8 ms·

I really don't think the problem is software patents, it seems to me the real problem is that most software patents seem to be ridiculously trivial. Here's an

I really don't think the problem is software patents, it seems to me the real problem is that most software patents seem to be ridiculously trivial. Here's an example: my employer, a rather small company has invested an enormous amount of resources in developing a specific algorithm. It took years of research and development, lots of money and brainpower, it's very far from being trivial (in fact it's so complicated I have to admit I don't understand some parts of it) and it's light years ahead (in terms of usefulness and speed) of any other algorithm with similar applications. Obviously the last thing we want now is simply handing it over to our competitors, after all this work. If the patent system didn't guarantee us monopoly on using the fruits of our own R&D, this algorithm probably would not have been developed for a very long time, if ever, by anyone. It would have been innovation that never happened. My point is that the threshold of what can patentable should by way higher, but I think we need a (software) patent system none the less.

ssp 15y ago
Have you filed patents on it, then? I'd be interested in seeing the applications.
reitzensteinm 15y ago
It would be a pretty messed up system if there were never any cases where patents were justified. Yours sounds like an edge case. The real question is what's best for the industry as a whole?
- scott_s 15y ago
  Rather than being an edge case, I think that what DrJokepu described is the intended case. But the system has been gamed.
  - reitzensteinm 15y ago
    You're right, that was a poor choice of words.
wslh 15y ago
I understand your viewpoint. My point is: what happens in the long term if people like Dijkstra patented their algorithms? I think we all have a stockpile of licenses to pay for almost any code, even a r"Hello World" (you are using a string, if it's unicode you need another patent!). So, I think there are more winners and innovation "sacrificing" the software inventors side.
- cop359 15y ago
  according to wiki, Dijkstra's algorithm was published in 1959. If he had a patent on it for 15 years and made enough money to cover his troubles, then I don't really see it as a bad thing. Now a days, you might say that 15 years is too long and maybe that's true, but sometimes it takes people a very long time to monetize their ideas (due to technical problems, bad business partners, time needed to develop connections in the business etc.)
  - wslh 15y ago
    But my argument is: if not only Dijkstra but a huge part of the comp sci community would have decided to patent their inventions it will be impossible to move forward in the field. Only the legal costs of "swimming" on this system will be prohibitive for the developers community.
    - sp332 15y ago
      More likely, institutions would license the patents and the field would move still move forward, just with more friction.
      - wslh 15y ago
        I don't see it that way. The comp sci field moved forward because of the democratization of the personal computer where you have the freedom to develop what you want and distribute it yourself. For example one of the mini revolutions was Turbo Pascal from Borland offering the compiler for few bucks when compilers were more expensive, with patents all prices go high and that means few people or companies can spend money on that.
dstein 15y ago
Is there any reason the algorithm couldn't remain a trade secret? And is there any possible scenario where the long term good that comes from this patent will outweigh the long term good that will come from untethered innovation?
- stcredzero 15y ago
  Is there any reason the algorithm couldn't remain a trade secret? Reverse engineering. And is there any possible scenario where the long term good that comes from this patent will outweigh the long term good that will come from untethered innovation? Perhaps. In those situations where there would otherwise be no commercial development at all without the prospect of the monopoly a patent affords. In software, I would posit that there's more good that comes from the innovation.
fiblye 15y ago
I think the issue that most people have is that people who coincidentally develop an algorithm that serves the same functionality or operates in a similar way as your algorithm will be sued to hell, even though they'd never even heard of the work you'd done. A bigger issue is with people that patent every single idea they've ever had, never fully develop anything, then sue any person that's made a profit with a similar idea. You have the right to keep your code confidential and protect it from theft, but people who are capable of making something similar with their own abilities shouldn't be punished.
rbanffy 15y ago
Another solution would be to create a very inexpensive way to defend against a patent lawsuit. If defending yourself is trivial, the incentive to agree to extortion schemes is dramatically reduced. A better one could be to make it profitable to defend against frivolous lawsuits like these. If the plaintiff is required to pay the legal bills of the defendant, plus a hefty margin awarded by the judge according to how frivolous the lawsuit was, we would turn the incentive backwards. That would prevent ridiculous patents like these to even reach a court and create a market for lawyers specialized in shooting down stupid lawsuits.
