Let Re-Discovery Evade Patents

In this post I’m going to explain why patents can be a good idea, why they often go wrong today, and a way to fix that problem. And I’ll do that all in the context of a situation you should understand well: finding a shorter route to drive from home to work. (This post is ~1600 words, and so longer than usual.)

Imagine that you usually take a particular route from home to work, and some firm offers to find you a better route. You tell them your current route, and they tell you that they have found a different route that will save you thirty seconds a day, which over a year adds up to eight hours. You can inspect their route to verify their claim, but only if you agree that you can’t use that route (or anything close) unless you pay them a mutually agreeable fee. (Assume they can enforce that, by seeing your car’s driving path records. And assume you can verify their claim somehow.) You agree, inspect and verify, and then agree to pay them one hundred dollars, which is well below your value of saving eight hours of driving, and above their cost of finding the route.

This example contains an info property right: once you agree not to use their route unless you pay for it, then they own a right to your use of that route. Since the route is info, what they own is info. The prospect of owning that info right gives the firm an incentive to work to find that route. And because they must find a mutually agreeable price, their incentive to work is neither too much nor too little. An agreeable price must lie between their cost of finding the route and your added value from using it.

Now imagine that you are one of hundreds of drivers who go from the same initial home area to the same final work destination. Now this route-finding firm wants to sell a better route to all of you. But there is a problem. Once this firm sells the route to a few of of you, the others may learn of that route from these few buyers, either by being told or by following their cars. In this case the total price the firm could get from all the drivers might be much less than the sum of driver values for using the better route. Thus the firm’s incentive to work to find a better route could be too low. That is, this group of drivers could be better off it they joined together to paid the firm more to find a better route. But joining is too hard, so it doesn’t happen.

This problem is often cited as the reason governments allow patents. As an analogy, imagine there was a public website where route-finding firms could post pairs of old and new routes, so drivers easily find the new route matching their old route. The new legal rule might be that once some firm had posted such a route pair, no driver may legally use the new route matching his or her old route without paying the firm a mutually agreeable fee. This new patent-like system would restore the good incentives of the case of just one driver and one firm; an agreeable price would again lie between a firm’s cost of finding the route and a driver’s value of using it.

At least, this new system would have good incentives if it was very easy for drivers to look up better routes on the public website, and if drivers had no better way to find faster routes, and if the legal system could accurately tell what routes drivers would have taken without the public website. To see how this could all go very wrong, consider some variations.

If drivers had GPS devices to tell them about better routes, drivers might still be required to pay hundreds of dollars to route finding firms to be allowed to use routes they could have learned about for pennies from their GPS. Or, imagine these rules applied to drivers even before they moved to this city or found a job, and the system assumed that without the website such drivers would have taken a very bad route that takes twice as long. Or that drivers would have walked instead. In such cases the route-finding firms might charge drivers for the value of being able to take any reasonable route to work, rather than just the added value of their better routes.

These variations would pay route-finding firms way too much. By being the first to register a route, a firm could charge drivers the value of using that route at all; the only check would be competition between firms with related routes. And if lots of route-finding firms got together to pool their routes, like firms do today with patent pools, the route-pool could charge drivers the value of being able to drive to work at all. So there’d be a huge wasteful rush to register routes, after which the route-pool would basically enslave all driving workers, yet add little in the way of social value.

The key problem here is that the social value of creating property rights in info comes from the value of using that created info, over and above the info such users would have had without that creation. If you instead assign them property rights in being able to use any remotely similar info, you can reward them far too much, crediting them with far more than the value they actually provide.

A very similar problem happens today with software patents. If there is a library somewhere with patented algorithms for sale, the social value of a software developer using that library to achieve some purpose is the difference between the cost to search that library for an algorithm that achieves their purpose, and the cost of just making up an algorithm that achieves their purpose. Usually that difference is negative – it is usually easier to make something up than to look for prior solutions. And even when the difference is positive, that value difference is usually much less than the value of being able to achieve the purpose somehow.

