Hindsight Bias

In my patent policy seminar at Boston University Law School, we recently studied hindsight bias and how it can play-out in the innovation sphere.  Professor Gregory Mandel at Albany Law School has a pair of great empirical papers on the topic here and here.  Mandel concludes that a juror’s knowledge of the invention makes the juror much more likely to believe that the invention would have been obvious.

To reach a proper non-obvious conclusion, the decision-maker must step backward in time to a moment when the invention was unknown. Unfortunately, Humans are cognitively incapable of ignoring what they have learned, as required for the proper ex ante analysis.

In class, I followed-suit and gave the students a test of their own hindsight bias.  In one group, students were given a problem and were also told how it was solved.  The other group was simply given the problem. Both groups were asked to rate the nonobviousness of a solution.

ForesightHindsightII

Although not statistically significant because of the small seminar size, my class results closely followed Mandel’s results.  Students who did not know of the solution were more likely to think that a solution was nonobvious.  However, after students learned how the problem was solved, they were then more likely to think that the solution was obvious. This is the crux of hindsight bias. 

Notes:

  • Something that makes the results even stronger is that before the class, all the students were assigned the Mandel hindsight bias reading and they also knew that the topic of the day was hindsight bias. 
  • Professor Joseph Miller at Lewis & Clark Law School is working on this issue as well — he gave me the idea for this specific problem-solution.
  • Peter Zura discusses hindsight bias here.

68 thoughts on “Hindsight Bias

  1. Why the candlestick? The problem only says he can’t reach both ropes “at the same time”. Therefore, why can’t he grab one, walk to other and then tie both together?

    Presumably, the ropes are too high off the ground to do this and thus require the chair and candlestick/pendulum approach, but the problem doesn’t say this.

  2. the question isn’t “Is the solution obvious to me?” or “Is the solution obvious to the average man?”, the question is “Is the solution obvious to a highly skilled, bright individual trained in solving problems of this type?” because, after all, this is the sort of person who would be asket by a company to solve a problem. Take the problem above. To me, the solution was obvious without reading it – it took me approximately 5 seconds to come up with the solution on seeing just the problem. And if you asked a selected group of individuals who might be hired to solve such a problem, I suspect they would all say the same – the solution is obvious.

  3. pds:

    “Exactly how I wrote it: “patenting around” with the quotes. Without the quotes and you get 811K hits. The term of “patent around” with the quotes yields 535 hits.”

    AND I hope you had the perception to see that the hits did support my position. Broadening the search without the quotes would’ve yielded more hits, which I noticed that you did not dispute in regards to the hits supporting my position. Boy, I hope you do a better job for your clients in the preliminary search assessment before taking their money.

    I THINK THIS PRETTY MUCH PROVES all of your arguments are dubious at best.

    “I understood your example perfectly. However, the terminology you use make it perfectly clear that your level of sophistication in the business is at an introductory level.”

    I don’t think you do because I had to really simplify all this discussion for you — and you still don’t get it. Your obsession with terminologies, even when you’re WRONG as the above demonstrates, shows how useless it is to have a discussion with someone who simply just can’t understand the substance of material.

    “… do you really understand what you are saying? Do you understand the meaning of negligence? If you had, you would have never used that term in your statement, but it doesn’t fit.”

    Here we go again, do YOU? Do you understand the context of the sentence? What did your magnificent search yield this time?

  4. “This thread is becoming a reprise of the one for KSR.”

    Has anyone been told to quit their job and get psychiatric help yet?

  5. “Wow, let me simplify further. If you have really broad claims, it IS economically and/or technologically feasible to patent/design around it with an obvious modification because the coverage would be essentially the same.”

    I don’t think you and I are on the same page with regard to what “really broad claims” mean. In my book, really broad claims mean that the claim coverage is very extensive. Examples: a broad claim would be to “a chair.” A narrower claim would be to a three-legged chair, a high-back chair, a leather chair, a foldable chair, etc. All the narrower claims would likely not be obvious in view of the broad claim to “a chair.” However, all these other claims would be covered by a claim to “a chair.”

    “How did you search? My search yielded half a million hits and my perusal of the first few pages of highlights indicate they are indeed relevant.”
    Exactly how I wrote it: “patenting around” with the quotes. Without the quotes and you get 811K hits. The term of “patent around” with the quotes yields 535 hits.

    Like examiners need to consider the claimed invention, as a whole, you need to consider the term, as a whole.

    “If you could not understand my example of stealing coverage from a broad patent [buffer] using an obvious patent [fifo], then it proves my point that the negligence of PHOSITA is ruining the system.”

    I understood your example perfectly. However, the terminology you use make it perfectly clear that your level of sophistication in the business is at an introductory level. One does not “[steal] coverage from a broad patent.” Two patents may cover the same feature (i.e., one patent covers the broad concept while the other patent covers a specific implementation of the broad concept). Regardless, coverage is not “stolen.” The patentee with the broad patent does not lose any coverage when subsequent patents are issued that cover subject matter also covered by the broad patent. For you to even insinuate that coverage is lost clearly indicates that you really need a better understanding of patent law.

    Regarding your statement that “the negligence of PHOSITA is ruining the system,” do you really understand what you are saying? Do you understand the meaning of negligence? If you had, you would have never used that term in your statement, but it doesn’t fit.

  6. “All of that functionality was anticipated by 1999 and however complex the software may be to accomplish it, that software will primarily be the application of techniques well known by 1999. There may or may not be anything new.”

    Really??? If “[a]ll that functionality was anticipated by 1999,” then why are we having software patents issued today. A rejection based upon anticipation (i.e., 35 U.S.C. § 102) doesn’t care about hindsight or require a teaching, suggestion, or motivation. If what you have argued is true, then there is no need for obviousness rejections.

    “And to Mr. Stonecker who is concerned that this thread is starting to echo KSR, I take his point and will post no further. However, “obviousness” goes to the heart of the purpose and current problems with the patent system. It is almost inevitable that discussions about the patent system will circle around to obviousness, particularly a discussion of such a closely related concept as hind sight bias.”
    We actually agree on something. Preventing impermissible hindsight (i.e., the topic of this thread) is a critical part of an obviousness rejection, or at least this is what has been argued by those supporting the respondents in KSR. As such, when discussing hindsight, it is almost impossible not to have the discussion not touch upon KSR since the SCOTUS will ultimately determine how important hindsight becomes.

  7. Returning to the question of “obvious”, I think that there’s a sliding scale.

    (1) If I know that a problem is commercially interesting, I might be motivated to think about it. The answer is not so obvious that it would come to me if I were not motivated to think about it.

    (2) If I know that a problem has a solution (perhaps it is impossible), I am more likely to be able to solve it. Simply being told that it is possible to join the two strings, makes a solution more likely.

    (3) The other dimension for “obvious” is whether a solution involves a single non-intuitive or surprising trick, or a large number of subproblems, none of which is in itself hard, but the exact path is not known (not “obvious”) until the problem is worked.

    The reason we have multiple similar patents filed in clumps, is that there’s enough smart people out there to solve non-obvious problems once (1) occurs.

  8. “… if you have really broad claims, then designing around the patent is not economically and/or technologically feasible. I thought I made the perfectly clear.”

    Wow, let me simplify further. If you have really broad claims, it IS economically and/or technologically feasible to patent/design around it with an obvious modification because the coverage would be essentially the same. This is highly possible under the current system of nonobviousness test requiring TSM because fear of hindsight has completely ignored the statute consideration of PHOSITA.