- kluikens 15y ago
  I'm having a hard time imagining how there could be a general way to introduce an inexpensive defense process. Do you suggest a public defender system? Would it be subsidized by additional patent application fees? Additionally, keep in mind that road goes both ways. Suppose you devise something that is arguably worth a patent and suppose you later find one or many people infringing on your patent, far over the line of fair use. If legal defense is now cheap (dollars/time-wise), you may be stalled in the courts trying to legitimately defend your patent. There is the approach of increasing punitive fines for unworthy suits or false claims. But patent law is messy. How do you prove the other party intentionally made false claims, wasting the courts time? Or what is an unworthy suit? Somewhat related to punitive enforcement measures, in one of our states a government official is automatically jailed for 1 or 2 nights if they close off an open-meeting no matter the reason. I personally think this is an appropriate punitive measure, but others may think it harsh. I don't think either approach would work -- both would likely make the system worse. What we need are both a court system and patent office that are sufficiently staffed and have a minimum competence. (I don't mean to insult -- I say this in the context where an official is utterly out of their realm of expertise).
fauigerzigerk 15y ago
What makes you think your company deserves to be protected against others independently developing a similar algorithm?
- atakan_gurkan 15y ago
  That situation is an unfortunate side effect. The point of patents is that the inventor requires an incentive (a limited term monopoly in case of patents) to make their methods publicly known. The protection is against the copiers, but since an independent developer is indistinguishable from a copier, they get excluded from using the idea as well. I always thought that this is precisely why the patentable ideas are non-trivial.
  - edanm 15y ago
    You're right. But even non-trivial ideas tend to occur to several people at once, which is a part of the problem.
- bostonpete 15y ago
  This argument can be made against patents in general. If people want to argue that all patents are bad, that's fine, but most people draw a tenuous distinction between software patents and other kinds of patents. I agree with DrJokepu -- the real problem (assuming that you're not opposed to all patents) is that patents are being given for ideas that are moderately to extremely obvious.
  - prodigal_erik 15y ago
    It can only be a net win for society if the total price everyone pays to license a patent is less than the total cost of making each user reinvent it for themselves for the next twenty years. That's why obvious software patents are so bad—the basics of the young field weren't already claimed decades ago, the cost of "reinvention" is so low (one-click requires what, a napkin, a crayon, and half a beer?), and the licensing prices devolve to pure rent-seeking based on the necessity of the obvious idea. What we need is some counter-balancing force that ensures patent license prices remain reasonable. Normally that's done by the threat of substitution or inventing-around, which doesn't work when USPTO screws up and allows claiming the whole problem space. The price especially needs to be reasonable for non-profit use, which is a software-specific issue made possible because our unit costs are also basically zero. It would also help if they tested patent claims for obviousness by giving someone the problem statement and seeing if they produce a solution off the cuff covered by the same claims.
shad0wfax 15y ago
What if your competitor came up with the same algorithm (or very similar) and performs as well both in time and space complexity... and all of this on their own, without actually taking a peek at your code. In that case would you go after them and sue them?
- DrJokepu 15y ago
  I imagine we would, simply because we can; we're a commercial entity after all. Personally (and I'm not representing my company here) I don't think that's a good thing; if they could prove that they independently reached the same results without using any of our research or work (or we couldn't prove the contrary), they should be protected. That's not the case though at the moment unfortunately.
  - shad0wfax 15y ago
    I agree, its not the same case. The problem might be in 15 years from now, your algorithm might just become a very natural way to solve that specific problem and at that point how your company plays this out matters. My guess is the companies that are in the business of making, creating, producing and then selling software themselves, will be ok with it. These trolls like Lodsys try to neither produce nor create. A very important rule for the any type of patent should enforce "active" product in use/development. Sadly, it might never happen.
  - shad0wfax 15y ago
