Slashdot Mirror
New USPTO Test Could Limit Software-Based Patents
bizwriter writes "The high tech industry has been waiting for a Supreme Court decision in the Bilski case to decide fundamental questions, like when you can patent software. But there's a new test from the Board of Patent Appeals and Interferences (PDF) that just became precedential, meaning that it offers new grounds on which the US Patent and Trademark Office can deny patents on machines that use mathematical algorithms."
I've been waiting for something like this.
The largest prime factor of my UID is 263267.
Well since all software is math, this would effectively ban all software patents.
"linux is just DOS with a UNIX like syntax" -- Galactic Dominator (944134)
Sure, the patent office has it rough, sorting a gajillion [technical term] applicants all wanting a patent for their "unique idea", with the majority really just wanting a foothold for litigation riches. Don't get me wrong, there are definitely honest attempts at securing one's interests and not getting your own pride and joy "unique idea" stolen, but I'll be damned if the USTPO hasn't awarded some of the dumbest, most wide-ranging generalizations to companies that patent spam thousands upon thousands a year. The latter mentioned issue has been going on long before the dawn of the digital, so I feel its too late to correct the problem in the current system.
'We are trying to prove ourselves wrong as quickly as possible, because only in that way can we find progress.' RPF
Probably ANY algorithm can be reworked to be a mathematical formula. (I wouldn't want to have to maintain the thing, but it could run.)
Table-ized A.I.
As eloquently illustrated here: http://xkcd.com/435/, everything is math. So, the question, as always is where does one draw the line. There is no reason to throw the baby out with the bathwater.
That's always been the law. Look up basic patent law. Those criteria have been fundamental since the patent system's inception. I suppose it takes a judge that's actually read the law to be able to rephrase it so that other judges can be reminded of what the law has stated for centuries. Perhaps now more patent judges will make rulings actually based on patent law. Glad to hear of the "development", but it's always been the case.
For those curious, you can find the patent here. Looks like a generic recommendation engine.
My question is, if this was decided in August, why is it only precedent now?
It that normal? Was it time for an appeal?
Comment forecast: Bits of genius surrounded by a sea of mediocrity.
As eloquently illustrated here: http://xkcd.com/435/, everything is math. So, the question, as always is where does one draw the line. There is no reason to throw the baby out with the bathwater.
How about this: you may not patent a work that is (or could be) protected by copyright. You can't copyright a physical invention like a new machine, drug, or industrial process. You can copyright manuals and other documents that describe those things, but those documents in and of themselves are not the subject of patents.
It is a miracle that curiosity survives formal education. - Einstein
That makes so much sense... Now if only it was true!
These comments are my own and do not necessarily reflect the views or opinions of my employer or colleagues...
Unfortunately, if this logic actually held, then software would never have been patentable. An algorithm, if you explain it in a certain way, can sound a whole lot like a machine.
You can't copyright a physical invention like a new machine, drug, or industrial process. You can copyright manuals and other documents that describe those things, but those documents in and of themselves are not the subject of patents.
But you can both copyright and patent a machine's visual design, which itself can be described in a document, but the copyright and patent would be on the visual design itself, not the document. So there is precedent for software being both copyrightable and patentable at the same time.
To really complicate things, visual designs can even be trademarked in some countries (including the U.S.)
My blog
A software program is simply an instruction manual created to be read by a computer. I've said it before, programs/algorithms are copyrightable, but should not be patentable. Period.
Every time I start to have faith in humanity, I ruin it by driving to work between 7 and 8 am.
You can't copyright a physical invention like a new machine, drug, or industrial process. You can copyright manuals and other documents that describe those things, but those documents in and of themselves are not the subject of patents.
But you can both copyright and patent a machine's visual design, which itself can be described in a document, but the copyright and patent would be on the visual design itself, not the document. So there is precedent for software being both copyrightable and patentable at the same time.
To really complicate things, visual designs can even be trademarked in some countries (including the U.S.)
What you describe is how the system currently is. What I offered was a proposal for how the system might be improved, so naturally it won't reflect the current reality. Still, I like the idea of choosing one. Anytime a work might have multiple forms of intellectual property protection (copyright, patent, trademark, etc), the owner may choose only one. So you may copyright your software but may not also patent it. Or you can patent your software, but then we're free to pirate it :-).