    “BTW: A google search of the term “patenting around” yield 218 hits, most of which weren’t relevant to this discussion. As such, I believe my questioning of the use of the phrase “patenting around” was justified.”

    How did you search? My search yielded half a million hits and my perusal of the first few pages of highlights indicate they are indeed relevant.

    I just realized that perhaps you’re not a software guy. Your remarks indicate that you really are a little naive concerning the innovation of software. If you could not understand my example of stealing coverage from a broad patent [buffer] using an obvious patent [fifo], then it proves my point that the negligence of PHOSITA is ruining the system.

  9. “Please don’t tell me that all these software-related technologies were available back in 1999. Do you have any idea as to the complexity of all the back end software that is used to support these websites and all the patents that have been procured on this software?”

    All of that functionality was anticipated by 1999 and however complex the software may be to accomplish it, that software will primarily be the application of techniques well known by 1999. There may or may not be anything new.

    And to Mr. Stonecker who is concerned that this thread is starting to echo KSR, I take his point and will post no further. However, “obviousness” goes to the heart of the purpose and current problems with the patent system. It is almost inevitable that discussions about the patent system will circle around to obviousness, particularly a discussion of such a closely related concept as hind sight bias.

  10. Some observations:

    This thread is becoming a reprise of the one for KSR. I do not believe Dennis intended it as such.

    “Spirited” differences of opinion seem to be leading to negative comments about persons holding contrary views. This seems inappropriate in a professional forum.

  11. “Let me make it very simple for you: you design around a patent and then patent that design, which would be patenting around.”
    Let me make it even simpler for you, if you have really broad claims, then designing around the patent is not economically and/or technologically feasible. I thought I made the perfectly clear.

    BTW: A google search of the term “patenting around” yield 218 hits, most of which weren’t relevant to this discussion. As such, I believe my questioning of the use of the phrase “patenting around” was justified.

    “An example: little Johnny has a truly ingenious patent (e.g., network address buffering). Some biggie company sees this patent and designed/patented around it (e.g., network address buffering using FIFO queue).”

    Nice example. However, I’ve seen the “little Johnnys” of the world make a boatload of cash, even with very limited resources. If little Johnny has enough money to get a patent drafted, prosecuted, and issued, he has enough money to have an attorney send a couple cease and desist letters (which may include offers for settlement). Once the first settlement check comes in, then little Johnny is off to the races.

    “Examiner, exhausted by hundred of other similar cases all screaming no TSM, finally issues the obvious patent.”
    I always like the straw man you guys introduce … the assumption that the patent is obvious. Once you’ve introduced that assumption, you can dream up scenarios to your heart’s content and play up the hindsight/TSM boogeyman.

    BTW: I have no idea how Biggie could “[trample] over little Johnny.” If, per your hypothetical, little Johnny’s patent is broader than Biggie’s patent while also covering Biggies patent, I doubt very much that any intelligent licensee is going to be willing to license Biggie’s patent.

    The first thing any smart licensee should do when facing a Biggie’s demand (particularly if the demand is for a considerable amount of money) is to order up a copy of the prosecution history and made a determination as to whether or not Biggie’s patent is valid. At the same time, because the PTO has used little Johnny’s patent during prosecution, the smart licensee would realize that a license on Biggie’s technology alone would be useless. I can continue on and on, but the point is that your example evidences that you really are a little naive as to what actually goes on.

  12. “There hasn’t been a fundamental change in the nature of personal computers since the Macintosh introduced mice and graphical interfaces to personal computers in 1984 – oh wait, that is right around the time software patents were first recognized: coincidence, I think not.”

    You are joking, are you not?

    Since you have posted on this website, I’m sure you have opened up a browser. Perhaps you have gone online and purchased things. Have you been to the USPTO website and filed anything online using PAIR? Please don’t tell me that all these software-related technologies were available back in 1999. Do you have any idea as to the complexity of all the back end software that is used to support these websites and all the patents that have been procured on this software?

    “Pick whatever simple test you like so long as it systematically errs on the side of freedom.”

    I’ve written several times, and I’ll continue to write it again. It is my opinion that very many of the people advocating the modification of the standard for obviousness test through KSR v. Teleflex are not doing it because they believe the standard is wrong. These people are doing it because they are generally hostile to intellectual property, in general.

    I’m young enough to be very comfortable with all the technology that is currently available, but I’m old enough in that I got used to having to pay for copies of my music and movies. Just a guess on my part, that many of the younger generation have this belief that intellectual property is just a hindrance and that free software, free songs, and free movies should be the norm.

  13. pds:
    “… One does not “patent around” anything … one designs around a patent. Also, broad claims are what prevents a big company (or any size company for that matter) design around a particular patent, which has nothing to do with whether a particular invention is obvious or not.”

    Boy, you sure are over your head and not very perceptive. Let me make it very simple for you: you design around a patent and then patent that design, which would be patenting around. The problem with the current system is that the fear of hindsight brought about the TSM test that has allowed obvious variations of a true invention to be issued. An example: little Johnny has a truly ingenious patent (e.g., network address buffering). Some biggie company sees this patent and designed/patented around it (e.g., network address buffering using FIFO queue). Some examiner at the PTO uses little Johnny’s patent with another patent to reject (e.g., FIFO queues can be used subbed for other buffers by PHOSITA). Biggie’s representatives screams no TSM, files 100 pages arguments, instigate phone interviews with no records, files appeals, all considered to be minimal investment compared to the potential strategic benefit of acquiring the patent. Examiner, exhausted by hundred of other similar cases all screaming no TSM, finally issues the obvious patent. Biggie then uses the patent with vastly more financial resource to hire a legal team to defend/enforce the patent, perhaps even trampling over little Johnny, who was supposed to have a “broad” patent (e.g., did not bother to disclose FIFO because it should be obvious to PHOSITA to see that FIFO can be substituted for other buffers). So, little Johnny, whom the patent system was supposed to defend, according to you, has a “lesser” patent than one that he could’ve had if the system did not issue “obvious” variations of true patents.

    I think your other remarks have been more than adequately answered by other knowledgeable PHOSITAs in software.

  14. pds

    “I don’t know about you, but if you compare today’s computer and the software running on that computer to a computer and software running back in 1999, I would imagine that you would find a HUGE difference. So much for all these patents being a hindrance and not promoting innovation.”

    As a matter of fact, I’m sitting here working on a computer running Windows 98 and no I have not seen a lot of change since 1999. Our printers are faster, our Ethernet is faster, my internet connection is faster but the software and my computer are fundamentally the same. I am using a different browser but the differences are minor.

    There hasn’t been a fundamental change in the nature of personal computers since the Macintosh introduced mice and graphical interfaces to personal computers in 1984 – oh wait, that is right around the time software patents were first recognized: coincidence, I think not.

    The biggest differences that I am aware of in computers over that time period are: (1) the increase in the clock speed of CPUs (2) the virtually complete conquest of super computer installations by various forms of Beowulf cluster (outside of the reach of patents).

    And I am perfectly prepared to believe that 90% of semiconductor patents are not novel when I consider some of the outright c**p that slips through.

    A large number of narrow patents can be every bit as restrictive as one big patent and harder to deal with. If there is one big patent you have one person to negotiate with. If there are twenty small patents you may have to negotiate with a dozen patent holders each holder convinced that his patent is the key to the whole project.

    It is no answer to say that a patent does not necessarily grant a monopoly because someone else may already have been granted a broader over-riding monopoly.

    I have some understanding of the frailties of the judicial process but the test for obviousness is up against a constitutional limit and if the rules are to be simplified so that they can be administered by technically illiterate lawyers and judges than the rules have to err on the side of freedom which means erring on the side of denying patents.