    I agree, its not the same case. The problem might be in 15 years from now, your algorithm might just become a very natural way to solve that specific problem and at that point how your company plays this out matters. My guess is the companies that are in the business of making, creating, producing and then selling software themselves, will be ok with it. These trolls like Lodsys try to neither produce nor create. A very important rule for the any type of patent should enforce "active" product in use/development. Sadly, it might never happen.
guelo 15y ago
Nope not good enough. Software patents are doing much more harm than the potential gain from protecting your company.
CodeMage 15y ago
Threshold is one problem. Duration is another. Here's a little story to illustrate why patent trolling is so evil. I used to work at a software company that has an incentive program for developers that come up with patentable ideas, e.g. algorithms. You get a token prize when "your" patent gets submitted -- a pencil or something like that, I don't remember anymore. Then, if "your" patent gets accepted by USPTO, you get a more substantial prize, along the lines of a week at some comfy resort. A lot of my co-workers asked the same question: "Yeah, okay, that's nice, but there's a good chance that the company is going to make shitloads of money off a successful patent. A week at some resort doesn't really compare to that. Why don't you guys offer something that does?" The guy who was explaining the incentive program responded: "Because you patent an idea and that's worth exactly nothing until you implement that idea in a product and sell that product to customers. You had the idea, but it's the company that has the means to implement it and market it and sell it." When you think about it, it's reasonable. Not precisely fair, but then again, things rarely are. The point of this story is that software patents, just like all patents, are there to give the inventor a chance to implement the invention and make money from it, as a reward for the inventor's genius and hard work that went into coming up with an invention. That's obviously a good thing, because it gives you an incentive to invent something without having to fear that someone will simply steal your invention and make money off it. Like every regulation with good intentions, it gets abused. What you said about trivial patents is just one way it gets abused. But another thing that makes it easy to abuse software patents is the duration of the patent. Twenty years is a hell of a long time in software. It practically invites patent trolling.
- Joeri 15y ago
  I've never heard a case where the do-or-do-not decision regarding a software project hinged on patents. That they are somehow needed to encourage innovation is a myth. Besides, they're unfair. A typical software product might have perhaps 1% patented matter, and 99% copied from the "standard requirements" of the market segment it's in. The competitors spent the money developing the market to figure out that 99%, so if you use a quid pro quo measure then they have an automatic right to the 1% innovation the new player brings to the table.
  - 15y ago
citadrianne 15y ago
I think it's hard to predict what abolishing patents would do to innovation. For example, there are no patents in fashion, which actually ends up putting pressure on creators to be superinnovative to stay ahead of the copycats. Ref: http://www.ted.com/talks/view/lang///id/866 http://www.ted.com/talks/view/lang///id/866
- mindcrime 15y ago
  My personal opinion - for what that's worth - is that this kind of scenario comes pretty close to representing what would happen if there were no patents in software (or any other field for that matter). Innovation would still happen, and would probably happen even faster than it does now, because of the "arms race" effect. Simply put, a firm would be forced to innovate constantly, quickly and fiercely, to stay in front of its competitors. Combined with the fact that I am appalled by the idea of use of government force to block someone from using an invention - that they may have arrived at independently - just because someone else patented it first; and my belief that there are too many smart people in the world for any really useful idea to stay undiscovered long, I advocate for simply abolishing the entire patent system.
  - biafra 15y ago
    In Europe there are no software patents. Do they develop software and innovate there?
Vivtek 15y ago
