are funded through private investments, venture capital, and through public offerings. Thirdly, there are active research and development activities within these companies. IBM, as an example has reported that it consistently spends well over one billion dollars in research and development specifically in the software area. Lastly, highly skilled personnel are employed in these companies and many have advanced Computer Science college degrees, including PhDs. And because of its complexity, many software products are built using software engineering disciplines.
There is six phases in the life cycle of software products: Definition, Design, Implementation, Delivery, Maintenance, and Enhancements. Let’s look a little closely at these phases and you will see how closely they resemble characteristics common to all manufacturing companies. Often, prior to the definition phase there is research as well as competitive analysis. During the definition phase software companies describe its functionality, its specifications, the environment in which it must operate, and its operating characteristics. During the design phase, it develops and defines all its interfaces, breaks down the functionality into modules, and does all the engineering so that the product can be properly implemented, maintained and enhanced during its lifecycle. During the implementation phase the software is debugged, tested, and goes thru quality assurance. During the delivery phase there is alpha and beta testing, documentation, installation, and training. Often software companies OEMs (Original Equipment Manufacturers) the product to other companies where the software becomes a component of a larger system and is re-packaged. During the maintenance phase the company warrants its workmanship, and guarantees the correction of errors and defects. Finally, during the enhancement phase the software is improved, enhanced, upgraded, and new models (releases) are announced.
Note these terms indicative of a manufactured product …..research, competitive analysis, functionality, specifications, operational environment, operating characteristics, interfaces, modules, engineering, implemented, debugged, tested, quality assurance, alpha and beta testing, documentation, installation, training, OEM, component, system, re-packaged, maintenance, warrants, workmanship, guarantees, errors, defects, improved, enhanced, upgraded, and models.
It is obvious that software products are not “software ideas”. But is software more like publishing a book or more like manufacturing and maintaining a machine? And if it’s more like a machine how can the Supreme Court deny the patenting of inventions[6] in software?
I believe the Courts should view software as a component of a general purpose computer (a machine) and that software transforms a general purpose computer into a special purpose computer (or machine).
[1] Patent # 3,380,029 Sorting System Issued April 23, 1968
[2] The Diamond vs. Diehr landmark decision by the Supreme Court in 1981 opened the door to the patenting of software when the court stated that “processes” were patentable, and that just because an invention used a formula, program, or computer, it was not necessarily unpatentable.
[3] The
[4] The Software Industry is made up Software Product and Software Service companies and is recognized as one of the top three manufacturing industries in the world with 2008 revenues of over 700 Billion dollars. See http://en.wikipedia.org/wiki/Software_industry
[5] My tenure in this Industry started in 1954. In 1959 I was one of its founders and spent almost 30 years at Applied Data Research (ADR), a publicly traded 200 Million software company that was acquired by Ameritech in 1986 and sold to Computer Associates (CA) in 1988. Many of ADR’s products initially developed over 30 years ago have been improved by ADR and CA and are still viable today. Today, I am a consultant and investor in software companies.
[6] An invention, in lay terms, can be a novel device, material, or technique which is new, inventive, and useful. It has been well established that machines, including computer hardware, contain patentable subject matter.
Icon construction chemicals are a Manufacturers and Suppliers of Construction Chemicals and Water Proofing contractors in Bangalore India.
“That must be why computer programming is taught in the math faculty. NO IT ISN’T. Math and logic are not the same thing. Applied math and applied logic are not he same thing as pure math and pure logic. Computer programming is a branch of engineering. It’s about getting machines to perform useful functions.”
The area computer programming is taught varies with the university.
link to google.com
That’s the sort of thoughtful contribution, from somebody who is not a patent practitioner, that I welcome. Thank you Chavoux Luyt.
There seems to be two different arguments going on here and I don’t think they are the same:
1. Is software patentable under current law? (“exactly what our forefathers meant in the Constitution”)
2. Should software be patentable?
It seems as if Mr. Goetz’s aticle only addresses the first question. “Software development is similar to general engineering, so it should be patentable.”
Question 2 is about whether the object of “exactly what our forefathers meant in the Constitution” are being met by software patents.
Scientific discovery of natural laws and mathematical algorithms/formulas (the fruit of scientists’ labours) are explicitly excluded. The implementation or use of that knowledge in a “useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof” is patentable. But it is still possible for anybody to use the laws of gravity or electronics freely develop any new “useful process, machine, manufacture, or composition of matter” of his own. And this is the problem with software patents. Because it is simply an mathematical algorithm implemented in some computer language, it can effectively prevent me from using that same algorithm to develop useful processes, machines etc.
Copyright could not protect physical inventions. Inventors could either keep something a trade secret (meaning that the knowledge of how their invention works dies with them and never becomes public knowledge) or make it publicly available (meaning the inventor doesn’t get any reward for his work). That problem was (imperfectly) solved by the patents option: making it possible for the inventor to make known how his invention works, but still being rewarded for it.
However, copyright can protect software. Implementing an algorithm (= writing software) is a lot of work and can be done in a lot of different ways. Copyright prevents me from stealing the hard work done by somebody who implemented an algorithm to solve a problem. Software patents, however effectively prevent me from using that known algorithm to solve the same problem.
For all those harping on software not being mathematics: I don’t think anybody (any programmer) will have a problem with software patents if it would still allow me to use the mathematical algorithm (nothing more) included in the patent to solve a problem. :-)
The current situation is such that probably all programmers has already used some patented process in their software without knowing it… and I can not see how that could ever have been the aim of patents in the first place.
I am sorry, but life cycle of software can’t possibly be nearly as large as the life cycle of hardware.
Also, if systems people view their role as making software machines, go ahead, make systems patentable, but keep off single things that are ridiculous to be patentable: Algorithms (which are a part of math) and single UI elements and visuals (which are way too trivial to be considered a software invention and in no way should be patentable)
While we are on it. Make sure that when patenting a “software invention” the source code of such invention is included. How else are people going to be able to make sure not to implement such invention? Patenting worked a long ago for machines because it was easy to recognize reuse of someone else’s invention. It is unfortunately ridiculous to think that we can be ever be sure that when coding -say an OS- you are not implementing one of the thousands of patents from the likes of MS and IBM, specially considering they do NOT provide source code…
“you’re not that smart,”
6, you are not qualified to make such a determination…
I think so, yes.
there’s also usefulness of a method.
wikipedia shows a novel and non-obvious exmaple of entertaining a cat with laser machine and claim that it is not useful (argument looks very odd to me). Well, if I attach a condition that the owner of the cat is entertained/happy by looking at cat playing with laser beams — that is clearly beneficial to humans.
Does it become useful to be patentable?
Frith,
PHOSITA is just one of the required properties for patent. Sure, there are many problems on what is sufficient and what is not, because it’s subjective human process, but at least it has some practical hint on how to do measure whether patent is disclosed enough.
However there are other criteria that need to be fulfilled:
1) novelty
2) non-obviousness
Non-obviouseness can also be bound to a human process like PHOSITA, because it is clearly subjective and imperative from human position, there is no natural/objective way to measure it by definition.