    Pick whatever simple test you like so long as it systematically errs on the side of freedom.

    The problem the patent industry has (consisting basically of lawyers and judges) is that one or two more NTP v. RIM cases and the lawmakers are going to do a root and branch reform. The system can reform itself or it can be swept away.

  15. “A patent confers no right to the patentee to practice the invention. A patentee may be prevented from practicing the patentee’s invention by an earlier, broader patent.”

    Ouch, that’s a buzzkill. You aren’t going to spur much innovation with talk like that.

  16. A snippet from your 7 year old (dated 11/99) academic research paper:

    “Yet, interestingly, some of the most innovative industries today—software, computers, and semiconductors—have historically had weak patent protection and have experienced rapid imitation of their products.”

    The footnote that accompanied that statement is the following:
    Software was routinely considered excluded from patent protection until court decisions in the late 80′s. Semiconductor and computer patent enforcement was very uneven until the organization of the Federal Circuit court in 1982. Both areas contend with substantial problems of prior art [Aharonian, OTA, 1992], and some experts contend that as many as 90% of semiconductor patents are not truly novel and therefore invalid [Taylor and Silberston, 1973]. These problems make consistent enforcement difficult.

    Seven years later and semiconductor patents, as well as software patents, are probably some of the most widely patented subject matter at the USPTO. I don’t know about you, but if you compare today’s computer and the software running on that computer to a computer and software running back in 1999, I would imagine that you would find a HUGE difference. So much for all these patents being a hindrance and not promoting innovation.

    BTW: You have to admit the authors of that article had chutzpah in citing to experts that believe “as many as 90% of semiconductor patents are not truly novel.” That is a pretty harsh (and very likely very unwarranted) criticism of the Patent Office.

    “Tom argues that patents are a way of breaking the monopolies of the large companies. If IBM is filing more software patents than anyone else then patents are more likely to allow the established players to lock out new entrants.”

    Another person that doesn’t understand the difference between narrow patents, broad patents, how broad patents can preclude the practice of narrow patents on the same subject matter, etc., etc., etc. Look at the type of patents for which many of the big companies file … most of them are fairly narrow and iterative. In contrast, the patent application filed by some of the smaller entities are much more likely to be broader and innovative. After practicing many, many years and working with both types of clients, that has been my observation. As such, regardless of the numbers of the patents filed by the big companies, the most important factor is the breadth of the patent claims, such that 1 broad claim in 1 patent may be worth more than 20,000 narrow claims in 1,000 patents.

    “As a general matter arguments that this or that test would be difficult to administer are not persuasive. The goal here is the greater good; the mechanism is restricting personal freedom by granting monopolies: judicial convenience should count for very little.”

    Let us just take a step back from reality and enter the ivory tower where all dispute can be easily resolved. If a fair test cannot be administered, then who is going to push for that test? In fact, of all the proposals I’ve heard being bandied about, that proposal hasn’t even made the “pie in the sky” list. Despite what you think about attorneys and the law, simplicity in administrating the law is valued. Unless you can come up with some fair way of administering this type of test, then you have ZERO chance of persuading anybody of championing this test before Congress, which is only entity capable of introducing this type of test.

    One last thing, all laws restrict personal freedom, so get used to it. This country happens to be based upon thousands and thousands laws that tell people what they cannot do. Also, you perpetuate the common myth that patents grant monopolies. A patents confers no right to the patentee to practice the invention. A patentee may be prevented from practicing the patentee’s invention by an earlier, broader patent. As such, how can the patentee have a monopoly if the patentee isn’t even necessarily permitted to practice the invention.

  17. pds said

    “You facts are woefully lacking. Do you know who is the top filer for patents each year in the U.S.? The answer is IBM. … a very large percentage of the patents IBM has filed and granted are related to software. There are software patents aplenty, so your assertion that “the software industry was innovating magnificently without patenting” needs to be corrected.”

    This is a non sequitor. The fact that IBM is now filing lots of patents does not negate the assertion that the software industry was innovating magnificently before patents.

    There is academic research that permitting software patents did not promote innovation:

    link to researchoninnovation.org

    Tom argues that patents are a way of breaking the monopolies of the large companies. If IBM is filing more software patents than anyone else then patents are more likely to allow the established players to lock out new entrants.

    As a general matter arguments that this or that test would be difficult to administer are not persuasive. The goal here is the greater good; the mechanism is restricting personal freedom by granting monopolies: judicial convenience should count for very little.

  18. “”Conversely, an idea may be hit upon by many people in close succession precisely because it’s important and basic. Does that mean that the basic idea should not be allowed a patent?”

    The answer to which in most cases is “yes”, the patent should be denied. The fact that one inventor was riding a faster horse and came up with the idea minutes, days or weeks ahead of the rest does not justify a 20 year monopoly.”

    How long should an inventor have to wait to see if anybody came up with the invention before being given their patent? 1 month? 1 year? 3 years? 10 years?

    What happens when an invention is made and it is presented in a paper at some conference, and the patent application is filed within 1 year of the public disclosure? Should someone, after reading the paper, later be able to say that they would have come up with the invention?

    Different hypothetical, but same question, what happens if a patent application was file one day and the next day the subject of the patent application was disclosed? Should someone, after hearing about the disclosure, be able to file something with the USPTO the day after saying that they would have also come up with the invention?

    Your suggestion, which is that patents shouldn’t be rewarded to invention that would have been developed independently by several inventors, although a nice concept, would be nearly impossible to implement.

    As an aside, besides conforming with the rest of the world, the USPTO is trying to go to a first-to-file regime (instead of the current first-to-invent regime) mostly to get rid of interferences. If the USPTO wants to get out of the interference business, I would imagine that they wouldn’t want to be forced to determine whether an invention has been independently developed by multiple inventors.

  19. “Trite question that is beginning to lose its credibility: what incentive would there be to inovate if Microsoft, SAP, Oracle, etc. can easily patent around the small company’s ingenuity because it’s not obvious for PHOSITA to make a minor change of (you fill in the blank)?”

    You are mixing apples and kumquats, and your use of the phraseology indicates you really are getting over your head. One does not “patent around” anything … one designs around a patent. Also, broad claims are what prevents a big company (or any size company for that matter) design around a particular patent, which has nothing to do with whether a particular invention is obvious or not.

    “This boogeyman hindsight is killing off not only PHOSITA, but pure common sense of a homo sapien.”
    If you think “hindsight” is a boogeyman, then there probably isn’t much of a chance to reach you. From where I sit, the use of hindsight is ubiquitous.

    “I agree with the above assessment about properly allocating the burden of producing evidence on the obviousness of an invention — that is the proper question and the central issue to all this obviousness debate. If you want to be a potential millionaire, prove your worth!”

    Please explain what evidence you would use to prove that someone wouldn’t have considered the invention was obvious. I’m glad you used the term “potential” millionaire, but I would imagine that 95%+ of issued patents are worth less than $1M.

    “BTW, the software industry was innovating magnificently without patenting. If you really want to correlate innovation with patents, I suggest you give us some cold hard facts and explain why the american auto industry, mired in patents, is stagnant in inovation in comparison to the software industry.”

    You facts are woefully lacking. Do you know who is the top filer for patents each year in the U.S.? The answer is IBM. Granted, IBM has lots of non-software technology, but a very large percentage of the patents IBM has filed and granted are related to software. There are software patents aplenty, so your assertion that “the software industry was innovating magnificently without patenting” needs to be corrected.