While this has no bearing on the law, I've always thought that works of software that represent an actual effort of development comparable to physical invention should enjoy the benefit of patents. This is precisely what the patent system was designed to do - protect large research investments. The abuse of the system for things that really don't require research and development is at issue here. I think proposals to abolish software patents entirely are probably wrong - but granting patents on XOR, or 1-click, or doing something obvious but on a computer system, are equally wrong. Put it this way: if I have a problem to be solved by software that is really hard (such as your algorithm), then I'm going to research it before spending five man-years developing it, find you guys, and happily license your solution. No question. And if somebody reverse engineers your solution and sells it, then by God they should be penalized for violating your patent. This is why we have patents. But if I have an idea like 1-click and pay an Indian developer $10 an hour for a week to develop it, this is simply not worthy of monopoly protection; your disclosing your solution isn't that valuable. If, seventeen years from now, after I've built an online shopping empire while you haven't, you sell your patent to Lodsys and they sue me for a billion dollars out of the blue, that is highway robbery, and it's not in any way helping society - and that is the situation we actually currently have.
- anamax 15y ago
  > I've always thought that works of software that represent an actual effort of development comparable to physical invention should enjoy the benefit of patents. There is no "effort" requirement for patents on physical inventions. There's also no "it was done by experts" requirement.
  - Vivtek 15y ago
    Yes. That's true. That's why I said it has no bearing on the law.
  - lurker19 15y ago
    Incorrect. http://en.m.wikipedia.org/wiki/Person_having_ordinary_skill_in_the_art http://en.m.wikipedia.org/wiki/Person_having_ordinary_skill_...
noonespecial 15y ago
Unfortunately those of you who really deserve patents will not be able to make them rise above the noise of this lunacy. We're eventually going to have to go with "no software patents at all" to fix this. As a society, we just haven't outgrown the greedy thuggery that could allow us to respect each others work without gaming the system into tragedy of the commons territory. A classic case of "see, this is why we can't have nice things".
- jpadvo 15y ago
  I don't think that is inevitable. There is not an inherent reason why the Patent Office needs to award trivial patents. If they tighten up (or apply) their standards, the problem would be tremendously reduced.
  - rapind 15y ago
    Coming up with relevant criteria to determine the triviality of a software patent application doesn't seem simple at all. This means you need to hire experts in the field. This is where it gets interesting. In most regulatory organizations where a high amount of expertise is required (like economics) you see a lot of corruption where companies hire out of and into these organizations specifically for the benefits of having representation on the inside. I.e. Henry Paulson. So unless you have a good way of preventing these employees from ever working for Microsoft / Facebook / Apple / Google etc. then it's really not going to work out as intended. IMO, just get rid of software patents. Sure it'll suck for some, but overall it's probably healthier for the industry. Plus it's nice and simple and doesn't employ a whole mess of lawyers.
  - seats 15y ago
    The USPTO is overwhelmed and understaffed, so I can understand why they have poor QC on patents, particularly when they require technical knowledge to determine triviality. What I don't get though is why it seems to be such a challenge to prove the invalidity of these crap patents in court. The google loss to bedrock (the linux kernel troll) is a high profile example of this. Maybe the standards that need to be looked at are around litigation and if enough crap is thrown out, all the incentive is gone and all of the sudden the USPTO application volume goes way down.
SoftwareMaven 15y ago
Coca-cola has managed to deal with this for a hundred years. Google has managed it for 10 years. These things are called trade secrets. You protect them through NDAs, being intelligent with who has access to the information, and not open-sourcing the code.
- aridiculous 15y ago
  In the case of Coca-cola, at least, the secret formula isn't even much of a concern anymore (I think it's been leaked a few times actually). Them protecting the formula is really just tradition and adds to their branding. What's important to them is keeping their brand power.
georgemcbay 15y ago
I'm OK with the idea of software patents in theory just like I'm OK with the idea of the death penalty for extreme crimes in theory. Because we live in an imperfect world made up of imperfect people, I don't support either in practice. Perfect software patent system > no software patent system > broke software patent system, but I believe the perfect system is impossible, so I'd rather see the software patent system dismantled completely than exist in a broken state.
- fredwilson 15y ago
  i love this analogy and i am with you on both
jellicle 15y ago
Except that you could already use and profit from the algorithm indefinitely without any patents in place whatsoever, simply by keeping the algorithm secret.
uiri 15y ago
I don't understand how you would be handing it over to your competitors: 1) if it is proprietary software, just add a clause which forbids reverse engineering. If it really is as complex as you describe however, I doubt reverse engineering it would be an option. 2) if it is open source software, I don't see the big deal. Even if they get the algorithm your software can still stay steps ahead of theirs. So what advantage would a patent even give you in this case?