Novelty is not so easy. We can go though prior art — OK, that’s pretty concrete, but how do you measure what is novel enough? is replacing variable names make it novel enough? is changing loop structure from “for” to “while” is novel enough? Swapping independent sequential operations? Ultimately just for comparison reasons one would convert it into abstract representation, using some formal/mathematical language. If you are lucky (to avoid undecidability problems) you can then deduce that perhaps two algorithms implement the same function (mapping from inputs to outputs), then you have a result that those two algorithms compute the same things, but this still does not answer the question: are these methods the same or are they different? Do you allow different ways of computing the same thing?
e.g. I can invent a method X to produce Y, somebody else invents Z which produces Y but in different ways/procedures (e.g. more efficient way).
Does Z method infringe on X patent?
The same argument applies on discussion where software+generic_CPU and dedicated_hardware achieving the same effect are compared. They both produce the same stuff, but are they infringing on each other? The structure of the machine reveal that their approaches are different by definition. Lawyers may not care about that fact, because they both appear as black-boxes to them and all they care is subjective fairness, a kind of observable equivalence which make their black-box argument valid, then engineer comes along, opens the box: look they are different machines.
In the material world, no two machines are completely the same, so if you open the machines sufficiently you can prove that they are different, some electron on some atom has a different spin!
If you consider software at abstract level, then apart from “math objections”, you will also face algorithm analysis decidability issues.
@bobo,
Point taken.
You said the checkout girl is doing “maths”.
Fact is, you don’t know what’s going on in her head and neither do I.
But if I had to take a bet, I would bet that what is *not* going through her head is something like: “Well let’s see if I apply Schrodinger’s equation to this current situation, obviously the photons are transitioning from the E1 state to the E2 level and …”
More likely what is going on her head is: “This is a boring job and I can’t wait til my shift is over. Hey look at that cute guy waiting in the back of the line …”
Cheers.
@step back
“The high school girl at the bar code automated checkout counter did not build the laser scanner or the software or anything else. There were other people who did the hard hard “maths” work (perhaps over them there “internets”) where that hard work led to formation of the checkout system at hand.”
What, like Lemelson you mean? Ha!
“These other people may have had PhD degrees in optics, in electronics, in system operations, etc. They were “inventors”. Maybe they got patents for their contributions to society and maybe they didn’t. But don’t go taking the credit away from them and mis-allocating it to the checkout girl. ”
I don’t know whether you are misunderstanding my point on purpose or not. I did not attribute any credit for any inventions to the checkout girl.
Whilst we may have opposing points of view and I appreciate your sense of humour do you have to pretend I said something that I didn’t?
Hmmm.. I think I will have to revise my position a little. I am getting most of my impressions on patents from the few actual patent documents I have read and the various discussions/websites on the subject. Still, I think that a law should be easily understandable (another utopia I won’t be holding my breath for!) in order to be considered ‘good’.
And I do think there is a place for patent law, but its application should be carefully applied in order to promote innovation.
I looked up PHOSITA, and as near as I can tell, the application for a patent must be clear enough that a PHOSITA would be able to implement it. This would make sense to me – you must be able to achieve your claim (otherwise you could patent all kinds of strange things and sue R&D departments who are actually trying to achieve these goals).
And just by looking at a few patents (a great way to find out how something works by the way!), it seems to me that just by reading the patent and having some skills in the area you should be able to construct the invention.
So by that standard, actual source code would not be required. But something pretty close would be needed. I mean, personally I can’t really see how you could get away without providing at least all the API’s the software uses. It might be argued that low-level design docs could get you over the line, but in many instances these essentially consist of the API’s.
Still, this would be a pretty fair approximation of my orignial ‘utopia’. As all the interfaces are publicly available, anyone can create their own implementation (obviously for commercial purposes they would have to license the interface, or create something different and compatible)
Thinking about this some more, I could see this as a great way for software to move into the future where digital copying makes physical scarcity impossible. Software companies could license the interface and the implementation separately – the bonus being that the (patentable) interface would be implementable by anyone, and hopefully this would be a fairly cheap license.
That would be pretty awesome really… (but again, not holding breath)
Noise above Law:
“I wanted to see what your answer would be to an extended version of the same question:
Would you say that software running on a virtual machine is fundamentally different from the same software running on a hardware machine, as well as if either is fundamentally different than a fully dedicated and hardwired machine with no software running to the same invention?”
Software running on a virtual machine and the same software on hardware are equivalent. They are both information in a computer, and they both implement the same algorithm.
Your hardwired/no software machine I would say is different, since the logic is expressed with physical components rather than information. It might not be able to run the same algorithm, either, since most algorithms require the use of memory space, which is disallowed by the ‘no software’ part of your definition.
There’s (at least) one other possibility — a general purpose computer with a hardwired memory and with any temporary rewritable memory space inaccessible from outside. From the inside, it looks like a general purpose computer, from the outside it looks like a hardwired device. This, to me, is pretty much the definition of a special purpose computer. I can’t see a real problem with this sort of device being patentable (as a whole).
Mike,
I was waiting for peanut gallery to reply, but I’m not sure a reply is forthcoming.
At your Dec 02, 2009 07:15 PM post, you asked,
“Are you aware of the concept of a virtual machine? (a computer implemented not in hardware, but in software running on another computer) Would you say that software running on a virtual machine is fundamentally different from the same software running on a hardware machine?
I wanted to see what your answer would be to an extended version of the same question:
Would you say that software running on a virtual machine is fundamentally different from the same software running on a hardware machine, as well as if either is fundamentally different than a fully dedicated and hardwired machine with no software running to the same invention?
“is or is not maths as defined so here in such a nonlegal sense leaves nothing comparable to a legal determination.”
Would math in the legal sense be like the legislative attempted redefinition of the value of pi to exactly 3.14?
“is or is not maths as defined so here in such a nonlegal sense leaves nothing comparable to a legal determination.”
You are raising an aspect of the law that get me scratching my head.
In the context of patent litigation, is the question whether something is maths an issue of fact where expert opinion must be considered? Or is it an issue of law where expert opinion is not relevant?
If it is an issue of law then I agree you can dismiss what I say as not being the law and this is the end of it. But if it is an issue of fact you may have to ask yourself what an expert testimony will look like. It is probable that a competent expert on this matter will be a mathematician or a theoretical computer scientist with expertise in maths.
there is more than enough money in the software industry to pressure Congress into adding software to 101.
No doubt the Senators from Maine, Nebraska and Mississippi will be heavily influenced by their states’ powerful pro-software patenting lobbyists.
is or is not maths as defined so here in such a nonlegal sense leaves nothing comparable to a legal determination.
it is just more mental mathurbation.
the smart guys just want to hear more William Shatner lamentations.
@step back
“If you allow anything from the real world to leak into the purely theoretical/ purely mathematical realm, you allow uncertainty and nondeterminism to enter the picture.”
Even if you stick to “purely theoretical/ purely mathematical realm” nondeterminism is in the picture. Nondeterminism has no influence on whether or not software is maths.
Your last example breaks the rules on mechanical aids. Only pencil and paper is permitted. Spit is not allowed.
If the rules on pencil and paper are tightened up to be limited to reading and writing symbols in a deterministic manner as the mathematicians intended, nondeterminism is out of the picture. But I prefer to leave it in because real computers sometimes have nondeterministic behavior.