    You state that the American auto industry is stagnant in innovation, yet you produce no facts for support. You probably have no idea as to the complexity and extent of the technology that goes into today’s automobiles. Compare an automobile of today with one from 20 years ago, and you’ll see vast differences.

  20. Tom

    You feel that making life difficult for Microsoft is a good thing in and of itself – I disagree.

    In any event, a patentable incremental innovation in word processing or spreadsheets is not going to break Microsofts market dominance – it will simply put the patent holder in a position to extract money from Microsoft and strengthen Microsoft’s hold on the market.

    You ask:

    “Conversely, an idea may be hit upon by many people in close succession precisely because it’s important and basic. Does that mean that the basic idea should not be allowed a patent?”

    The answer to which in most cases is “yes”, the patent should be denied. The fact that one inventor was riding a faster horse and came up with the idea minutes, days or weeks ahead of the rest does not justify a 20 year monopoly. The inducement of a possible monopoly was not necesary to bring forth the discovery.

    You can raise in terrorem arguments about Microsoft dominating certain narrow categories of software indefinitel but history has taught us that technical supremacy is a fleeting thing. The same things were being said about IBM in 1976 and their supremacy was crumbling by 1991.

  21. Joe Smith,

    I’ll try to respond to more of your points later, but I did want to note one thing you argue that relates pretty directly to the issue of obviousness.

    You assert that the relevant notion of obviousness should be tied to the issue of whether the idea has been independently invented by a number of people in a relatively short time period. I think this likely is just the wrong approach.

    One reason that something may not be conceived but infrequently is that it is of only small importance. Conversely, an idea may be hit upon by many people in close succession precisely because it’s important and basic. Does that mean that the basic idea should not be allowed a patent? If, again, one is basically concerned with the good of society, that hardly follows. It may well be exactly the basic ideas that are most important to keep out of the hands of the big companies, and to allow to be developed by another company with another, distinctive point of view. In the larger scheme of things, it may not be crucially important which of the early inventors gets the claim, only that it be staked out independently of the pre-existing powers that be. I simply presume here that decentralization is an overriding good in promoting innovation; this is why monopolies are generally regarded as damaging to innovation.

    Indeed, one of the great criticisms of the patent system is that it offers “monopolies” to the patent holder. But what people fail to acknowledge is that economic forces by themselves grant monopolies to certain companies, most especially, because of network effects, in the software industry (Bill Gates once, before his attorneys shut him up, opined that Microsoft formed a “natural monopoly”). There is simply no other mechanism than patents to counteract this effect. And the “natural monopolies” in software have no time limit of 20 years — they might go on for generations, for all we know.

    Microsoft has essentially owned the desktop OS, word processing, spreadsheets, and presentation software since the very early 90s, and looks to own them for decades more. Are we to believe that not one of these areas is capable of real innovation from now until eternity? The car has been around for over a century — did innovation stop after a decade or two?

  22. pds:

    “… what incentive would a small software company have to innovate if patent protection cannot be obtained for a new idea, and Microsoft, SAP, Oracle, etc. can come right in and exploit the idea?”

    Trite question that is beginning to lose its credibility: what incentive would there be to inovate if Microsoft, SAP, Oracle, etc. can easily patent around the small company’s ingenuity because it’s not obvious for PHOSITA to make a minor change of (you fill in the blank)? This boogeyman hindsight is killing off not only PHOSITA, but pure common sense of a homo sapien. I agree with the above assessment about properly allocating the burden of producing evidence on the obviousness of an invention — that is the proper question and the central issue to all this obviousness debate. If you want to be a potential millionaire, prove your worth!

    BTW, the software industry was innovating magnificently without patenting. If you really want to correlate innovation with patents, I suggest you give us some cold hard facts and explain why the american auto industry, mired in patents, is stagnant in inovation in comparison to the software industry.

  23. “MM could explain how focusing the attention of the public education system will improve the quality of innovation? ”

    People who can’t read, write or do basic math are less likely to invent anything worth patenting in 2006 relative to those who can read, write and do basic math.

    Am I surprised that you would seriously doubt this? No. Say, the holidays are coming up. Is there a charity for small inventors where I can donate some canned food?

  24. MM:

    Always has a way with words …

    “The failure for the government to award deserving systems some special award is a common everday occurence. Nobody gives a hoot. It’s pretty much expected.”

    We shouldn’t care that someone spent anywhere between 5K-20K (if not much more) to protect their invention and is denied a patent even though they may be deserving? Perhaps the person could have gotten the application allowed if they had a better attorney and/or appealed the case. However, most inventors and even companies are not an endless fountain of money. Eventually, the money runs out.

    “An error in (2) means that an **undeserving applicant** gets a 20 year monopoly with which he/she can theoretically cripple an industry and affect the lives of millions of people.”

    I’m glad that MM threw in the “theoretically.” Otherwise, his assertions would have been found wanting for lack of factual support. He might as well argue that the drug to cure cancer, AIDS, and the common cold might be held hostage by a patent owner demanding too much money for this life-saving drug. It would make not agreeing with his arguments more scarier.

    MM presumes that the applicant is **undeserving**, but who actually makes that determination? The person arguing that the patent is obvious will always argue that the patent owner is undeserving, regardless of the facts. This undeserving patentee is MM’s straw man.

    However, even if the patentee is undeserving, any patent that can “theoretically cripple an industry and affect the lives of millions of people,” will provide people sufficient incentive to search for better prior art in invalidating the patent. As evident from reading the posts on this board from the last couple of weeks with regard to a different subject (i.e., non-patent literature), it is pretty well accepted that most examiners do not use all the tools available to them in search for prior art. As a result, if the patentee is undeserving, it is very possible that invalidating prior art is out there that hasn’t been found.

    If a million people are affected, and each chip in $1, then these people are able to perform $1M search that should find any and all potential prior art. Certainly, the $1M search will be magnitudes better than what the patent office could perform. If, after spending $1M, it is determined that better prior art cannot be found, then perhaps the patentee was not undeserving. However, if better prior art is found, isn’t it better to have the people affected by the “underserved” patent pay a small amount to have it invalidated than it is to stifle innovation by erring on the side of preventing good patents from issuing?

    “I think the American public believes that (2) is clearly worse. I agree. The bogeymen floated by those who think that (1) is worse is that “people will stop filing patents” and “innovation will be stifled.”"
    I wonder if MM could be kind enough to provide evidence for his first statement. Of course, anybody could point out to the vast desire (particularly in the internet community) to eliminate the copyright protection enjoyed the authors of the many songs and movies. The fact that most people enjoy not having to pay for something that someone came up with, be it a movie, a song, a patented product, a script for a play, etc., to me, isn’t a good reason to err on the side of not granting patents.

    “Poppycock. If the failure to innovate is such a terrible problem, I suggest those folks who are so worried focus their attention on the public education system in this country rather than on the patent office.”

    Nice use of misdirection … any magician and/or politician would be proud of that statement. Perhaps MM could explain how focusing the attention of the public education system will improve the quality of innovation? Can innovation be book learned? Can the public (or private) school system teach people how to invent better?

    Invention, and the disclosure thereof, is all about incentive. Without incentives, many people will stop innovating and of those that innovate, many will stop disclosing. Making it harder to obtain a patent achieves reduces the incentives. As noted by an earlier poster, what incentive would a small software company have to innovate if patent protection cannot be obtained for a new idea, and Microsoft, SAP, Oracle, etc. can come right in and exploit the idea?

  25. Tom

    You pose a false choice between the status quo and anarchy.

    As a former computer programmer I am extremely sceptical about software patents. I am not however opposed in principle.