It is a miracle that curiosity survives formal education. - Einstein
Definitely a move in the right direction to address the now prophetic "untold consequences" foreseen by Judge Archer and Judge Nies in their dissenting opinion in In Re Alappat, No. 92-1381 (Fed. Cir. July 29, 1994).
Unfortunately, as with the majority decision in the 1994 Tektronix appeals case, the tests provided to determine patent-ability of software algorithms continues to leave the door wide open to incessant lawyering not for the purpose of upholding the constitution and promoting "the Progress of Science and useful Arts".
No, instead we will continue to waste investment resources to stifle competition in the name of profit margins and monopolies.
Most people likely will not read the dissenting opinion so I'll quote the conclusion from the dissenting opinion here with emphasis added so others can see the prophecy for themselves:
Please don't mod me a troll for asking an honest question. IANaIPL. If i had the answer i wouldn't be asking.
Why allow software to be patented instead of copyrighted?
- Patents should be for THINGS (concrete stuff).
- Copyright should cover TEXT (abstract stuff).
It make sense to me that you should be able to protect the way YOU managed to execute some process, but not the idea of being able to DO the process at all. It seems to me that software companies are trying to patent vehicle direction input devices (as a concept) rather than just 'our particular design for a steering wheel'(a specific implementation).
Seems to me that these systems should exist to prevent others from stealing your specific work, the result of *your* labor; not to prevent others from engaging in the same line of business. Which seems to be the point of much of software patenting.
And why allow it to be patented instead of having it copyrighted?
Am i missing something?
Tangent/Rant:
If the intent of copyright and patenting is to encourage innovation, i think it has failed. i have tons of doodles and outlines for things i'd like to see on the market or share. But when i look at what would be involved... i'd rather do anything else. Wash dishes, scoop the litter box or watch TV... than hire a lawyer and go through the years of waiting, piles of paper work and enormous expense of dealing with the nightmare of IP. Even if i do it all and do it all correctly it could still be taken from me by a better paid lawyer.
Utilizing the synergization of benchmark e-solutions to pre-workaround action items!
Is the claim limited so as to not encompass substantially all practical applications of the mathematical algorithm5 either “in all fields” of use of the algorithm or even in “only one field?”6
if (claim) !(( practical algorithm application) (( fields) || (fields x 1))) then no patent
Neither are easy to read - but I think the idea is you cant patent a mathematical formula by itself
As eloquently illustrated here: http://xkcd.com/435/, everything is math.
No. We can describe/approximate almost all sciences using math, just like we can describe them using English or any other language. That does not mean that everything is math or linguistics.
So, the question, as always is where does one draw the line. There is no reason to throw the baby out with the bathwater.
That baby already drowned ages ago.
Donate free food here
But what is the point of making that arbitrary distinction? Simply imposing arbitrary rules with no purpose other than to make the system appear simpler, is meaningless. Most everyone here is a technical person in some way: networks, software, hardware... The same principles apply here as in any complicated system. Usually you have to have a complicated system in order for it to appear simple to the user. Many here seem to want to believe problem with IP law is merely that it is too complicated, and thus offer simplistic solutions to, well, simplify the system. The real problem is that IP law is like a legacy mainframe system that has had decades of hacks piled on top of it and then had a web based interface slapped onto it as well. It needs to be ripped out and something else designed from scratch to replace it. Not in that order, of course.
The final solution may not be so very simple as many here often propose. But it should be coherent, and that is the key.
Perhaps true. But we still need to give it an official burial and make a new baby.
Our scientific fields all use mathematics as a language to describe the phenomena they investigate. That does not mean those phenomena ARE math, or are composed of numbers. Like logic, math is a descriptive and deductive tool. I believe you are confusing the map with the territory on this one.
To put it a more facetious way, I can use English to describe how to build a house. That does not mean the home is made of my words, for they are not nearly so physically sturdy as wood, metal, and concrete.
It is a miracle that curiosity survives formal education. - Einstein
If slashdot was still giving out mod points to people who don't use javascript, I would have modded you up.