In most patent violation court cases, people are accused of accidentally violating a patent. In the US, patent holders can actually get triple damages if they can show that someone violated a patent on purpose. But even with that added incentive, such suits are rare. So the usual case is where someone tried to solve a problem by themselves, and the solution they hit upon just happens to be one someone else found before and patented.

We have little grounds for thinking patents have much social value in such cases. After all, the developer considered looking up and buying a patented solution, and decided that it would be cheaper and easier to just make up their own solution. Yes they might have made a mistake, but they are unlikely to be systematically biased. Yes, perhaps the patent fee made all the difference; if they thought they could use a patented solution without paying its fee, maybe they’d have searched for it in a library. But given the huge ranges over which relevant parameters vary, such cases must be rare.

Most likely, the developer made a socially-good choice to not search. In which case, the prior discovery of the patented idea actually had no social value for this application. And even if it did have a social value, that value would often be far smaller than the value of using the rediscovered idea to solve their problem, which is the high amount patent holders usually try to charge. That high amount seems much worse than zero as an approximation to the social value the patent holder could have provided in this case.

I think this gives us a great reason to excuse such developers from patent violations. That is, independently re-discovering a technique should excuse developers from using that technique without permission of the patent holder. Such re-discoverers shouldn’t be able to sell the use of that technique to others, but they should be able to use it themselves. And freeing developers to use any solutions they can figure out for themselves would greatly free the software industry to more open vigorous competition, competition that patent pools now discourage.

Now you might claim that even though patent holders rarely sue to get the triple damages available for intentional violations, most violations really are on purpose – such things are just much harder to prove. But at best that would justify a presumption that claimed rediscoveries are no such things unless supported by concrete data.

I’m told that today, developers are sometimes isolated from outside news in an attempt to avoid any accusations of intentional patent violation. But that seems pretty painful to me.  Today’s surveillance society offers an easier solution. Imagine that a developer recorded all the articles they read, the talks they hear, conversations with outside developers, and intermediate versions of software they develop. All of this might be made available for inspections during a patent violation lawsuit. If a patent holder can’t find any holes in their re-discovery story after looking at all that, the story is probably true.

This general approach probably works well for most non-software patents as well, but I feel less confident of that. I did software for many years, and think I at least have a good feeling for how that works.

I was emboldened to write this post in part by a talk with Alex Tabarrok (who posted on it here), and by finding similar ideas endorsed by Mike Masnick here.

GD Star Rating
loading...
Tagged as: ,
Trackback URL:
  • roystgnr

    “shouldn’t be able to sell the use of that technique to others” – does that include “selling” it by releasing it under a copyleft license, or even simply giving it away by releasing it unrestricted? If open source can’t include easily reinvented patented ideas, then the patent system remains deeply flawed. But if open source programmers can reinvent and release such ideas, the “damage” to the patent holder is probably little less than if the reinventions could be sold too.

    • IMASBA

      That’s a problem you don’t have with patentable stuff other than software, with software you could just refrain from making your own re-invention public, but you’d first have to know the piece of software was patented in the first place (this is hard to find out and that’s one of the reasons so many people get sued without knowing they’ve done anything against the law). It kinda sucks that you pretty much can’t do anything in software unless you’re so rich you can afford to buy all the licenses or you stick to (usually completely useless) stuff that no one else has ever published software about. Really makes you wonder if having software patents really benefits society.

  • IMASBA

    “That is, independently re-discovering a technique should excuse developers from using that technique without permission of the patent holder. Such re-discoverers shouldn’t be able to sell the use of that technique to others, but they should be able to use it themselves.”

    Surely this is already allowed, at least for non-software patents? If not the US patent system is even more f*cked than I thought (American patent laws already allow for the patenting of ideas without a working prototype, geometric shapes, button functions, obvious derivations of other patents that only serve to continue control over the original product and other such BS, and don’t even get me started about 100+ years copyright terms).