    The test for obviousness goes to the heart of the tradeoff inherent in a patent system: by granting a patent we may bring forward a particular innovation more quickly but there will be costs (transaction costs; deadweight costs resulting from the restrictions on the use of the patented invention; possible delay of follow-on developments).

    The problem with obviousness is not limited to software but it is accute in software because of: the large number of potential innovators and the relatively low start-up costs; the complex nature of software with each piece of non-trivial software incorporating hundreds of separate innovations; and the continuing rapid pace of innovation.

    You ask:

    “You … can NEVER answer the question of how to protect ANY small innovator from being put out of business by large companies. I’m not sure that I’ve EVER heard an honest response to this question from any of the dyed-in-the-wool opponents of software patents.”

    Protecting every small would be innovator is not the goal. A patent system should not be designed to favor small innovators at the expense of the greater social good. An innovator who brings something truly important to the marketplace can succeed and can displace the established players (and may well qualify for a patent). The whole history of computers and software is that yesterday’s dominant player is tomorrow’s also ran. Google is a recent example that a small innovator can succeed in the face of established players and a determined effort by Microsoft to crush them.

    You suggest:

    “In fact, most people invest their time, energy, and money in endeavors when they envision a personal payoff. If they can’t see anything like a realistic path whereby their efforts are repaid, they will simply move on to another potential endeavor.”

    Quite so. From a broad social perspective, I do not want a lot of small innovators spending time trying to develop incremental improvements to word processors. Their skills would be better dedicated to less crowded fields of endeavor.

    You conclude:

    “The result? Hosts of innovations simply may never come to pass.”

    Software patents have been recognized for about twenty years. The corollary of your suggestion is that the patent system has brought forth “hosts” of innovations from small companies during that period. Can you point to a software patent which actually illustrates the points that you make: an innovation worthy of a patent which would not have been developed independently within a short period of time but for patent protection and for which patent protection was effective? (Please don’t refer to the public key cryptography patent or the method for solving linear programming problems – they were non-obvious and major mathematical feats but raise a completely different set of questions about the proper scope of the patent system.)

    I apologize to all, and especially to Dennis, for being long winded.

  26. … (2) an erroneous conclusion that an actually obvious invention is non-obvious.

    Obvious/not-obvious to whom and under what circumstances?

    Example:
    The clock is ticking.
    The answer to this riddle is immediately obvious to 8 out of 10 kindergardners.
    You are much smarter & therefore can “obviously” answer faster. Go.

    What is greater than God, More evil than the Devil, The poor have it, The rich want it, And if you eat it you die?

    Well? Well? Time’s up.

  27. Shubha

    “The point of the second set of questions is to emphasize that one cannot answer the question of what is nonobvious, least of all a legal test for answering that question, without a better sense of the policies for why we have the requirement. At the heart of nonobviousness is the notion that just because an invention is new does not mean it is deserving of patent. So the question becomes does an inventor deserve a patent for (1) ingenuity, (2) hard word, or (3) solving a problem that benefits society, however, measured. The current TSM approach adopts an administratively workable solution that does not address the reason for having a nonobviousness requirement in the first place.”

    Word.

  28. There are two types of potential error in each non-obvious decision: (1) an erroneous conclusion that an actually non-obvious invention is obvious, and (2) an erroneous conclusion that an actually obvious invention is non-obvious. Each is problematic for achieving the goals of patent law; it is not clear that one is worse.

    Let’s see. An error in (1) means that a deserving applicant is denied a very very special reward by the government. The failure for the government to award deserving systems some special award is a common everday occurence. Nobody gives a hoot. It’s pretty much expected.

    An error in (2) means that an **undeserving applicant** gets a 20 year monopoly with which he/she can theoretically cripple an industry and affect the lives of millions of people.

    I think the American public believes that (2) is clearly worse. I agree. The bogeymen floated by those who think that (1) is worse is that “people will stop filing patents” and “innovation will be stifled.”

    Poppycock. If the failure to innovate is such a terrible problem, I suggest those folks who are so worried focus their attention on the public education system in this country rather than on the patent office.

  29. Greg

    “The results were that the suggestion test did not mitigate the hindsight bias for the inventions tested, but also that it does not lead to excess conclusions that an invention is non-obvious (contrary to the central charge of its critics).”

    What do you mean when you say “excess”?

  30. Thank you, Dennis, for the post.

    A couple posters mentioned the EPO’s problem and solution approach. I propose a method for bifurcating the non-obvious issue before juries and at the PTO in the articles that is somewhat similar. Very briefly, the idea is that, to the extent possible, the decision-maker would not know of the invention, but only the problem, prior art, and level of ordinary skill in the art, when they judge non-obviousness. I plan to conduct additional studies to test the EPO approach and my bifurcation proposal. (My guess is that the EPO approach still suffers hindsight bias because the decision-maker knows of the invention, but perhaps it is somewhat reduced).

    Concerning the suggestion test, some of my studies specifically tested its impact on the hindsight bias (see the Patently Non-Obvious II article). The results were that the suggestion test did not mitigate the hindsight bias for the inventions tested, but also that it does not lead to excess conclusions that an invention is non-obvious (contrary to the central charge of its critics). Based on the studies, I support the suggestion test. There are two types of potential error in each non-obvious decision: (1) an erroneous conclusion that an actually non-obvious invention is obvious, and (2) an erroneous conclusion that an actually obvious invention is non-obvious. Each is problematic for achieving the goals of patent law; it is not clear that one is worse. The hindsight bias studies teach that the former errors are expected to be much more prevalent than the latter. In addition, some analyses by Professor Christopher Cotropia identify a category of invention, which would not have been picked-up in my studies, for which the suggestion test potentially is beneficial: complex technology inventions. My studies purposely focused on simple technologies (for reasons discussed in the studies). The suggestion test certainly is not ideal, but it does not appear to cause the harms hypothesized by its critics, it is the only doctrinal fix to the hindsight problem identified, it potentially reduces the bias for complex technology inventions, and the problem the test confronts (erroneous decisions that an invention is obvious in hindsight) is known and significant, while the problem the test is alleged to create (improper application of the non-obvious standard) appears, at a minimum, to be overstated.

  31. Joe Smith,

    You compare patenting to a land rush. In certain respects, I think that may be a fair comparison. While you imagine that a land rush amounts simply to “squatting”, and waiting for someone else to develop the property, that really does miss the point of governmentally sponsored land rushes in the past.

    To begin with, both patents and land rushes do, of course, have constraints that the property in question be developed, at least to a degree. One can dispute, in either case I’d guess, whether an adequate requirement for working the property is really imposed. But what is very hard to dispute is that those who profit the greatest are those who work the property most assiduously and effectively. So the issue is, how many land rush winners are simply exploiting the rules to bring in a windfall with the bare minimum of work, and how many are doing their utmost to make their property productive? On balance, is society better or worse off?

    The answer to the last question is particularly difficult, because one must deal with the alternative: namely, if there is no governmentally sponsored land rush, then (under the analogy) only large companies will be in a good position to take over the land, because they can easily push out any small player and take over their land and crops. Can it be good for society to enable the large companies to squeeze out small innovative individuals and companies virtually at will?

    You see, the last point brings up the fatal flaw of your argument, and the argument of all the anti software patent crowd. You focus on the inequities of the land rush, but can NEVER answer the question of how to protect ANY small innovator from being put out of business by large companies. I’m not sure that I’ve EVER heard an honest response to this question from any of the dyed-in-the-wool opponents of software patents.