“Note that a perfect random number generator in the purely mathematical world is not identical to a real world generator. A real world generator could never be that perfect.”
We can do maths with nondeterminism without knowledge of the statistical distribution. The “perfection” of the random number generator has no influence on whether or not software is maths.
At the end of the day it is turtles all the way.
@PoIR
If you allow anything from the real world to leak into the purely theoretical/ purely mathematical realm, you allow uncertainty and nondeterminism to enter the picture.
Say for argument sake that you didn’t let me have the spinning pencil or the origami papers folded into dice and you only let me have a human being doing the work in his head. In that case you still let a physical human being enter the picture. I can then have the human perform any of a large number of tasks with his body that create nondeterminism. For example, I can have him drool some spit from his mouth and count in his head the time it takes for the spit to break surface tension and hit the ground. You know … 1 Mississippi, 2 Mississippi, 3 … No “machines” involved, and yet the human body is a machine whose parts obey the laws of physics, including the uncertainty principle.
Note that a perfect random number generator in the purely mathematical world is not identical to a real world generator. A real world generator could never be that perfect.
“Go talk it over with the “leaders in your field”, they’ll explain it to you and they might decide to explain to you how he cheated. ”
I don’t think I was cheated. Next time someone tells me the computer isn’t carrying out the same process as a human with pencil and paper because computers have nondeterminism I can raise the spinning pencil and prove him wrong.
Nondeterminism is allowed in mathematics. It is not included in Turing machines though. The spinning pencil will probably bridge that gap. I am consulting to find out exactly how. I am careful. I don’t jump the gun and say something silly. I check my facts first.
Most of the arguments are about whether there should be any right to patents or to software patents in particular. The first was answered in 1789. Madison took the position taken by the anti-patent people. Jefferson wanted a complete legal monopoly. They compromised and limited the time. Maybe it was a bad bargain but it was the bargain nonetheless. Maybe a right against self-incrimination or forced confessions hinders the police in fighting crime. Maybe the right of free speech permits people to use profanity in public. To argue against these rights is senseless. Those are the rules we work under.
As to software patents, if the SCt rules that software is not patentable matter, there is more than enough money in the software industry to pressure Congress into adding software to 101. We will be back where we are today.
Who am I?
Posted by: Champernowne Constant | Dec 02, 2009 at 07:56 PM
Oh Oh Oh!
I know I know
You’re Santa Claus!
“I disagree.”
I don’t care if you disagree, you’re not that smart, and you apparently don’t even understand your own thesis that you just wrote the other day. Mostly because you were just parrotting what others have already said. Go talk it over with the “leaders in your field”, they’ll explain it to you and they might decide to explain to you how he cheated.
And step, new Kirk got caught. He also got caught in the original iirc.
It’s hard to bs a bser.
MM the baboon said: >>Are you f*ckin’ nuts?
To a baboon yes. To a human no.
@bobo
Once again your crowd of blind men round the elephant want to mis-allocate credit for a job well done away from the person who did the actual work (the inventor) and to someone who is free-riding on the coattails of the inventor.
The high school girl at the bar code automated checkout counter did not build the laser scanner or the software or anything else. There were other people who did the hard hard “maths” work (perhaps over them there “internets”) where that hard work led to formation of the checkout system at hand.
These other people may have had PhD degrees in optics, in electronics, in system operations, etc. They were “inventors”. Maybe they got patents for their contributions to society and maybe they didn’t. But don’t go taking the credit away from them and mis-allocating it to the checkout girl.
Just because you and yours don’t want to give credit where credit is due, that doesn’t mean the rest of us have to join your gang of willfully blind bandits dancing around the bobo elephant.
We have eyes wide open and we do “recognize” what you are trying to do.
@peanut gallery
“The commenter who said that writing a computer program is doing math has no idea what he/she is saying.”
The girl on the check-out might not realise she is doing maths when processing your items, but she is.
When a pitcher throws a curve ball is he doing physics?
When a cook creates a dish is he doing chemistry?
Just because you don’t recognise something a little more fundamental in what you’re doing does not mean you aren’t doing it.
phantom gordian knots – its not a knot if the one string is the law and the other is what you want the law to be, but is not. you only knot yourself when you try to conflate the two.
so-called smart people superficially shoehorning amorphous definitions of maths into actual law to serve an agenda.
are we quite done yet with the mental mathurbations?
PoIR,
“By the way, how do you do maths if you don’t write any symbols? In your push button or diabetes examples the people find out about the result of the calculation because the result is written or otherwise represented with appropriate symbols somewhere.”
Let me be clear, if someone can start with initial information and use pencil nd paper to arrive at a value that determines what button to push or whether someone has diabetes, that is a process that is patentable.
stepback: It’s all about the money. You and your anti-patent cohorts want to exploit another person’s invention without compensating them and without getting their permission. That’s what this really all about.
Are you f*ckin’ nuts?
@PoIR,
When the average lay person walks into Best Buy or wherever to buy a CD on which “software” is inscribed in the form of pits along tracks or phase changed areas, that person is not buying “maths” (whatever that coined word means). He is buying a tangible physical object that he (or she) will insert into the CD player on his/her computer so that the computer takes on a different configuration and so that the computer will appear to be working as a new machine for that user.
The user is not going to be doing “maths”. The user is going to be playing an action video game, or composing a musical piece with aid of software, or designing a Field Programmable Gate Array (FPGA) with aid of provided design software (CAD), etc.
That’s the user’s definition of what useful software is. You have your own Chesire Cat definition and bless you for that. But what does your Alice in Wonderland definition have to do with the legal landscape? Please just keep your definition in Alice’s Wonderland, where it belongs.
I did, BTW, much enjoy designing the spinning pencil game. Thanks
6,
“You straight up broke the rules and he didn’t catch you”
Does that mean I get my name enshrined next to that of James T Kirk for having defeated the Kobayashi Maru scenario? Wow! How cool is that?
–LOL
“You didn’t, you merely pulled one over on POIR by utilizing “ingenuity” in all of your examples even though it is forbidden.”
I disagree. Ingenuity is permitted when making up the steps for the method. It is in the application of the steps that it is forbidden. Using a pencil as a random number generator requires only moving the pencil and reading a number. This is no more ingenuity than what occurs in ordinary arithmetic calculations.
“The pencil and a paper is provided to make calculations only, not play with in any fashion involving ingenuity that you can come up with.”
This is the implicit intent. But there is a valid point in observing that a rotating pencil random number generator doesn’t require more ingenuity that this.
@step back
“You keep insisting that we are out to get you for engaging in academic exercise of free thought and in transmission of mere ideas.”
This is not true. I am discussing whether or not software is maths.
“But that of course is a badly feigned ruse. It’s all about the money. You and your anti-patent cohorts want to exploit another person’s invention without compensating them and without getting their permission. That’s what this really all about.”
This is baseless accusations. No exploitation of other people’s work is wanted.
If it is found that software is unpatentable mathematics then there is no exploitation of other people works because: 1) patents are not applicable to ideas and 2) software provides ample protection. The proper course of action is to let the controversy follow its course until it is resolved. Then the dice will fall where they belong.