When information is power, privacy is freedom.
As RMS once pointed out, machines are more and more being implemented in software instead of hardware these days. Where gears and spring-triggers may have been used before, now sensors, cameras, and servos are used more and more often instead. Thus, ending software patents may end up ending patents all-together.
Some may feel that's a good thing. But, many feel that research should be rewarded beyond just being first-to-market. The problem is separating "good ideas" from patent trolling. How can the legal system distinguish? I doubt there are any set of clear rules for such. The best we can have may be some form of "techie jury" that ranks the innovation level.
The mechanical world was in general easier to manage with patents because equivalency was difficult to obtain without adding extra parts or movement. However, software is much more malleable such that there are many different ways to achieve something sufficiently similar without critical loss of efficiency. But this also encourages the use of overly-broad patents that can be interpreted to cover just about anything roughly similar. Equivalency allows the patent defender to interpret the alleged violator's process as being the same as the patent.
When everything is everything, nothing is nothing, and the most cleverest-at-tongue lawyer wins in the end. This is why big companies tend to favor wide patents: they have the best lawyers.
Table-ized A.I.
Horse crap. That comic says "purity" which is not the same thing anyway.
Math is the language of expression for the results of research in science/engineering (compute the curve of the graph) and is used to predict future results. However, patents don't cover the math itself in those fields, they cover a specific application of it, probably with some actual creativity thrown in (shapes that are more practical like the hexagon design of a nut instead of being circular; also, realising that you could apply the formula to produce the useful result in the first place). Software on the other hand is the math, there is usually no clear separation between the implementation and the derivation that lead to it.
And, again, software is already protected by copyright and is the only field to "enjoy" double protections. [You can't patent a story idea (eg. romance-comedy) and you can't copyright a vacuum cleaner design, why does software deserve both?]
Anyway, this is a sweet move on the USPTO's part, it's nice to know that they aren't completely asleep at the wheel. These new criteria seem to be specifically for the purpose of forcing applicants to narrow the scope of their patent applications, hopefully we won't see more vague crap like 'a mechanism for using a data-entry interface to enter uniquely identifying credentials to access a system' (passwords, also would cover smartcards and just about anything else) [I made this example up but it isn't far off].
Well since all software is math, this would effectively ban all software patents.
Correspondence is not equivalence. If you digitize the Mona Lisa you could say it's a binary string, or a number, or an image, or the definition of a Turing machine. The context is what defines the significance. Saying that every grain of sand on a beach has a corresponding number doesn't mean that a sandy beach is really just a bunch of numbers, and saying that every computer program has a corresponding mathematical algorithm doesn't mean that computer programming is just mathematics.
None of that means the USPO won't take the chance to stop patenting software, but that'll be a political decision, not something based on an abstract correspondence.
tomorrow who's gonna fuss
The entire concept of copyright, patent, etc. is completely arbitrary. It's not the natural product of the rules of physics. It is something human beings made up for their own purposes. Therefore, the important question is not whether it's arbitrary, but whether it continues to serve those purposes. Making it as simple and robust as possible is a step in the right direction.
Why do you suppose a legacy mainframe with hack after hack, patch after patch applied to it would be a problem? That's easy. Because the constant hacking and patching introduces additional complexity. The complexity arises from the need to add new functionality that the original creator did not anticipate while at the same time not breaking the existing code. By contrast, redesigning something from scratch and accounting for everything you learned from all the hacking and patching gives you an opportunity to design it as simply as possible.
I agree that the IP system needs to be replaced. We can do that all at once, but this is not realistic politically. Realistically, the best we can do is to take steps in that direction. Throwing out some of the existing complexity and replacing it with simplicity is a step in that direction.
It is a miracle that curiosity survives formal education. - Einstein
No, not everything is math. Entropy is something math cannot yet account for. Until mathematics can model physical systems perfectly, there'll be a huge gap between mathematics and reality. And that's what this comic strip is actually trying to say.
"If a nation expects to be ignorant and free in a state of civilization, it expects what never was and never will be."
Writing code is creating a 'thing', and falls under patent law.