    Anyway software patents are something peculiar, they fall right in the middle between copyright and patents. Software can be created on demand (and quite cheaply too, most of the costs come from marketing which wouldn’t be necessary without patents in the first place), unlike art, many rich organizations (corporations, governments, universities) would develop it anyway, even without patents, the source code can easily be kept a secret and the probability of coincidental re-invention is relatively high, much higher than with copyrightable materials, and if a piece of software is so revolutionary only a single group can make it then it shouldn’t be difficult to effectively hold a patent by keepin the source code secret (if some other group reinvents the source code before the other group makes a profit that’s just proof the software wasn’t so revolutionary after all), so maybe we shouldn’t have patents for software at all. Just let videogames stay under copyrightable material and I’m sure the world won’t end.

  • Douglas Knight

    Yes; I’d go farther and say that without any change in law, rediscovery should count as prima facie proof of obviousness, invalidating the patent.

    So the usual case is where someone tried to solve a problem by themselves, and the solution they hit upon just happens to be one someone else found before and patented.

    That’s just not true. In the vast majority (90%?) of software patent lawsuits, the solution is completely different, but the patent is broad to the point of meaninglessness.

    So
    the usual case is where someone tried to solve a problem by themselves,
    and the solution they hit upon just happens to be one someone else
    found before and patented. – See more at:
    http://www.overcomingbias.com/2013/09/let-re-discovery-evade-patents.html#sthash.l3wBGUem.dpuf
    So
    the usual case is where someone tried to solve a problem by themselves,
    and the solution they hit upon just happens to be one someone else
    found before and patented. – See more at:
    http://www.overcomingbias.com/2013/09/let-re-discovery-evade-patents.html#sthash.l3wBGUem.dpuf
    So
    the usual case is where someone tried to solve a problem by themselves,
    and the solution they hit upon just happens to be one someone else
    found before and patented. – See more at:
    http://www.overcomingbias.com/2013/09/let-re-discovery-evade-patents.html#sthash.l3wBGUem.dpuf
    So
    the usual case is where someone tried to solve a problem by themselves,
    and the solution they hit upon just happens to be one someone else
    found before and patented. – See more at:
    http://www.overcomingbias.com/2013/09/let-re-discovery-evade-patents.html#sthash.l3wBGUem.dpuf

    • http://overcomingbias.com RobinHanson

      Too broad perhaps, but hardly meaningless. That would imply that who is said to violate which patents is completely random.

  • Douglas Knight

    Also, don’t be a jerk.

  • Robert Koslover

    Independently rediscovering a technique does not, and probably should not (in my view), allow you to simply ignore the IP rights of a pre-existing patent holder. In my own work, I quite often create “original” inventions, only to discover after subsequent searches in technical literature that my “original” invention is, wholly or partially, pre-existing. When this happens, I either: (a) console myself with the knowledge that “great minds think alike,” or (b) see if I can possibly advance the technology into a new realm or towards a capability that the earlier inventor simply did not anticipate. If/when the latter approach is successful (and sometimes it is), I may be well-positioned to obtain a patent of my own. And I think that is fair. That said, I’m not sure that there is any kind of reworked patent system possible that could ever be 100% fair to everyone. But a practical patent-law framework does seem to be important to the advance of science and technology. Our founding fathers certainly thought so, and they were right about a lot of other things. See http://en.wikipedia.org/wiki/Copyright_Clause .

    • http://overcomingbias.com RobinHanson

      I don’t much care what is fair. I care what is efficient. I outlined efficiency arguments for excusing re-discovery.