You seem to deny the possibility of a controversy with a genuine basis. I think you are wrong on this. Suppose for a minute that like I argue the mathematicians’ view of mathematics doesn’t match the tests lawyers use in patent law. What happens? Can’t the people who notice speak out? Isn’t there an issue of finding out who has the right definition? Isn’t there an issue of whether abstract ideas are actually patented even though patent lawyers don’t intend to? Such controversies are legitimate.
“I didn’t mean to upset your perfect theory of the perfect Computation Theory. ”
You didn’t, you merely pulled one over on POIR by utilizing “ingenuity” in all of your examples even though it is forbidden.
The pencil and a paper is provided to make calculations only, not play with in any fashion involving ingenuity that you can come up with.
You didn’t mess with his words or screw around with syntax or blah blah blah, you straight up broke the rules and he didn’t catch you.
Either way, I don’t really care for his argument anyway.
@PoIR,
Patent lawyers are not interested in getting patent claims for wholly preemepting abstract thought. If you want to engage in abstract thought, go ahead and do so. I’m not stopping you (and I can’t).
You keep insisting that we are out to get you for engaging in academic exercise of free thought and in transmission of mere ideas.
But that of course is a badly feigned ruse. It’s all about the money. You and your anti-patent cohorts want to exploit another person’s invention without compensating them and without getting their permission. That’s what this really all about.
The “Bilski” case has no software in it. Those who are against their own self-generated idea of what a “software patent” is, are hoping that the US Supreme Court will fashion a definition of the phrase “any new and useful process” so that the new definition leaves out ALL new and useful computer-implemented operations (where by computer-implemented, I mean implemented with use of a real world physical computer and not some fantasy Turing mechanism having massless tape of infinite length and infinite advancement speed).
I didn’t mean to upset your perfect theory of the perfect Computation Theory. Just change your real world paper and pencil to virtual paper and pencil, change your real world person to an immortal hypothetical person and you be able to escape my lawyerly clutches. The above argument about the spinning pencil was just shark’s chum that I intended to throw back in the sea. I meant no harm to your Computational Theory. Sorry if I upset you that much.
_______
BTW there could be many variations on theme. I could have used Origami skills to fashion paper dice out of the real world paper; or I could have just cut up (oops torn I meant) a bunch of paper squares and flipped them to land heads or tails up with different symbols on opposed sides. The pencil itself was not essential to rigging up a nondeterminstic mechanism.
I am the child of David Champernowne, Alan Turing’s good buddy. I own all possible software and data. I can emulate a hard drive infinitely bigger than the universe, using an equation that fits in one line of code. I also have all possible music files. My father was a musician, and I am prior art to every CD, MP3, WAV. I am very musical myself. I can sing any song by any singer in any language that was or ever will be sung or performed. I am doing so at this very moment. I know exactly how many 4-minute sounds there are and have played every single one. Nobody has ever made a digital recording that I have not already made. My one line equation proves that by telling you exactly when and where I played a newly recorded song before it was sung. SONY can not sue me for infringement because they owe me for damages to my computer which was hacked into via their rootkit. A 6 year old child has been listening to me making my music since the 1970s. My work is all public domain. The child has grown up and is furious that people are suing people for listening to his favorite songs without me. So he has been telling everyone how to make my music without copying it. He has been selling software and machines that make my music, but not copies of my music. All digital music was made first by me. And I can tell you how to make it too, for free. A one line equation, all simple arithmetic. The 6 year old child who listened to me knew my equation. I also have written every possible computer program and all of Shakespeare, so I can even teach it to a Monkey. You don’t even need a computer to do my math and see my video and listen to my songs and read my books, because an infant and a small machine can easily imitate me. Only God has made more things than I have. I am free. Who am I?
@step back
I did respond to this part of your last post.
“Much in the same way that I twisted the meanings you intended about “pencil and paper”, you and your anti-patent friends are twisting the meanings of what we are trying to convey from the pro-patent side.
No one on the pro-patent side is saying that patent law can be used to suppress your Constitutional right of free speech, and what inherently goes with it, namely free thinking; including thinking in terms of mathematics.
No one on the pro-patent side is saying that patent law can be used to suppress the free flow of “ideas”.
One cannot patent a raw “idea”.”
Of course no one on the pro patent side wants to suppress the free flow of ideas. The issue as I understand it is not what the pro patent camp says or wants. It is whether some patents will have the effect of suppressing the free flow of ideas.
When you consider that maths is made of symbols having a meaning, it is conceivable that patents on mathematical processes, the real world kind with real world symbols, raises issues of free speech. I am not competent to discuss these issues. I can only point out we are in fertile grounds for such discussions because symbols and meanings are constituents of speech.
In my humble opinion the best cure is exactly the one you point out. One should not patent a raw idea. But I see there is a disagreement on a critical issue. What constitutes a raw idea? Are patents on real world processes that manipulate symbols the same as patents on raw ideas? This issue was discussed tangentially in Benson except that SCOTUS did see the Benson patent was processing symbols in this particular case.
I think I did acknowledge what was said by the pro-patent camp on 101. I have acknowledged that the plain reading of patent law means real world processes are patentable. This is one of the prongs of my Gordian knot observation.
Whoever said it:
“The key question (to my mind) is whether an algorithm implemented on a general purpose computer is essentially different from the same algorithm implemented manually. If it is not, I suggest that you need to look for patentable subject matter elsewhere.”
OK, so the problem is that the algorithm may be applied manually and we don’t want or can’t patent a manual process which would contradict to free will/expression.
Here is a trick the lawyer would apply:
claim this algorithm on >511bits.
That’s it. If the algorithm is too simple, he can always push the limit of required memory even further until it becomes not feasible for manual/mental/pencil labor, thus if the algorithm novel and is useful (and it is very likely to be useful when lots of memory is needed) it becomes patentable as any other invention. Also if the author applies for the patent he automatically surrenders copyrights. See, no magic and everybody is happy. For real examples look for patents on cryptography.
I believe there are many many more legal tricks like that, and a good lawyer has a big collection of those.
P.S. I am playing devil’s advocate in this post, IANAL and I might be wrong.
peanut gallery:
“Humans can simulate the actions of a computer, but not actually perform them.”
This confuses me. How is simulating the actions of a computer not actually performing them?
Are you aware of the concept of a virtual machine? (a computer implemented not in hardware, but in software running on another computer) Would you say that software running on a virtual machine is fundamentally different from the same software running on a hardware machine?
@step back
I have always listened to facts. I just did right in front of you.
You think you played a game with words and meaning. I noticed that. I didn’t want to go into this discussion because I don’t think it would help address the substantial issues. This is not a game of who has the best rhetoric and what kind of rhetoric is permissible. I think facts are more important. And in case you didn’t noticed, I saw more than a game of words and meaning in what you said.
You raised a fact, one I didn’t expect and found relevant. You pointed out that one can use pencil and paper as a random number generator. You pointed out this is matching the definition of effective method. And you pointed out that computers likewise have sometimes a non deterministic behavior. These points are facts. They are important ones because they may have consequences on the accuracy of the Church-Turing thesis. I acknowledged them. I thanked you for teaching me something. And I have to follow up on them with a specialist in the field. And I responded to the facts you brought with more facts.