Copying that program and selling it falls under copyright law.
of course, software falling under patent law is pure bullshit, not only for all the reasons that dozens of posts on this thread will mention, but also because nobody is actually disclosing usable source code in their patents. A patent is supposed to explain an exact method for doing something, not just vague hand waving and hints about how something is done.
I would have far less of a problem with software patents if they actually disclosed fully functional source code with the patent. The way things work now, they get the benefits of a patent, and the secrecy of having a trade secret.
HA! I just wasted some of your bandwidth with a frivolous sig!
This patent used the common tricks of "An ordinary computer implementing 'insert unpatentable material here'" and "A computer-readable medium containing code implementing 'insert unpatentable material here'", and while the board rejected both, it did so with very nitpicky analysis (particularly in the second case).
What the board should have done is categorically rejected them, particularly the second one. A claim like "A computer-readable medium containing X" should be no more valid than a claim of "A human-readable medium containing X". And the latter case would make the patent applications themselves patentable, which is obviously absurd.
In my observation of software patents, those that are really mathematical in nature like a new compressing algorithm, 3-D rendering algorithm, etc are actually the ones that are most genuine, non-obvious, creative. Come on, if those are trivial, can you come up with one that beats the existing ones now? On the other hand, the low quality / troll patents, like those involving UI or famous EOLA browser plugin patent, are usually non-mathematical in the the usual sense. You could argue it is still math, but even a mathematician would not look at it that way. Are we doing it the opposite way?
But you can both copyright and patent a machine's visual design
Don't confuse a design patent with a utility patent, they are really not the same thing.
While I do agree with you on a technical level, the person to whom I was replying had said that "all software is math." I was simply showing the logical extension of that statement and way of thinking. My point, in the end, is that regardless of whether software is math or not, a more considered line needs to be drawn as to whether something should be patentable or not. Although I agree that most software patents are total BS, I do not believe that NO software should be patentable. There are some software approaches that are truly unique and not obvious which someone worked very hard to figure out and perfect. I believe those should be protected (if the inventor wants to protect them). Unfortunately, the USPTO has been giving out patents for the idea of having a certain feature rather than a particular implimentation of that feature.
Ok, I feel really stupid reading that thing.
It's in a form of legalese which is something that always makes my mind retreat to the dark recesses of my brain to hide in fear.
Can someone please provide a normal-speak (or even geek-speak) version of that mess, maybe with an example or two?
I'd certainly appreciate that, and I'm sure a lot of other readers would as well.
Of course this is a bit of a fallacy of equivocation on the word "arbitrary." In general, and in the way I used it, "arbitrary" means "with no reasonable consideration, given the context." You are redefining it to mean "not adhering to immutable laws of physics." There is quite a difference there. Then you are using that "straw man" to argue against.
I am sure we can all agree that the definition of "Arbitrary" itself can, in a way, be arbitrary. If it is applied in a reasonable way within the given context then "arbitrary" would not be arbitrary.
That said, I agree with the rest of what you said. In an ideal world we could do as I suggest: Bury our dead baby and make a new one. Just as we would not simply rip out a mainframe, we could not simply scrap all of current IP law. However, simplistic solutions will definitely NOT solve the problem and I grow weary of hearing them.
The natural world may only be approximated using a mathematical framework (at least at present), but the devices that we create based on that knowledge is a direct application of the mathematical models. And it is those devices that are under question when one applies for a patent, not science or nature as a whole.
So say you have an algorithm, and claim that it is nothing more than math. You can implement that algorithm in a programming language, convert it to machine language for a general purpose processor, implement it as a giant logic table in an FPGA, implement it using TTL logic circuits, or make a custom ASIC of the same logic. Many algorithms, such as signal processing, can be implemented to within a desired precision with either digital or analog filters. There are also mechanical implementations of the same ideas - mechanical calculators are obvious, but even simple things like automatic valves may be more robust in for certain applications than converting to the electronic/digital realm and back again. At what point do you draw the line and say that it is no longer just a mathematical algorithm?
You don't seem to understand the Curry-Howard correspondence.
You don't seem to understand that I was only pointing that forbidding patents on math does not automatically remove all science fields mentioned in the XKCD comic from patentability, in order to defang the argument that opposing software patents because they are math patents is equivalent to opposing patents on anything, because all sciences are supposedly math.