    Returning to the inequities of the land rush, it’s certainly true that the rules whereby one person and not another gets deed to a piece of land isn’t always “fair”, in some cosmic sense of the word. A rich spot of land might be first staked by a slacker, or someone otherwise less worthy than another person who also angled for the land. But in the larger scheme of things, such inequities are tolerated because, as a system, and in aggregate, the effects for society are better with the system in place than without any such system at all.

    Likewise, it may seem terribly unfair that only one inventor should get claim to an idea via patents, when there are other inventors who hit on the idea independently. But this local unfairness pales by comparison to the global distortions introduced when there are no protections whatever for new ideas. The alternative, again, is that large companies simply grab up everything they desire.

    Now perhaps you think that that is in fact the right solution: just let the small companies do the innovation, let large companies pluck the ideas they want, and society is better off. But this rests on the assumption that small innovators and their investors are, and will remain, delusional. In fact, most people invest their time, energy, and money in endeavors when they envision a personal payoff. If they can’t see anything like a realistic path whereby their efforts are repaid, they will simply move on to another potential endeavor. The result? Hosts of innovations simply may never come to pass.

  32. This link takes you to a list of the Gregory Mandel papers at SSRN. Both papers can be found there: link to papers.ssrn.com

    If that link does not work, click on the “here” link in this blog, then click on the Gregory Mandel link near the top of the page in SSRN and that will take you to the list of his papers.

  33. This is a great discussion, but two points come through.

    First, it is hard to deny that hindsight bias of some degree exists in the nonobviousness analysis, just as it exists in other areas of the law (the negligence analysis, for a non-patent example). The problem is what to do about it. Simply shifting the time frame is a nice linguistic solution, but if the bias is persistent, the way the time frame is constructed would also be affected by the bias. Alternatively, one could seek objective evidence of nonobviousness, which is the tack of the TSM approach. While this approach seems to safely shift the time frame, there is still the problem of making the proper inference from silence when there is no teaching, suggestion, or motivation. Some things may be too obvious to state.

    The latter is not a problem of hindsight bias, but indicates the problem of properly allocating the burden of producing evidence on the obviousness of an invention. Who has that burden? Society as a whole that may be adversely affected by an improperly granted patent? The inventor who is denied a deserving patent? The agent who is supposed to administer the process while ensuring that the statutory requirements are met?

    The point of the second set of questions is to emphasize that one cannot answer the question of what is nonobvious, least of all a legal test for answering that question, without a better sense of the policies for why we have the requirement. At the heart of nonobviousness is the notion that just because an invention is new does not mean it is deserving of patent. So the question becomes does an inventor deserve a patent for (1) ingenuity, (2) hard word, or (3) solving a problem that benefits society, however, measured. The current TSM approach adopts an administratively workable solution that does not address the reason for having a nonobviousness requirement in the first place. The Supreme Court has done a sem-decent job in addressing questions of patent policy in eBay and Independent Ink. It will be interesting to see if, and how, they address the policies of nonobviousness, perhaps even more interesting than what they do with the TSM approach.

  34. Joe Smith laments, “It seems that there is an idea in software development that if you scratch out an idea on an envelope, write a hundred lines of code and file a patent someone else owes you a hundred million dollars.”

    I think you give too much credit to them: some don’t write any code.

  35. Tom on your doubt about RIM NTP

    “Campana developed pager inventions in the early 1980s and started working on pagers with text messages around 1985. The trial court judge noted: “It is apparent that RIM developed and conceived its BlackBerry products entirely independent of the Campana patents.” NTP, Inc. v. Research in
    Motion, Ltd., 270 F. Supp. 2d 751, 755 (D. Va. 2003).”

    taken from link to students.law.umich.edu

  36. Let’s go one step further with the pictorial problem:

    1. Take the Genie’s lamp out of the room. (That was another non-obvious solution –rubbing the lamp and wishing. I liked the shard one also for the mirror-in room problem.).

    2. Pose the above picture question again (guy in a room with two hanging strings & a chair, no lamp) to two groups, one that had seen the swinging lamp solution and one that had not.

    How obvious is the answer now?

    Two answers written in reverse direction below:
    ffo tleb ruoy ekat, ffo eohs ruoy ekat

  37. Tom

    Do not misunderstand me. I believe that there should be a patent law. I just think that the current test for “obviousness” sets the bar too low (in fact unconstitutionally low) and I think that we need to do serious rethink on the remedies for infringement.

    The current system is like a land rush where a few people are investing in faster horses so they can get out there and lay claim to a piece of what should be a public intellectual domain and then they wait for someone else to make their claim valuable.

    If you look at Forgent’s claims against jpeg, Silicon Graphics’s claims to the use of floating point arithmetic in graphics cards, or Eolas’s claim to plug-ins: these are all patents which should never have been granted. It seems that there is an idea in software development that if you scratch out an idea on an envelope, write a hundred lines of code and file a patent someone else owes you a hundred million dollars. The fact is that in most cases the value is created not by that idea but in developing a complete product and reaching customers with that product.

    There is lots of incentive to innovate in the broader market. That there is little innovation in word processors or spreadsheets is not important – both are mature products which are already grossly overloaded with features. Would be innovators should look to greener pastures.

    What your comments about Microsoft show is that Microsoft’s real strength has been not in technical innovation (afterall they bought DOS from someone else and most of their software is badly written) but in effective marketing.

  38. Here’s the problem with this “test” and it’s relationship to the patent law issue: it’s not the patent law situation. Let me explain.

    This is supposed to measure whether people think something is easier after someone gives them a plausible solution. I think that most people would agree the problem isn’t that difficult to solve and therefore most of the commenters see it as “obvious.” Indeed, there are a number of “solutions”: tie the candlestick & swing it, use the guy’s belt to make one longer, ditto for the shirt, pants or shoes laces, etc. I’m not sure who said that the solution “isn’t obvious,” but when pressed for awhile to think about it, they would probably generate some answer. The question though was “is it MORE obvious” than before you’re given a solution.

    The problem assumes away the more significant patent law issue: what made someone want to tie the two strings together. A patent law scenario might look like this: “a guy in a room with two strings hanging down, a candlestick and a chair.” Then “swinging a candlestick on a rope in order to grab the other to tie them together” is far less obvious.

    Indeed, the problem as stated looks more like the test for obviousness perhaps ought to be: given the problem (tying two strings together), would a PHOSITA have figured out the problem this way? Of course, this begs the question “was the problem itself obvious?”

  39. “It would be good to see just how many people hit upon the solution who have not been tainted by previous exposure to the problem.”

    I feel deeply sorrow for adults who have not encountered the generic mundane problem here (“I can’t reach!”) at some point in their lives.

  40. I should think that if a sizable number of BU law students, presumably pretty bright, did NOT come up with the solution, it would bepretty hard to assert that the solution was “obvious”.

    Are you serious?

  41. Geez…this was not an “empirical study”. It was a simple question to see if there was a tendency to for those who were given the answer to consider the solution more apparent than those who were not given the answer.

    Only attorneys can take a simple test for psychological predisposition and turn it into a call for empirical data and a debate about level of skill and obviousness. Clearly, this was not the purpose of the test.

  42. Dennis hasn’t yet divulged a critical piece of information regarding how his two groups in the seminar responded. Namely how many of the group who were NOT told the solution INDEPENDENTLY came up with it?

    I’d guess that it would be hard for a member of this group who failed to hit upon the solution to declare it obvious. And for outside observers like us, I should think that if a sizable number of BU law students, presumably pretty bright, did NOT come up with the solution, it would bepretty hard to assert that the solution was “obvious”.