Don’t you see there are facts behind my points too? The definition of effective method has been written by a mathematician for mathematical purposes. It has not been written to withstand an assault from a lawyer. If you make abstraction of this deficiency, what is your answer on the facts I raised?
Paul says:
“OK, here’s one for all you legal geniuses out there. If I take a patent claim which is purely a description of an algorithm”
Only the most hard core are going to object to applying 101 to invalidate such a claim unless there is some dipute about whether the description is actually an algorithm. The more interesting and controversial cases are pretty far removed from your example.
@Big Guy,
Also you should mention that there is no such thing in the real world as a “general purpose” machine (GPM). It is as mythical as is the programmer’s man month and the Turing machine.
“BASIC is an “ALL PURPOSE INSTRUCTION CODE” for a “GENERAL PURPOSE MACHINE” for TELLING IT WHAT TO DO. Therefore, all that has been or can be written in BASIC is unpatentable.”
Steel is a versatile raw material for fabricating lots of different gadgets. Therefore, all that has been or can be fabricated from steel is unpatentable. Gee, non sequiturs are fun!
link to diva.library.cmu.edu
This is a link to Allen Newell’s paper that NWPA mentioned. It’s the OCR’d version of the PDF; the link to the PDF is at the top of the page.
I don’t see anything in it about ‘not being able to figure it out’ exactly, though he does acknowledge being a CS type, rather than a lawyer. The basic premise of the article is basically the same as PolR’s ‘Gordian knot’ point.
It appears to be missing the second and third pages, but it has the rest, and it is a very interesting and well-written article. I have only read it very quickly once, I plan on giving it another, more thorough read later.
@PoIR,
I think you are begrudgingly having to admit to having slightly more respect for us stoo-p-id patent attorneys who beforehand appeared to not know anything about anything when it comes to computer science.
But you are still missing the point I was trying to make. I can play games with words and meanings just like you try to. I understand what you were trying to say about pencil and paper. However, as a lawyer, I quickly saw that this was an opening for twisting what you were trying to say on Grocklaw about Computational Theory. So as a dyed-in-the-wool lawyer I couldn’t resist pouncing on the opportunity.
Much in the same way that I twisted the meanings you intended about “pencil and paper”, you and your anti-patent friends are twisting the meanings of what we are trying to convey from the pro-patent side.
No one on the pro-patent side is saying that patent law can be used to suppress your Constitutional right of free speech, and what inherently goes with it, namely free thinking; including thinking in terms of mathematics.
No one on the pro-patent side is saying that patent law can be used to suppress the free flow of “ideas”.
One cannot patent a raw “idea”.
We have repeatedly pointed you to the written law, 35 USC 101 and you have repeatedly refused to acknowledge what we are saying. You have your own model of how the evil patent lawyers operate and no amount of facts is going to dissuade you from your chosen set of beliefs, wrong as they are.
BASIC is an “ALL PURPOSE INSTRUCTION CODE” for a “GENERAL PURPOSE MACHINE” for TELLING IT WHAT TO DO. Therefore, all that has been or can be written in BASIC is unpatentable. Now some companies use non-GPMs for Specific purposes. This in no way prevents the use of a GPM for all purposes. Software patents cannot stand on GPMs because a GPM cannot be legislated into a Non-GPM. If businesses want to sell each other machines that can’t do anything but play Solitaire, well what is that worth, when individuals already have GPMs that can do everything including play Solitaire.
The point is really, a person is told to do math. The person can use a calculator to do math. The person can give a list of maths to do for a computer to automatically calculate. What you see on TV or a computer screen isn’t really there. On TV it might be somewhere else. On a computer it is calculated.
Before electronic computers and GPM types were invented, doing math and following a list of rules was a job for a person. The machine replaced that person. The machine has already been invented. You don’t invent anything just by telling a machine or a person to do it. And again, software patents do nothing but tell you what you CAN’T tell a machine or person to do, whose job it already is to do that. The end result is like inventing new colors of nails that are illegal to hit with a hammer without your permission, thus all the carpenters can only freely build houses of cards (play Solitaire).
@breadcrumbs
The point I am making now is not about applied maths. I am saying there is no way to do maths at all without some sort of real world process.
In terms of established law the trilogy Benson, Flook and Diehr has not been overturned. In Benson and Flook there are some real world processes that have been found non patentable because in the opinion of SCOTUS they are maths. Some of the posters here disagree with these rulings but they are still current law.
Diehr states the exclusion of maths processes still stands. The Diehr patent was allowed because it was found not to be maths and the mere use of maths standing alone doesn’t disqualify a process from patentability.
During our discussions back then I did accept that the application of maths could be patentable when the requirements of patent law are met and I didn’t change my mind. The application of maths is not the same thing as maths itself. Diehr makes this clear. I have always agreed with this view.
My point about the Gordian knot is about what happens when the real world process itself is maths, not when it is an application of maths.
As for my comments on Groklaw I made a mistake and learned the lesson. I will change my tone. But I will keep arguing my opinions. There are cases where rulings have been made based on an incorrect understanding of technology and I intend to keep pointing this out.
“If you are going to take the position that a process that transforms represented information is maths, then why can’t a process that transforms a grain of wheat be maths?”
Because of that pesky definition of “maths”?
Oh yeah, that.
“The commenter who said that writing a computer program is doing math has no idea what he/she is saying.”
You’re right, it’s actually more akin to writting a novel :(
“Mr. VariableX can you please be compared to Mr. VariableY?” Said the computer programmer to Mr. VariableX. “After that, could you please move over to register position 100?” “I thank you kind sir that will be all for now”.
It’s going to be hilarious when someone comes up with a widely used system that uses as its computer programming language simple English. Perhaps then this bogus nonsense about software patents will go away.
“I have said it before in this thread. In this intersection of real world and maths there is a Gordian knot that can be cut in two ways. Either the real world aspects trump the maths aspects and mathematical processes are patentable, or the maths aspects trump the real world aspect and some real world processes are not patentable because they are maths. It is a choice the law must make. But whichever way the law goes it will remain that some real world processes are maths.”
PoIR – a year ago we danced to a similar tune – one called applied math. Why is it a year later and you are just now begrudging that real world aspects trump the maths aspects and mathematical processes are patentable. This is not new. This is established law.
Will I visit a groklaw page in the near future to see more derisive comments about the Law and those who understand it?
anonymous,
the reason why is that your line of thought is synonomous with the general anti-patent line of thought. Besides being rooted in philosophy and overt political overtones, there is no traction with that line of thought with serious practicioners.
Ah Pook: I have a degree in applied mathematics and computer science, and both programs were run by the faculty of mathematics. You’re wrong.
Everyone: How come there isn’t more interest paid to the question of the utility of patents brought up near the beginning of the conversation? If the evidence suggests that software patents (or patents in general) are counter-productive, why insist on them? What if not having patents ends up counter-intuitively generating more money? Patents are simply a legal construct with a specific end in mind, and should never take precedence over that end.
@step back
This is not more information than I want to hear. On the contrary this is going to the gist of the issues.