I did not say or even intend to imply that software isn't math. Just google for my name (from my email address) in combination with "software patents" and you'll see that I would be one of the last people in the world to advocate software patentability.
Well, that or you intended to reply to the GP instead of to me.
Donate free food here
Is there a baby in the bathwater?
Unfortunately, the patent office is so inept in the examination of software that it thinks the algorithmic equivalent of a simple lever is a dazzlingly brilliant new invention. Further, it thinks that simple combinations equivalent to using a rope and pulley to actuate a lever is an ingenious invention rather than just a blindingly obvious combination of two ancient ideas.
While I do agree with you on a technical level, the person to whom I was replying had said that "all software is math." I was simply showing the logical extension of that statement and way of thinking. My point, in the end, is that regardless of whether software is math or not, a more considered line needs to be drawn as to whether something should be patentable or not. Although I agree that most software patents are total BS, I do not believe that NO software should be patentable. There are some software approaches that are truly unique and not obvious which someone worked very hard to figure out and perfect. I believe those should be protected (if the inventor wants to protect them). Unfortunately, the USPTO has been giving out patents for the idea of having a certain feature rather than a particular implimentation of that feature.
"Everything is math" as stated in the post to which I replied != "all software is math", for the reason that "software" != "everything". My response was on that basis alone.
About whether software should be patentable, we need not speculate too much. There are European countries where software cannot be patented. The relevant question is whether this has caused their software and IT industries to collapse, and it has not. So it has already been demonstrated that software patents are not an essential component of a working economy. The burden of proof is therefore on anyone who suggests that the USA is a special case, that it cannot live without them even though other countries can.
That's good enough for me, although those with vested interests (and those who believe the marketing of same) won't like the simplicity of this observation.
It is a miracle that curiosity survives formal education. - Einstein
Next time quote with context please. The quote with context looks like this:
I said that the question is whether intellectual property concepts serve the purposes for which they were created. That is a reasonable consideration.
That might apply to overly simplistic solutions. What I advocate is that it be as simple as possible but no simpler. Any degree of complexity that is truly necessary is acceptable from this point of view. That's why I said "Making it as simple and robust as possible is a step in the right direction." It's also why I did not say "making it too simple to where it cannot serve its purpose is a great idea." Is this non-controversial only to me?
It is a miracle that curiosity survives formal education. - Einstein
There are European countries where software cannot be patented. The relevant question is whether this has caused their software and IT industries to collapse, and it has not. So it has already been demonstrated that software patents are not an essential component of a working economy. The burden of proof is therefore on anyone who suggests that the USA is a special case, that it cannot live without them even though other countries can.
Very good point. The first step, in designing any new system, is asking, "Do we need this system at all?"
But it's not arbitrary. Patents by design were suppose to be for tangible things, with Copyrights for intangible things; never the two should intertwine. That actually works very very well; until you get the lawyers and PHB's involved that want to protect everything in every possible way - so they intertwined them and now we have the chaotic mess that we have. We need to separate them again.
Truth is like the sun. You can shut it out for a time, but it ain't goin' away. - Elvis Presley (source: imdb.com)
There are European countries where software cannot be patented. The relevant question is whether this has caused their software and IT industries to collapse, and it has not. So it has already been demonstrated that software patents are not an essential component of a working economy. The burden of proof is therefore on anyone who suggests that the USA is a special case, that it cannot live without them even though other countries can.
Very good point. The first step, in designing any new system, is asking, "Do we need this system at all?"
I think acknowledging the truth of this means that you and I are never going to have a career in politics :-).
It is a miracle that curiosity survives formal education. - Einstein
It would be worth it to take down business process patents as well, for they are merely algorithms.
I'm pretty sure most machines could also be so represented, though.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
That's the most screwed up logic I've heard today... software is a machine because software + computer == machine? You realize that's like saying water is a plant cause water + seed == plant.
Every time I start to have faith in humanity, I ruin it by driving to work between 7 and 8 am.
Here's one for Microsoft & IBM. If you don't actually build the idea, the design, the original design idea, then it doesn't even exist. Just because the idea exists doesn't make it yours or real. If you haven't made it, you can't patent it.