      • Robert Koslover

        Exempting rediscovery is problematic in my view, since I doubt one can prove reliably (at least, with present technology) the provenance of ideas in the re-discoverer’s mind. Meanwhile, it is worth emphasizing that a patent is issued not merely as a reward, but as an exchange of valuable entities. In exchange for openly *teaching* (and this is critical) others about the detailed nature of one’s invention (via the published patent), which is in fact required to show clearly, to others, how to reproduce it, and (eventually) for it to be used freely by others, the inventor is granted (but only temporarily) exclusive rights to make/produce/sell it. The inventor is free to license those rights to others, also. The Founders’ intentions in creating this patent system was to encourage the use and spread of technological knowledge, not to hold it back. When inventors are granted few or no rights (and yes, I have known of such circumstances), they have great incentives to hide the details of their inventions from others (and yes, I have known them to do exactly that). The real challenge is thus to establish the best (most “efficient,” if you prefer) balance of rights to be granted in exchange for “teaching” this knowledge to others. Hence, the ongoing arguments over how long patent rights, once granted, should last. I do not maintain that the status quo is perfect (or even excellent), but it’s purpose, and considerable successes, need to be appreciated carefully if one is going to succeed in making it better.

      • IMASBA

        “Exempting rediscovery is problematic in my view, since I doubt one can prove reliably (at least, with present technology) the provenance of ideas in the re-discoverer’s mind.”

        Actually with software it’s often easier to re-invent something than to find the existing patent (I find it completely plausible the majority of people getting sued really did not know about the patent and if it’s so easy to re-invent the software it’s not revolutionary and did not necessitate a large investment, meaning it does not require patent protection) and with non-software patents re-creating something for your own personal use is often so difficult and time consuming that it becomes silly to sue the re-creator to disincentivize them, plus you’d get into absurd situations, especially when you apply your thinking to copyright as well, it would basically mean you could get sued for singing a copyrighted song or educating students by letting them program a patented algorithm. A patent is a legalized monopoly, it has to have limits and drawbacks, the US patent system is already the worst of pretty much the entire world (where else can you patent the rectangle?)

        Also, please don’t draw “the founding fathers” into this: they’re the same guys who didn’t care much for the rights of women and blacks, heck, even of every white male who was not a landowner, plus it’s not going to convince non-Americans like myself.

      • Maximum Liberty

        Robert:
        The notion that software developers (at least) would ever read patents to learn how to build something is far fetched. Here is my evidence for that proposition:
        1. I have worked for technology companies of all sizes for 18 years and have never seen it happen or heard of it happening.
        2. Patents are inscrutable. Lawyers and patent agents write them with a view to maximizing the value of the patent by using terms that could, in the future, be interpreted to mean something like what was invented, but not exactly what was invented. This makes them very difficult to understand. If we really wanted understanding, then patent applications would be required to include several concrete examples, including (for example) actual software code.
        3. In part because of point #2, lawyers at software companies often warn their developers not to read patent applications, because doing so opens the compan up to triple damages in the unlikely scenario that one of those ambiguous patents gets applied to their business. This reinforces point #1.
        I like the idea that Robin describes. While he disclaims fairness, it seems fairer than the current approach.
        Also, I think that Robin’s point about re-discoveres being prohibited from licensing the re-discovered invention helps reduce the importance of the lower incentive to teach about your patent. It might actually improve the teaching, since patent-holders would then have a stronger incentive to give software developers and easier way to implement the patented invention, compared to re-discovering it.
        Max

  • Ely Spears

    Here is another idea: anytime prosecution occurs over an accused violation of a patent, automatically review the validity of the patent. Any patent case could then have three outcomes by default: (1) the accuser is vindicated, the accused is forced to pay damages for patent violation, and the patent is upheld in current form; (2) The accuser is rebutted, the accused is vindicated and not made to pay damages, and the patent is upheld in current form, (3) the patent is deemed inappropriate due to e.g. being too vague; no damages are awarded but the patent must be reformatted and refiled so as to make more specific and clear the protections offered by it. (I’ll consider other possibilities, such as the accuser being awarded damages but also made to re-format the patent to be very unlikely.)

    This might give inventors/firms more of an incentive to acquire very specific patents and give them some pause to consider the seriousness of an accusation of patent violation they plan to make — since having the patent reformatted to something less favorable is always automatically on the table.