Nondeterminism alone doesn’t move a computational process outside of the bounds of mathematics because there is no requirement in maths that everything must be deterministic. On the contrary nondeterminism is explicitly included in many areas of mathematics. The reference to probabilistic Turing machines was meant to make this point. Paul Leopardi made it another way.
There is a difference between a machine and the computation it runs. The machine is not maths. The computation is.
I agree computations made by a computer are in the real world. Computations made with pencil and paper are also in the real world. How do you expect mathematicians to practice their art if they don’t use real world tools? Absent the real world tools nobody will ever be able to do any mathematics. The intersection of real world processes and mathematics is not an empty set.
I have said it before in this thread. In this intersection of real world and maths there is a Gordian knot that can be cut in two ways. Either the real world aspects trump the maths aspects and mathematical processes are patentable, or the maths aspects trump the real world aspect and some real world processes are not patentable because they are maths. It is a choice the law must make. But whichever way the law goes it will remain that some real world processes are maths.
The commenter who said that writing a computer program is doing math has no idea what he/she is saying. I would guess that their only lesson on computer science was from the Benson opinion, written by someone who also had no idea what he was saying.
“The key question (to my mind) is whether an algorithm implemented on a general purpose computer is essentially different from the same algorithm implemented manually. If it is not, I suggest that you need to look for patentable subject matter elsewhere.”
I find this question strange. First, if it is the “same algorithm” then how can it be “essentially different”? To me, the same algorithm means the same actions. Humans can simulate the actions of a computer, but not actually perform them.
The second and bigger problem is that it assumes that a manually-performed process is de facto unpatentable. Even under the Fed. Circuit’s Bilski opinion, that is not true. If the manual process transforms an object, it is patentable subject matter.
>>The computer is not the computational process. >>The computer is not maths. The computational >>process is.
You can argue that. But, I would say that it goes to what the definition of maths is. And, that if you take broad definition of maths then it likely encompasses many other areas of technology including for example grain processing, metal processing, etc. If you are going to take the position that a process that transforms represented information is maths, then why can’t a process that transforms a grain of wheat be maths?
“Software is already overly protected by copyright.”
Hah! Are you forgetting the copyright battles between Apple and Microsoft? I don’t think anyone at Apple was pleased with the excessive protection that their copyrights afforded to them.
@PoIR,
Irrespective of Paul Leopardi being an actual mathematician, his analysis of the “method” I put together (rather quickly I might say) is not fully correct.
But first let me say this directly to Paul: you bet your bottom coin flip I know what a “nondeterministic” computational “process” is. Matter of fact I have written patent applications for computer “scientists” on their methods for solving the problem.
For those out there who are not aware of this, computers do not (for a variety of reasons) always output the same answer for a same given set of inputs. This is known as non-determinism and it is a real thing (problem) that computer scientists are well aware of.
What it means for you is that when your bank runs your account numbers through its computers, the answers can come out different on different runs. How comfortable does that make you feel about the correctness of computer outputs?
Now back to where Paul made his missteps: He forgot to account for the weight of my blood spots as I marked my leaded (and now loaded) spinning pencil. He also made a false assumption of what menial tasks I can load into my unlimited number of LUTs (lookup tables). I can have my menial man servant fold one of the papers into a paper airplane (and I’ll include the step-by-step instructions if you insist) and have him launch that airplane into chaotically turbulent air and then perform next steps based on how the airplane flew.
The minute we introduce real world items like wooden pencils and cellulose paper into the “process” (yes I did think about having my man child rub the pencils together and start a fire) we are no longer in the realm of pure mathematics. we are now in the physical world. The “process” is real rather than purely abstract.
When we run an actual computer process in an actual computer (not in a fantasy and physically impossible “Turing” machine) the same thing is true. We are no longer in the world of purely abstract math. In a real world computer system, hard disks crash, cosmic rays flip memory bits, flip flops fail to flip through their metastable states, and so on and so on.
Probably more information than you wanted to hear.
:-)
Okay, as a proponent of patents on software-related inventions, I agree that the article ended without fully flushing out its arguments.
The gist of it is: software-related inventions are developed like other types of inventions, and produce the same results/benefits of other types of inventions (particularly hardware implementations of the same techniques). Therefore, there is no reasonable argument for treating them differently.
The fact that some of them can be invented and implemented in a day by one person is no different than many simple inventions in other fields. (some are the result of literally millions of man-hours)
The fact that sometimes the products that embodiy them have a short shelf life is no different than many inventions in other fields. (some have extremely long lives – the most valuable software-related patents are the oldest ones)
Saying that “software” is not a statutory category is like saying “engines” are not a statutory category. You don’t need to have a statutory category named after your field in order to be patentable. (e.g. there is no “bio-tech” category). Software inventions are protectable both as new and useful processes, and as machines configures to perform the new and useful processes. Storage media that stores software is patentable as a manufacture.
The Mad Hatter argues: “The argument has now moved from whether “software patents” should be allowed onto whether or not “patents” have any positive effect on society.”
The Mad Hatter must come out of the looking glass into this world. No court opinion (even Benson, which was the worst), ever hinted that the patent system should end. The current argument in this world is whether business method patents are good for society. It is true that we have largely moved beyond the issue of whether software is patentable. Since Alappat, the answer to that is yes. However, people that are stuck in the 70s Benson mentality still like to bring it up when occasion permits.
I don’t doubt that the benefits of patents as a whole is a hot topic in academic circles, but everything (no matter how unrealistic) is fodder for discussion there.
@NWPA
“Do you understand that just because a computable function is being implemented that does not mean that it is math.”
Maths is not only the function. It is also the computation itself.
Example: when you do arithmetic with pencil and paper like you learned in school, you are doing a computation and it is maths. Pocket calculators can do this sort of things and so can the computers.
“A computer is a machine that is implementing a process operating on represented information. It is not “math.”
The computer is not the computational process. The computer is not maths. The computational process is.
@step back
You probably have noticed by now, but just to make sure you don’t miss it, Paul Leopardi is an actual mathematician. His analysis of your method is correct.
The definition of effective method is not something I made up. It is how mathematicians saw computer algorithms in the 1920s and early 1930s before the works of Alan Turing.
You cannot read out the “in practice or in principle” clause from the definition like you do. Mathematical concepts are not contingent on human limitations. In the definition of effective method the reference to a human is a filter that is used because in the 1920s mathematicians knew no other way to separate what is computable from what isn’t. Nowadays we use a Turing machine for the same purpose.
Translated in Turing machine terms the method of spinning the pencil you have provided corresponds to a probabilistic Turing machine. This is maths.
link to itl.nist.gov
I agree the people that wrote the original definition didn’t have this spinning of the pencil in mind. Had you raised the point in the 1920s it would have been a valid contribution to the progress of mathematics. Perhaps it still is. Perhaps nobody saw that the definition of effective method literally allows this sort of things. Perhaps from now on mathematical textbooks will have to quote your procedure when discussing the Church-Turing thesis. I will raise the point to a specialist in the field.
I have learned something. Thanks a lot.
Newell, a famous theoretical computer scientist, tried to figure out patent law and ended up basically saying he couldn’t. I don’t have the cite handy, but he wrote a paper about it.