ssp 15y ago

Have you filed patents on it, then? I'd be interested in seeing the applications.

reitzensteinm 15y ago

It would be a pretty messed up system if there were never any cases where patents were justified. Yours sounds like an edge case. The real question is what's best for the industry as a whole?

scott_s 15y ago
Rather than being an edge case, I think that what DrJokepu described is the intended case. But the system has been gamed.
- reitzensteinm 15y ago
  You're right, that was a poor choice of words.

wslh 15y ago

I understand your viewpoint. My point is: what happens in the long term if people like Dijkstra patented their algorithms? I think we all have a stockpile of licenses to pay for almost any code, even a r"Hello World" (you are using a string, if it's unicode you need another patent!). So, I think there are more winners and innovation "sacrificing" the software inventors side.

cop359 15y ago
according to wiki, Dijkstra's algorithm was published in 1959. If he had a patent on it for 15 years and made enough money to cover his troubles, then I don't really see it as a bad thing. Now a days, you might say that 15 years is too long and maybe that's true, but sometimes it takes people a very long time to monetize their ideas (due to technical problems, bad business partners, time needed to develop connections in the business etc.)
- wslh 15y ago
  But my argument is: if not only Dijkstra but a huge part of the comp sci community would have decided to patent their inventions it will be impossible to move forward in the field. Only the legal costs of "swimming" on this system will be prohibitive for the developers community.
  - sp332 15y ago
    More likely, institutions would license the patents and the field would move still move forward, just with more friction.
    - wslh 15y ago
      I don't see it that way. The comp sci field moved forward because of the democratization of the personal computer where you have the freedom to develop what you want and distribute it yourself. For example one of the mini revolutions was Turbo Pascal from Borland offering the compiler for few bucks when compilers were more expensive, with patents all prices go high and that means few people or companies can spend money on that.

dstein 15y ago

Is there any reason the algorithm couldn't remain a trade secret? And is there any possible scenario where the long term good that comes from this patent will outweigh the long term good that will come from untethered innovation?

stcredzero 15y ago
Is there any reason the algorithm couldn't remain a trade secret? Reverse engineering. And is there any possible scenario where the long term good that comes from this patent will outweigh the long term good that will come from untethered innovation? Perhaps. In those situations where there would otherwise be no commercial development at all without the prospect of the monopoly a patent affords. In software, I would posit that there's more good that comes from the innovation.

fiblye 15y ago

I think the issue that most people have is that people who coincidentally develop an algorithm that serves the same functionality or operates in a similar way as your algorithm will be sued to hell, even though they'd never even heard of the work you'd done. A bigger issue is with people that patent every single idea they've ever had, never fully develop anything, then sue any person that's made a profit with a similar idea. You have the right to keep your code confidential and protect it from theft, but people who are capable of making something similar with their own abilities shouldn't be punished.

rbanffy 15y ago

Another solution would be to create a very inexpensive way to defend against a patent lawsuit. If defending yourself is trivial, the incentive to agree to extortion schemes is dramatically reduced. A better one could be to make it profitable to defend against frivolous lawsuits like these. If the plaintiff is required to pay the legal bills of the defendant, plus a hefty margin awarded by the judge according to how frivolous the lawsuit was, we would turn the incentive backwards. That would prevent ridiculous patents like these to even reach a court and create a market for lawyers specialized in shooting down stupid lawsuits.