    While I myself came up with the solution fairly quickly, I know at least a sufficient reason for my having done so: I have seen the problem before, a great many years ago. Likely, a lot of people who went through a period in their life when they were interested in puzzles have seen this one.

    It would be good to see just how many people hit upon the solution who have not been tainted by previous exposure to the problem. THAT would be genuine empirical evidence of “obviousness”, and ONLY that — bland assertions that no “science or math undergraduate with a B average or greater could find the solution non-obvious” are pure speculation.

  43. I don’t think Dennis’ hypothetical attempts to take into account ordinary skill in the art, except possibly implicitly.

    It is just used to demonstrate what should be painfully “obvious” to all of us and that is that the hindsight bias is extremely strong and must be filtered out somehow in the obviousness/inventive step inquiry.

    I’m not sold on the EP approach either. In practice, it has seemed to me to be very little different from the US standard except that the European Examiners tend to be more technically competent and willing to stand on a level of ordinary skill in the art to bridge the multiple reference gaps.

  44. Wow. I hadn’t looked at this “problem” until after my previous comment. By chance did you divide the responders in your class by background and undergraduate GPA?

    It’s impossible for me to imagine that a science or math undergraduate with a B average or greater could find the solution non-obvious.

    That test might be useful for determining who doesn’t get to sit next to the emergency exit door on the airplane.

    Here’s an interesting twist, though: give half the class that problem but also mention that “A paper in the room states that a wooden ruler was used to bring two wires together in a prison cell in 1962.”

    A reasonable person would find that the paper makes the solution obvious. A diehard “pro-patent” freak would cry, “Improper hindsight! Hindsight! This isn’t a prison cell! This is a candlestick, not a ruler! They are different! Where’s the suggestion??? Improper hindsight!!! Improper hindsight!!! I’m going to appeeaaaaaaallllll!!!!!”

  45. Well, for those who demand “hard” evidence for this sort of thing, I guess Dennis has provided some. As others have noted, I do not believe that anyone seriously doubted whether hindsight bias existed. I’d file this data next to the data on the non-efficacy of distant prayer.

    I will also second the notion that the question is whether the problems caused by a strict/explicit TSM test outweight its alleged ability to minimize “hindsight”.

    And I agree EPO has the superior approach.

  46. Joe Smith,

    In answer to your points:

    1. You miss the point. A large company is NOT going to be able to copy an idea that never makes it into the market because no one has an incentive to develop it and promote it. Seriously, how much innovation has Microsoft added to word processing, or spreadsheets, or presentation software, or you-name-it since they took over those markets for good? What is even their incentive to innovate on a grand scale in those areas where they already dominate because of network effects? All they need is just enough innnovation to assure that people upgrade — and that is generally pretty trivial to achieve, perhaps requiring as little as introducing new file formats that old programs can’t read.

    2. The question of whether software is just an “idea” is pretty uninteresting from an economic point of view. Clearly software patents are intended in any case to be applied to software as realized in the physical world, in devices that do undeniably useful things.

    3. Certainly a lot of important and useful software ideas were introduced before patents could protect them (though I dispute that only less interesting software is now being developed). And no one fully understood early on just what would happen to their ideas in the industry as time progressed, and the incentive for innovation seemed very much alive: people believed that if they built a better mousetrap, they would win in the marketplace. But now time has progressed. what do see as a result? That all those early ideas were absorbed into essentially one company, Microsoft, and precisely because those ideas were NOT protected by anything more than copyright and trade secrets, which were grossly inadequate to the task.

    4. NTP may or may not have a causal link to the progress of wireless email. I gather the people behind NTP were very early on advocating these ideas, so I’m not sure how easily one can make a claim for fully independent discovery by RIM. But the more important point is that independent discovery does NOT imply that it would be good AS A SYSTEM to deny patents for software. Even supposing that the economy would be better off if NTP in particular did not get any money for its patent, one can’t pretend that denying all patents gets rid ONLY of the “bad” cases of patent enforcement. What you and your friends NEVER do is present anything like a plausible case in which something other than patents can protect small innovative companies from having their ideas copied wholesale by a company like Microsoft. $600M seems like a big number, but it pales by comparison to the roughly $14B in profits that Microsoft makes every year based on technology that was, indisputably, invented, developed and promoted originally by other companies. I’d like to know how diverting all those vast monies into a demonstrably UNinnovative company, and parking them there for as long as network effects can keep them there (which may be nearly forever), can possibly be the best way in which to promote innovation. Yes, Microsoft wrote software independently to perform these functions, but in the larger scheme of things, of what significance is that? Just about ANY competent company could write software to do, say, word processing — and many companies have. But who has any incentive to innovate in this area, or any other area Microsoft now dominates? You provide potential innovators with NOTHING as an incentive.

    The overarching point is, if you can’t provide innovators with an incentive, innovation dries up. That’s why patents were introduced in the first place, and, again, you offer innovators nothing else that might possibly do the job.

  47. Tom said:

    “Look, one of the deep problems with failing to protect inventions is that there is generally NOTHING to stop large firms from simply ripping off smaller, innovative firms.”

    and went on to talk about software. To which I reply as follows:

    1. If the inventions were going to be inevitable then the social interest may be in the ideas being copied by the large company.
    2. There is no natural property in an idea.
    3. The most productive period in software was before software patents were recognized. (I’ve been writing software since the early 1970s.)
    4. The problem with the current system is that we are getting squatters staking out intellectual property which should be in the public domain and extracting economic rents from those who do the actual investing to produce a marketable product. NTP contributed exactly NOTHING to the eventual success of wireless email and yet they were able to extract $600 Million from an independent inventor who was able to actually turn the idea into a valuable product – how does that make society better off? How does NTP deserve that $600 Million when they admitted that there was no causal link between Campana’s work and the BlackBerry?

  48. Joe Smith wrote, “It seems to me that you are confusing the popular meaning of ‘obvious’ with the technical meaning of ‘obvious.’”

    True. But the point about hindsight biasing opinions is still valid.

    Perhaps the question could have been phrased to ask whether they thought they would have solved that problem given a particular time period to solve it.

  49. “If independent discovery within a time period of a few years is inevitable then a patent system will do more harm than good in encouraging progress and as a result, in the US, would be unconstitutional.”

    This is just simplistic. In reality, there are a variety of ways in which progress can be encouraged. Reducing obviousness simply to the question of whether it might be possible for more than one party to come up with the invention at roughly the same time is a perversion of its real meaning, and its real consequences.

    Look, one of the deep problems with failing to protect inventions is that there is generally NOTHING to stop large firms from simply ripping off smaller, innovative firms. Of what matter is it that there may be a number of smaller firms that come to the same idea at more or less the same time? While patents may protect only one of these firms — the one granted the relevant patent — the alternative of denying patents altogether simply enables large firms to sit back, wait for others to discover and invest in new ideas, then step in and copy the idea wholesale and take over the market. In such a scenario, what incentive is there for ANY small firm to innovate and take the innovation to market? How can progress be encouraged in such a scenario?

    Please don’t tell me such a thing is not possible, that large firms will always come up with the innovations anyway. Microsoft is a notorious case in point. No one who has followed the progress of the software industry will fail to note that not a single major innovation — not one — has ever first been put forth by Microsoft. In every such case, Microsoft has been the highly successful “fast copier” — a term which only in the software industry, with its relatively poor patent protections, has been considered a good, even laudable strategy. But those who argue this can’t come up with real incentives to motivate innovation other than to pretend that somehow copyright protection and trade secrets will do the trick — despite the overwhelming evidence that these “protections” did absolutely no good in preventing Microsoft, a very late comer to many of the technologies, from taking over the markets entirely.