>>Further, my point is that today we have >>situation described as #1, and you guys are >>running in circles in search for that >>patentability method — whatever it is,
That is how the process of law works. Not just patent law but all of the laws. It’s an ugly business.
Night Writer Patent Attorney:
“Do you understand that just because a computable function is being implemented that does not mean that it is math. Do you think that you are performing any task that is not a computable function? Are you math?”
Yes, I do understand and agree.
So what?
I also agree that:
“a software is a method of transforming information”.
I also realize that this might overstep human rights, free speech etc, but I also found some tricks that allow patents still transform the information and not overstep those rights — and everybody seems happy.
I even try to help you to find a coherent way to resolve the problem, I even hypothesize that there exists a method for testing patentability of software. How much further can I come?
My point is that the patentability test method, you guys are looking for, is an algorithm itself, therefore it lends itself to laws of computer science. Therefore Gödel’s incompleteness theorem applies:
either:
1) you have complete/universal coverage as you lawyers tend to think that law applies everywhere — there is nothing wrong with that, but the consequence is that such system is inconsistent, incoherent, and logic does not apply there (universally as in premise).
OR:
2) keep the system coherent, based on logic, but not universally applied. And the way to achieve it is to define and limit the scope. The scope is still infinite, but it cannot be universal.
Further, my point is that today we have situation described as #1, and you guys are running in circles in search for that patentability method — whatever it is, it is going to be bound by decidability problems, and while doing that attorneys are going to waste everybodies resources without reaching definite (coherent) conclusions.
So it all boils down to the following:
the patent law as it is serves well only to lawyers, and it puts into mutual destruction the rest of the world.
Well, I can even accept that as natural fact, but do you think THAT is fair? is this justice? is this how the legal system is supposed to work?
@lionel Hutz
“In the absence of a greater context the result is meaningless. Mathematics is merely a language developed by scientists to describe physical reality. ”
You should talk to mathematicians about their art. You have something to learn about maths.
Mathematics is a discipline that stands in its own right. Sometimes it may be applied in a physical context, sometimes it is done for its own sake. The laws of algebra stand by themselves for example.
Mathematics doesn’t always describe physical reality. It may also describe other abstract ideas. Look for these keywords on wikipedia: “symbolic logic” and “metamathematics” for examples.
“Your position is like saying random keyboard tappings constitute a sentence. ”
I have said no such thing. For evidence that mathematics is a written discipline look for “formal systems” on wikipedia. You will see this is no random keyboard tapping. The various branches of mathematics use formal systems as their foundations.
By the way, how do you do maths if you don’t write any symbols? In your push button or diabetes examples the people find out about the result of the calculation because the result is written or otherwise represented with appropriate symbols somewhere.
I don’t have the information to make an opinion on whether or not your examples are patentable. It seems you are probing me on my opinions on useful results v. patentatibility. I will say that the useful, concrete and tangible test is unable to filter out abstract mathematics. Even before you get to consider the push button or diabetes steps, the written information from the calculation is useful, concrete and tangible.
>>Anony Mouse
Do you understand that just because a computable function is being implemented that does not mean that it is math. Do you think that you are performing any task that is not a computable function? Are you math?
A computer is a machine that is implementing a process operating on represented information. It is not “math.”
I can see why the term Agony Mouse is so fittng.
The comment as to source code ghost was not to you. Please read more carefully as it was to Frith. I mean seriously. Egocentric is one thing…
A hint for you – don’t speak to a lawyer about patent law when you are ignorant on the subject.
for those who want to see how deep the rabit hole goes for software patents, please lookup the most simple Turing-powerful machines:
The simplest Turing-universal machine:
link to blog.wolfram.com
it just needs 2 states and 3 letters and you can run *any* algorithm.
two counter machines: if the description of the algorithm includes just three elements, you can construct an algorithm that is Turing-powerful and hence impossible to analyse such algorithms in general.
If you are still with me, you have two options:
1) stick current state of patent affairs and be doomed doing infinite analysis and exhausting funds.
2) limit the scope of patent descriptions. in software case, this is very hard, in fact current state-of-the-art of computer science is balancing here: you can take the most primitive approach, like disallow two counter machines and you will kill all software patents, or try to find better formalisms like finite automata, petri nets, etc, but it can never be as general and as powerful as Turing machine (or generic computing device) — you will immediately fall back in to option #1.
Noise above Law:
“Look into a little concept called PHOSITA before you proclaim (incorrectly) that actual source code is required, and why you are so blatently wrong when it comes to the “standard” for hardware, as no design blurprint for hardware is actually required.”
Thanks for PHOSITA, but where are you getting supposedely mine “proclaim” about “actual source code” *requirement* ?
Are you talking to ghosts?
I mean seriously. I never claimed that source code is *required*, never. Some time ago I was suggesting that source code could be a way of disclosing the invention. So you say PHOSITA, OK, fair enough.
But how does that help in devising a (leagl) method for testing patentability?
Are you implying that PHOSITA-formulated requirement is enough? so we can take any algorithm, go through PHOSITA test and we grant a patent?
What about novelty and usefulnes? This is what the patent law is about, isn’t it?
Here are a few hints for you:
1) PHOSITA is too loose requirement for description of an algorithm if one needs to determine any property about algorithm, let alone patentability. In fact it is made purposely to be general.
2) if you want a feasible method of determining a property about the algorithm, you have to be very precise and specific what modeling concepts are allowed in the description of the algorithm, and the purpose of those restrictions is to limit the power of expressiveness of the description. You can see this is the exact opposite of PHOSITA.
So you either stick with #1 and be doomed in searching for properties and methods forever until funding is exhausted. Effectively you get patent trolls, corporate extortion and
Or you stick with #2, and revise guidelines on what are suitable formalisms/models/descriptions for software patents, which effectively limits the scope of (patentable) space of innovation. And I don’t care what they are: a source code, a diagram, picture or human language description — my point that PHOSITA is too loose stands independently.
There is no middle ground or third solution, go lookup Turing halting problem, Rice’s theorem and Gödel’s imcompleteness theorem. I did my part and learned PHOSITA, In Re Bilski.
Anything else?
I am not asking you to study whole computer science starting with Peano theorems, so don’t expect me to go to education on Law just to figure fundamental principles behind patents. And no, the article on Grawklaw does not go deep enough, it provides some technical details, I think they are irrelevent with respect to innovation in general.
In Re Bilski judges — the experts — disagree.
Feeling as being a better expert?
Touché MaxDrie,
I had in mind the typical sites of software and electronic arts (e.g. groklaw, cnet, zdnet)
Noise, just one point you might clarify for me please. You write:
“… do tax lawyers habitually visit the technical sites”
How does a tax lawyer know, when he is entering a site that is, as you say, “technical”? Such sites are not labelled as such, are they. Might he not (in his baleful ignorance) have supposed that his home site is just as “technical” to him as yours is to you?
If you can come up with a succinct definition of “technical” maybe the EPO would be pleased to hear from you. But then again, maybe not. The EPO has already told us (thousands of times) what it means when it uses the word “technical” in its decisions on patentability.