    Homo hypocritus thinking suggests to me that firms and cabals would very much want vague, broad patents and would relish the ability to bring patent accusations against anyone at any time for seemingly random reasons. Patent trials should function somewhat like the Salem Witch Trials. But if there were frequent, implied opportunities to fix broken old patents, it might make it less worthwhile to prosecute e.g. re-discoverers who only plan to use the re-discovery privately. Only if that private use was really worth risking the current form of the patent would it be worth taking legal action.

    • IMASBA

      Or, you know, just implement a patent law that actually belongs in the 21st century… Just copy the European law if you have to.

    • lemmycaution

      patent validity is a currently a big part of almost all patent litigations.

  • lemmycaution

    Most efficient solution is to totally get rid of software patents. Rediscovery is so common in this area that it makes no sense to pretend that patents spur development in software. Software patents are an after thought.

  • Tom Breton

    I blogged about fixing the re-invention problem here (http://tehom-blog.blogspot.com/2013/09/patent-fix-1.html)

    Basically, the idea is for the inventor to pose the problem first and let it sit for a while. We’d use idea markets and a random bounty to estimate which problem’s solutions are likely to be reinvented and act accordingly.

  • Jess Riedel

    > This general approach probably works well for most non-software patents as well, but I feel less confident of that. I did software for many years, and think I at least have a good feeling for how that works.

    Regardless of your personal experiences, shouldn’t any proposed solution to the problem in software patents explain why the solution is particular to software? Any solution that should work just as well for non-software immediately prompts the questions “If this was a good idea, wouldn’t it have been tried before?” and “If software patents are unusually broken and you are applying a fix that isn’t software-specific, isn’t it unlikely that you’re fixing the underlying problem?”

    • Robin Hanson

      In software it is much harder to find out what solutions have been patented and if the solution you have in mind is one of them. Which makes reinvention more common and believable there. Patent law only changes very slowly, so it is easy to believe it isn’t optimal.

  • Pingback: Recomendaciones | intelib

  • http://www.selfishmeme.com/ The Watchmaker

    For what it’s worth, I spoke with a couple patent attorneys a week ago about IP protection for my start-up company. The impression of both law firms was that patents for “abstract ideas” are no longer worth filing. While 1998 through 2010 were a heyday of patent creation, neither law firm thought they would be enforceable in the current legal climate.

    Apparently the Supreme Court undermined the previous standard for software patentability in Bilski v. Kappos in 2010. Other relevant anti-patent results include Mayo v. Prometheus and CLS Bank (not yet seen in the Supreme Court).

    http://en.wikipedia.org/wiki/Bilski_v._Kappos

    http://en.wikipedia.org/wiki/Mayo_v._Prometheus

    http://www.crowell.com/NewsEvents/All/Patent-Law-Alert-A-Deeply-Divided-Federal-Circuit-Fails-to-Bring-Clarity-to-Patent-Eligibility-of-Software-Related-Inventions-in-Closely-Watched-Case#.UjdQbcbbNyU

    My point in sharing these findings is that we’ve moved far to the left on Tabarrok’s curve, and no one seems to have acknowledged that fact. For better or worse, software patents are already almost entirely eradicated.

    • http://overcomingbias.com RobinHanson

      Recent rulings makes less clear what will be enforced. Which is very different from a world without software patents.

  • Pingback: Put Up Your Code or Shut Up: A Modest Proposal for Software Patent Reform | The Ümlaut

  • Ronfar

    If you were planning on intentionally stealing a patent, it’s hard for the patent holder to prove that it *wasn’t* an accidental rediscovery…

  • Ronfar

    A non-obvious problem with patent litigation: whether or not one infringed a patent may depend on specific technical details – and the people who have to make a decision are twelve laymen with no knowledge of the field. Good luck explaining the difference between a phase-locked loop and a ring oscillator to a bunch of English majors!