    The irony of course is that patents are often considered particularly damaging in the software industry. Yet it’s especially in the software industry that the power of the dominant force in a given market can simply push aside smaller innovators. What is distinctive about the software industry is how completely it is governed by so-called network effects. Whoever has the network effects going for them in a given segment of the industry can easily let others do the innovation, and then implement those innovations independently at their leisure, knowing that they’ve got the field tilted in their favor. Virtually the only way for new, and large companies to develop is to find an entirely new area of the industry where the network effects have not yet set in, and become the dominant force in that area. Thus, Microsoft dominates desktop OS and applications, Google dominates search, Ebay online auctions, Amazon bookselling, etc. Yet within these segments, how does further innovation ever happen? Apart from patents, why should anyone pursue a new idea in those fields? Having been successfully taken over by a single company, are they lost to innovation forever?

    If one is genuinely interested in “encouraging progress”, one must take into account the larger economic consequences of patents, in any industry.

  50. 1. If you have a cluttered room, the string problem is actually easier, since all you need is an object to either extend your reach or extend the length of a string. It would also illustrate the typical obviousness inquiry a little bit better. Suppose that there’s a candlestick, and also a flashlight. Someone solves the problem with the candlestick in the way indicated, and then someone else uses the flashlight in the same way, seeks a patent, and pounds the table about hindsight bias and the lack of a suggestion in the prior art that a flashlight would work just as well as a candlestick. I’m skeptical.

    2. If I recall, the obviousness question isn’t whether a law student could figure out a riddle in 5 minutes. It’s whether the solution would be apparent to someone of skill in the art, focusing on the problem to be solved. Try sending the students home for a week with the problem and see how many of them are at a loss for an answer that actually works.

    3. What is this meant to show? I think the law of obviousness acknowledges the hindsight problem. The more difficult question is how to solve it and whether the TSM test provides any help. It could be that this shows that lay opinion of the patent system may be affected by hindsight bias, but so what?

  51. “So how many years would the time period of inevitable invention have to be before a patent does less harm than good … 20+ years?”

    It would depend on the benefit from the invention being discovered earlier balanced against the deadweight loss from the time of the inevitable invention to the expiry of the patent all taking into account an appropriate discount factor. I expect that the socially optimal test for “obviousness” would fall somewhere between “highly likely that the invention would have been discovered in any event within two years” to “50/50 chance it would have been discovered within five years.”

    None of my comments apply to pharmaceuticals because for pharamceuticals we are using the patent system to really protect the investment in regulatory approval.

  52. …and if you are in the physical world

    1. Stand on the table.
    2. Lift the mirror over your head.
    3. Smash the mirror on the floor.
    4. Pick up a shard of the mirror.
    5. Slit your wrists with the shard.
    6. Wait.

    Soon you will no longer be in the room.

  53. Answer to my riddle:
    1. Look in the mirror. See what you saw.
    2. Take the saw and cut the table in half.
    3. Two halves make a whole. Climb out of the hole.

    That was obvious wasn’t it?
    Who said the riddle takes place in the physical world?

  54. Joe Smith wrote, “If independent discovery within a time period of a few years is inevitable then a patent system will do more harm than good in encouraging progress and as a result, in the US, would be unconstitutional.”

    So how many years would the time period of inevitable invention have to be before a patent does less harm than good … 20+ years?

  55. How quickly we all forget taking exams in school. I cannot begin to count the number of times the answer to a test question jumped out at me as being “obvious” once I saw the correct answer. If only this had been the case when all I had was a question and a space to place my answer.

  56. Independent making of the same invention in different places at much the same time is unremarkable, and nothing new. The former boss of the Exxon patent operation in Europe (can’t remember his name)was showing many years ago a nice powerpoint (of a real life case) with about 15 columns, each representing a patent application filed by a different petro company, all for the same invention, and all filed within a brief time period. Most inventors lost out. One came to dominate, outside USA. It was the one that was first into a Patent Office with an enabling disclosure of the commercially important subject matter. This is normal, isn’t it?

    Seems to me that a proper patent system should be founded on the idea that, when the time is ripe for a step forward, there is going to be independent conception. That might be evidence of obviousness, whereas the absence of multiple independent conceptions is evidence of non-obviousness.

    Watching all this hand-wringing on hindsight, I should like to know why the 25 year old EPO problem and solution approach, which eats hindsight for breakfast, is not getting any serious consideration in the current debate at the Supreme Court. By now it is unassailable at the EPO, and has silenced the whinging about hindsight that used to waft across from the patent litigation community on the island of GB. My suspicion is that it is because nobody involved in the debate before the Supreme Court has any real understanding of it. Shame.

  57. It seems to me that you are confusing the popular meaning of “obvious” with the technical meaning of “obvious”. It also seems to me that the CAFC has fallen into the same error. It may be that Congress made a poor choice of words in choosing “obvious” in identifying the underlying concept.

    The whole question of obviousness in the technical sense has to be tied back to the question of the likelihood of independent discovery.

    I was watching a PBS program about the development of airpower the other night and it said that radar was invented independently in EIGHT different countries during the 1930s – and every one kept it a secret and thought they were the only ones who had it.

    History teaches us when a significant invention becomes possible it will be invented independently: transistors were invented three or more times; the integrated circuit was invented at least twice; the jet engine was invented at least twice.

    If independent discovery within a time period of a few years is inevitable then a patent system will do more harm than good in encouraging progress and as a result, in the US, would be unconstitutional.

  58. Tom, your concerns actually make these results even more powerful. I used a problem whose solution seems obvious to many people. Despite that, my students found that the solution was much more obvious once I told them how the guy solved the problem.

  59. Dennis said:
    “The more difficult a problem, the more possibility there is for hindsight bias.”

    I think the question can be better stated: “For a solution that is not obvious, how likely is hindsight bias?”

    Where the problem solution is obvious (as if anyone could ever agree on that!), I believe the results are skewed because the results will include those who rightly conclude it is obvious along with those who used hindsight bias. How do you separate the two?

  60. I agree with Stepback, I think your problem is flawed. A solution is obvious if one of ordinary skill in the art can come up with the same solution without undue effort. Not being one skilled in the art of being in a room with two strings, a chair, and a lamp, I quickly came to the solution you identified. If one not skilled in the art can readily come up with the solution, then that solution must be obvious to one who is skilled in the art. I cannot imagine how someone in either group could say the solution was not obvious. By prominently showing the items necessary to solve the problem, you were setting up the problem to obtain the answer you got.

    Why not use a solution that is not so obvious? For example, using the same illustration, if the solution was to tie the person’s shoelaces to his belt and then tie that to one of the strings in order to pull the string to the other string, then that solution would not be so obvious. Even less obvious is for the person to pull his hair out, braid it into a third string, which is then tied to one of the two strings so the two strings can be pulled together. It would be interesting to see if the results change.

  61. You’re in a lit room. No windows. No doors. Just a table and a mirror. How do you get out?

    P.S. The pendulum scenario is unfair since only the references for solving the problem are shown, thus making it obvious to combine. Why not add a lot of clutter to the scene? Make it a McGiver problem. Then it is far less obvious. i.e.: You’re in a drugstore with less than 5 minutes to shop and you must gather just the right things to help you save the quicksand victim in the lot next door. The store does not have a “rope” aisle.

  62. After hearing the results of the Mandel study, it was obvious to me that he would come up with the results that he did.

    Seriously, it seems to me that the Mandel study is an inovative work of scholarship. I’m looking forward to seeing the results of Joe Miller’s research.

Comments are closed.