Dear step back,
Your effective method is perfectly legitimate, and non-deterministic. Although your insistence in 3. is peculiar, it is essentially the same as terminating in a random but finite number of steps, with some probability distribution, within some upper bound.
The answer to “what does the method compute” is essentially some character string in the alphabet you specified, whose distribution can be estimated from the description you have given.
I cannot understand why you would say that a mathematician or computer scientist would say “no fair”? Have you not heard of nondeterministic algorithms?
In fact, there are quite legitimate uses for the generation of a random string in a finite alphabet. One of them is testing the tests for random number generators. See my paper on this subject.
“agony mouse” (to steal a moniker),
Your conclusion and manner of reaching it exemplifies the worst in your ability to handle the thing you seem to prize most – logic.
…as if mathematicians are any less subjective than lawyers.
Frith, please do hold your breath until you actually understand what patent law is and what it requires. Refrain from making up your own “standards”. Look into a little concept called PHOSITA before you proclaim (incorrectly) that actual source code is required, and why you are so blatently wrong when it comes to the “standard” for hardware, as no design blurprint for hardware is actually required.
For all you anti-patent tech types, do tax lawyers habitually visit the technical sites and crap all over them with discussions of topics they don’t understand? I say tax rather than patent, because patent lawyers at least have a technical base – to steal from another poster, patent attorneys have a foot in both worlds.
The EPO enjoys a reputation as a killer of software patents and a proliferator of software patents. Simultaneously. Depending which interested circles one frequents.
I am asked to cite an EPO Decision that reveals something of the way the EPO Technical Boards of Appeal go about their work, in the field of software inventions, on 101 and 103 issues.
I cite T_1143/06, happily dated April 01, 2009.
You can download it as a pdf from http://www.epo.org You can also read the entire file of the case in http://www.epoline.org by inputting the application number 00900249 into the search field (no password or registration needed).
It is a Decision of Board 3.5.01. Note how it distances itself from a precedent (T_0049/04) from another TBA, 3.4.03, and speaks of the evolution of the meaning of “technical” over 20 years of TBA jurisprudence. Note the ability of the Board in the English language (none of the Board members are English native speakers).
Judge for yourselves whether (as at least one commentator above alleges) the Technical Boards of Appeal at the EPO are “ridiculous” in the way that they interpret the statutory provisions of the European Patent Convention, as they relate to software patents.
Just to repeat, the Decision of the TBA to refuse the patent application by British Telecom is FINAL. There is no possibility to appeal that refusal to any court of law. The EPO is master in its own house.
@Got a Patent
“The vast, VAST majority of software patents issued are bleedingly obvious to any of us who work in the field. Most of them deserve to be ridden out of town on a rail to derisive laughter for what they “invent”.”
Then they should be rejected under 103 and not 101.
@PoIR,
“Mathematicians always produce a useful output when they work. They write symbols on a piece of paper. This is useful and this is output. It can be a computation that evaluates the value a formula.”
This is a very silly argument and you appear to be smarter than that. In the absence of a greater context the result is meaningless. Mathematics is merely a language developed by scientists to describe physical reality. Your position is like saying random keyboard tappings constitute a sentence.
Math neither pushes a button nor diagnoses diabetes, so please tell me why the two software examples I listed should not be patentable.
NameChange,
All of us disagree — this is obvious as night and day.
I am not talking about this blog (which is dominated by lawyers and hence subjective; if you want oposition, try slashdot — the majority there have completely the opposite consensus).
The disagreement is at the highest level you can get: even judges at supreme court show the rainbow of opinions, see In Re Bilski.
So far, this confirms my thesis that most of the disputes (about software patents at least) will not have any logical or coherent conclusion, and the only finite and hence feasible end of such legal process is financial extortion and exhaustion of smaller companies by large corporations.
“So are software companies more like publishers of books or more like manufacturers of machines?”
That’s precisely the problem with software patents, software companies want for their products both the protections that publishers of books (copyright) and manufacturers of machines (patents) enjoy. If Mr. Goetz wants to be taken seriously, he should put his money where his mouth is and argue for software to be patentable and not copyrightable. Otherwise he’s just a greedy hypocrite.
An Explanation of Computation Theory for Lawyers
By PolR
*****************
see
link to groklaw.net
*****************
OK
So let me understand this “Effective Method” algorithm of yours:
“3. M can (in practice or in principle) be carried out by a human being unaided by any machinery save paper and pencil”
OK. I insist that M be carried out by a mortal human who is guaranteed to die in finite time and that he is provided with as many pencils and papers as is necessary. Fair enough right?
“2. M will, if carried out without error, always produce the desired result in a finite number of steps.”
Fair enough. I desire the result to be whatever it is when my mortal human dies. It will be a finite number of steps for sure because my mortal human will not live forever.
“1. M is set out in terms of a finite number of exact instructions (each instruction being expressed by means of a finite number of symbols)”
This not really a limitation because I can choose my finite number of symbols to be 10^100 right? But I’ll be kind to you guys and not go to extremes.
“4. M demands no insight or ingenuity on the part of the human being carrying it out.”
Trust me, my mortal human servant is no genius.
OK let us begin.
You said we can use paper and pencil, correct?
Instruction #1 to my menial man servant:
Label the first page as “page 1” and enter the following: “Lookup (LUT) Table 00”
if input =00 then action= same as spacebar, but on paper of course,
if input= 01 then action= letter a
if ….
if input= 42 (in base 6) then action= letter z
if input= 43 (in base 6) then action= same as toggling caps lock key;
…
if input= 55 (in base 6) then action= switchover to next LUT in circular list of LUTs and get next input
(Here we can insert as many LUTs as we want into the circle each with entries 00 to 55 in base 6)
Set LUT 00 as curent LUT
Instruction #2 to my menial man servant:
Using your teeth or finger nails (or blood) mark the six sides of your No. 2 pencil with indicia representing the symbols 0 to 5.
Instruction #3 Do this twice: Hold pencil upright with point down and spin it like a top. Record the indicia that lands on top. Use the two digits obtained as input into current LUT and perform indicated action on current output page.
Of course your reaction is probably Hey no fair! Lawyer making up his own rules, twisting my words and their meanings.
Well ditto for you when you make up your own rules about what patent law says and means. Did the pencil make its point yet? Probably not.
The hardware == software point, even if taken at face value, does not indicate that software should be patentable in the way that is being done now.
Consider what a patent is awarded to: an implementation of an idea. The idea is categorically not patentable. In the exact same way that I can patent a hardware-based implementation of voice recognition, but I can’t patent Hidden Markov chains.
So if this argument is the keystone, then yes, software is patentable. That is, in order to gain the patent you must provide the source code used to create the program to the patent office for public display. No one may use this source code, but anyone may look at it and write their own with improvements (using the same algorithms if so desired), submit their own source code and have that patented.
This is the standard required for a hardware design – the design must be submitted for public viewing. And anyone is permitted to make their own copy for personal or experimental reasons. (Consider how much easier it would be for the WINE project under these conditions!)
Of course, I doubt software companies would submit to this kind of patenting, but this is the scenario that would keep the purpose of patent law alive – that of creative endeavour using the publicly available designs of current state of the art.
I’ll not hold my breath for this utopia to ever exist though.