kluikens 15y ago
I'm having a hard time imagining how there could be a general way to introduce an inexpensive defense process. Do you suggest a public defender system? Would it be subsidized by additional patent application fees? Additionally, keep in mind that road goes both ways. Suppose you devise something that is arguably worth a patent and suppose you later find one or many people infringing on your patent, far over the line of fair use. If legal defense is now cheap (dollars/time-wise), you may be stalled in the courts trying to legitimately defend your patent. There is the approach of increasing punitive fines for unworthy suits or false claims. But patent law is messy. How do you prove the other party intentionally made false claims, wasting the courts time? Or what is an unworthy suit? Somewhat related to punitive enforcement measures, in one of our states a government official is automatically jailed for 1 or 2 nights if they close off an open-meeting no matter the reason. I personally think this is an appropriate punitive measure, but others may think it harsh. I don't think either approach would work -- both would likely make the system worse. What we need are both a court system and patent office that are sufficiently staffed and have a minimum competence. (I don't mean to insult -- I say this in the context where an official is utterly out of their realm of expertise).

fauigerzigerk 15y ago

What makes you think your company deserves to be protected against others independently developing a similar algorithm?

atakan_gurkan 15y ago
That situation is an unfortunate side effect. The point of patents is that the inventor requires an incentive (a limited term monopoly in case of patents) to make their methods publicly known. The protection is against the copiers, but since an independent developer is indistinguishable from a copier, they get excluded from using the idea as well. I always thought that this is precisely why the patentable ideas are non-trivial.
- edanm 15y ago
  You're right. But even non-trivial ideas tend to occur to several people at once, which is a part of the problem.
bostonpete 15y ago
This argument can be made against patents in general. If people want to argue that all patents are bad, that's fine, but most people draw a tenuous distinction between software patents and other kinds of patents. I agree with DrJokepu -- the real problem (assuming that you're not opposed to all patents) is that patents are being given for ideas that are moderately to extremely obvious.
- prodigal_erik 15y ago
  It can only be a net win for society if the total price everyone pays to license a patent is less than the total cost of making each user reinvent it for themselves for the next twenty years. That's why obvious software patents are so bad—the basics of the young field weren't already claimed decades ago, the cost of "reinvention" is so low (one-click requires what, a napkin, a crayon, and half a beer?), and the licensing prices devolve to pure rent-seeking based on the necessity of the obvious idea. What we need is some counter-balancing force that ensures patent license prices remain reasonable. Normally that's done by the threat of substitution or inventing-around, which doesn't work when USPTO screws up and allows claiming the whole problem space. The price especially needs to be reasonable for non-profit use, which is a software-specific issue made possible because our unit costs are also basically zero. It would also help if they tested patent claims for obviousness by giving someone the problem statement and seeing if they produce a solution off the cuff covered by the same claims.

shad0wfax 15y ago

What if your competitor came up with the same algorithm (or very similar) and performs as well both in time and space complexity... and all of this on their own, without actually taking a peek at your code. In that case would you go after them and sue them?

DrJokepu 15y ago
I imagine we would, simply because we can; we're a commercial entity after all. Personally (and I'm not representing my company here) I don't think that's a good thing; if they could prove that they independently reached the same results without using any of our research or work (or we couldn't prove the contrary), they should be protected. That's not the case though at the moment unfortunately.
- shad0wfax 15y ago
  I agree, its not the same case. The problem might be in 15 years from now, your algorithm might just become a very natural way to solve that specific problem and at that point how your company plays this out matters. My guess is the companies that are in the business of making, creating, producing and then selling software themselves, will be ok with it. These trolls like Lodsys try to neither produce nor create. A very important rule for the any type of patent should enforce "active" product in use/development. Sadly, it might never happen.
- shad0wfax 15y ago
  I agree, its not the same case. The problem might be in 15 years from now, your algorithm might just become a very natural way to solve that specific problem and at that point how your company plays this out matters. My guess is the companies that are in the business of making, creating, producing and then selling software themselves, will be ok with it. These trolls like Lodsys try to neither produce nor create. A very important rule for the any type of patent should enforce "active" product in use/development. Sadly, it might never happen.

guelo 15y ago

Nope not good enough. Software patents are doing much more harm than the potential gain from protecting your company.

CodeMage 15y ago

Threshold is one problem. Duration is another. Here's a little story to illustrate why patent trolling is so evil. I used to work at a software company that has an incentive program for developers that come up with patentable ideas, e.g. algorithms. You get a token prize when "your" patent gets submitted -- a pencil or something like that, I don't remember anymore. Then, if "your" patent gets accepted by USPTO, you get a more substantial prize, along the lines of a week at some comfy resort. A lot of my co-workers asked the same question: "Yeah, okay, that's nice, but there's a good chance that the company is going to make shitloads of money off a successful patent. A week at some resort doesn't really compare to that. Why don't you guys offer something that does?" The guy who was explaining the incentive program responded: "Because you patent an idea and that's worth exactly nothing until you implement that idea in a product and sell that product to customers. You had the idea, but it's the company that has the means to implement it and market it and sell it." When you think about it, it's reasonable. Not precisely fair, but then again, things rarely are. The point of this story is that software patents, just like all patents, are there to give the inventor a chance to implement the invention and make money from it, as a reward for the inventor's genius and hard work that went into coming up with an invention. That's obviously a good thing, because it gives you an incentive to invent something without having to fear that someone will simply steal your invention and make money off it. Like every regulation with good intentions, it gets abused. What you said about trivial patents is just one way it gets abused. But another thing that makes it easy to abuse software patents is the duration of the patent. Twenty years is a hell of a long time in software. It practically invites patent trolling.

Joeri 15y ago
I've never heard a case where the do-or-do-not decision regarding a software project hinged on patents. That they are somehow needed to encourage innovation is a myth. Besides, they're unfair. A typical software product might have perhaps 1% patented matter, and 99% copied from the "standard requirements" of the market segment it's in. The competitors spent the money developing the market to figure out that 99%, so if you use a quid pro quo measure then they have an automatic right to the 1% innovation the new player brings to the table.
- 15y ago

citadrianne 15y ago

I think it's hard to predict what abolishing patents would do to innovation. For example, there are no patents in fashion, which actually ends up putting pressure on creators to be superinnovative to stay ahead of the copycats. Ref: http://www.ted.com/talks/view/lang///id/866 http://www.ted.com/talks/view/lang///id/866

mindcrime 15y ago
My personal opinion - for what that's worth - is that this kind of scenario comes pretty close to representing what would happen if there were no patents in software (or any other field for that matter). Innovation would still happen, and would probably happen even faster than it does now, because of the "arms race" effect. Simply put, a firm would be forced to innovate constantly, quickly and fiercely, to stay in front of its competitors. Combined with the fact that I am appalled by the idea of use of government force to block someone from using an invention - that they may have arrived at independently - just because someone else patented it first; and my belief that there are too many smart people in the world for any really useful idea to stay undiscovered long, I advocate for simply abolishing the entire patent system.
- biafra 15y ago
  In Europe there are no software patents. Do they develop software and innovate there?

Vivtek 15y ago

While this has no bearing on the law, I've always thought that works of software that represent an actual effort of development comparable to physical invention should enjoy the benefit of patents. This is precisely what the patent system was designed to do - protect large research investments. The abuse of the system for things that really don't require research and development is at issue here. I think proposals to abolish software patents entirely are probably wrong - but granting patents on XOR, or 1-click, or doing something obvious but on a computer system, are equally wrong. Put it this way: if I have a problem to be solved by software that is really hard (such as your algorithm), then I'm going to research it before spending five man-years developing it, find you guys, and happily license your solution. No question. And if somebody reverse engineers your solution and sells it, then by God they should be penalized for violating your patent. This is why we have patents. But if I have an idea like 1-click and pay an Indian developer $10 an hour for a week to develop it, this is simply not worthy of monopoly protection; your disclosing your solution isn't that valuable. If, seventeen years from now, after I've built an online shopping empire while you haven't, you sell your patent to Lodsys and they sue me for a billion dollars out of the blue, that is highway robbery, and it's not in any way helping society - and that is the situation we actually currently have.

anamax 15y ago
> I've always thought that works of software that represent an actual effort of development comparable to physical invention should enjoy the benefit of patents. There is no "effort" requirement for patents on physical inventions. There's also no "it was done by experts" requirement.
- Vivtek 15y ago
  Yes. That's true. That's why I said it has no bearing on the law.
- lurker19 15y ago
  Incorrect. http://en.m.wikipedia.org/wiki/Person_having_ordinary_skill_in_the_art http://en.m.wikipedia.org/wiki/Person_having_ordinary_skill_...

noonespecial 15y ago

Unfortunately those of you who really deserve patents will not be able to make them rise above the noise of this lunacy. We're eventually going to have to go with "no software patents at all" to fix this. As a society, we just haven't outgrown the greedy thuggery that could allow us to respect each others work without gaming the system into tragedy of the commons territory. A classic case of "see, this is why we can't have nice things".

jpadvo 15y ago
I don't think that is inevitable. There is not an inherent reason why the Patent Office needs to award trivial patents. If they tighten up (or apply) their standards, the problem would be tremendously reduced.
- rapind 15y ago
  Coming up with relevant criteria to determine the triviality of a software patent application doesn't seem simple at all. This means you need to hire experts in the field. This is where it gets interesting. In most regulatory organizations where a high amount of expertise is required (like economics) you see a lot of corruption where companies hire out of and into these organizations specifically for the benefits of having representation on the inside. I.e. Henry Paulson. So unless you have a good way of preventing these employees from ever working for Microsoft / Facebook / Apple / Google etc. then it's really not going to work out as intended. IMO, just get rid of software patents. Sure it'll suck for some, but overall it's probably healthier for the industry. Plus it's nice and simple and doesn't employ a whole mess of lawyers.
- seats 15y ago
  The USPTO is overwhelmed and understaffed, so I can understand why they have poor QC on patents, particularly when they require technical knowledge to determine triviality. What I don't get though is why it seems to be such a challenge to prove the invalidity of these crap patents in court. The google loss to bedrock (the linux kernel troll) is a high profile example of this. Maybe the standards that need to be looked at are around litigation and if enough crap is thrown out, all the incentive is gone and all of the sudden the USPTO application volume goes way down.

SoftwareMaven 15y ago

Coca-cola has managed to deal with this for a hundred years. Google has managed it for 10 years. These things are called trade secrets. You protect them through NDAs, being intelligent with who has access to the information, and not open-sourcing the code.

aridiculous 15y ago
In the case of Coca-cola, at least, the secret formula isn't even much of a concern anymore (I think it's been leaked a few times actually). Them protecting the formula is really just tradition and adds to their branding. What's important to them is keeping their brand power.

georgemcbay 15y ago

I'm OK with the idea of software patents in theory just like I'm OK with the idea of the death penalty for extreme crimes in theory. Because we live in an imperfect world made up of imperfect people, I don't support either in practice. Perfect software patent system > no software patent system > broke software patent system, but I believe the perfect system is impossible, so I'd rather see the software patent system dismantled completely than exist in a broken state.

fredwilson 15y ago
i love this analogy and i am with you on both

jellicle 15y ago

Except that you could already use and profit from the algorithm indefinitely without any patents in place whatsoever, simply by keeping the algorithm secret.

uiri 15y ago

I don't understand how you would be handing it over to your competitors: 1) if it is proprietary software, just add a clause which forbids reverse engineering. If it really is as complex as you describe however, I doubt reverse engineering it would be an option. 2) if it is open source software, I don't see the big deal. Even if they get the algorithm your software can still stay steps ahead of theirs. So what advantage would a patent even give you in this case?