5 ms·

>I would actually be interested in taking this exchange to email. Sure. I would like to, but only if you treat my points charitably (you haven't :(.) - the obj

by ABCLAW 8y ago

>I would actually be interested in taking this exchange to email. Sure. I would like to, but only if you treat my points charitably (you haven't :(.) - the objective is to learn. If that's not an issue, though, would the btilly email address in your profile work? I can send a full reply there. Some brief notes to think about in the meantime: Patents have multiple social functions. Encouraging progress can be performed by achieving other related, intermediary, policy steps. For instance, if we weaken industry-peripheral patents, do we just cause firms to resort to secrecy (and is that better for society over the long run)? Are there other solutions other than 'yes patent'/'no patent' depending on the patent type? Do patent licenses required for industry-participation have better solutions than just 'no patents'? See: Standards collectives/FRAND licensing terms/etc? Obviousness is VERY difficult to judge years later after having read a patent. Much like seeing the answer to a puzzle or math problem, it's hard to go back and feign ignorance in your analysis, and once you read a lot of patents, suddenly you see the sweep of technological development very differently. Obviousness pt 2: What if patents are needed to secure the benefit from an R&D effort. Is it worthwhile to say 'we shouldn't research this [New Medicine] because it seems obvious'? How else could we reasonably finance that R&D work? Obviousness analysis is done in reference to a 'person having ordinary skill in the art', which is an incredibly flexible legal fiction. Again, do you have issues with flexibility being used to achieve policy goals you don't agree with? Note that the artificiality of this test is amplified because we analyze this person at the time the patent was filed. Next, since 2016, KSR is the leading case in the states regarding obviousness. The KSR test performs a series of comparisons between what the field knew and what typical improvements or techniques the field uses to determine if the improvement was obvious or not. This test is heavily criticized for making innovations in slow but consistently developed fields far harder to patent, despite the fact that patents in those fields are less controversial than those in rapidly iterating fields. What specific elements of our legal regime make patent trolling too easy and profitable? Is it the cost of litigation? Is it how IP rights can be assigned at all (and what are the ramifications if we prevent assignment)? Is this a quantitative problem that we can fix by tweaking the numbers a bit or a categorical one which requires fundamental changes in the system? Btw, how large do you think the patent trolling problem is, dollar wise? Is it a priority as far as reforms go? What else do you think are competing legislative/regulatory priorities? Prior art is already a bar to patentability. Re: Unclean hands. Vexatious litigant protections exist in the court system independently of patent litigation. They aren't used much. Why do you think that is the case? Software is not the only industry that interacts with the patent system, and other industries have different needs and are influenced by patents differently (which, btw, is a core flaw with PG's note). If you want to change core elements of the patent system, how do those modifications affect other industries?

btilly 8y ago
Yes, that email address will work. I'm sorry that you do not believe that I have treated your points charitably. First on encouraging progress, it is a complex set of issues. For basic research I'm in agreement with https://issues.org/perspective-is-the-bayh-dole-act-stifling-biomedical-innovation/ https://issues.org/perspective-is-the-bayh-dole-act-stifling... and the successful approach outlined at https://www.chronicle.com/article/Michael-Milkens-Attack-on/83353 https://www.chronicle.com/article/Michael-Milkens-Attack-on/... that progress is maximized when we fund lots of ideas, make them share information quickly, and let them build on each other. The incentives of patents run exactly opposite from that. If we're paying for basic research, I do not want it locked away in patents. (Obviously private research is a different story, companies need an incentive to engage in that.) Next there is the problem of pharma. The challenge here is that proving effectiveness and safety to the FDA's satisfaction is extremely expensive. Patents are their route to recouping that. The fix that I would like requires legislation, but it is to give a patent monopoly to the company that began the FDA certification process first. That would both meet the needs of medicine, and also remove the current broken incentives that make it impossible to get anyone to pay for testing of treatments that are not covered by patents. (As an example, helminthic therapy for a variety of autoimmune disorders is extremely unlikely to go anywhere because there is no patent protection under which the costs of FDA trials could be recouped.) Now on obviousness, I agree that it is hard to judge after the fact, once you have seen the idea. Secondly what is obvious very much depends on who it is who it is obvious to. With my math background, things are often obvious to me that aren't to other programmers. But which are obvious to anyone else with a similar background. Of which many exist. What is the appropriate peer group to judge, say, a contribution that I made to an advertising system? This is why I would prefer a system where the patenter has to supply evidence that it could have been discovered, but wasn't. Contrary to the KSR test, this should make it easier to get a patent in a field that is slow moving rather than fast moving. Which I think is appropriate. A field that is already seeing innovation doesn't need patents to encourage more, and will be hurt more by typical patent periods. This is reversed in slow moving fields, where long patent terms are not such a problem and which could probably use the incentive more. On patent trolling, I've seen estimates of the cost of patent trolling to the tune of around $30 billion a year (see https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2091210 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2091210 for a citation). If you assume that there are 4 million software developers in the US with an average salary of $90k/year, that's 1/12 of the total salaries of all software developers. This is a pretty big priority. Now not all of that money would have been spent on research, but it is a pretty safe guess that patent trolling does reduce month available for research by several billion dollars a year. Why is patent trolling too easy and profitable? Well because the bar to getting bad patents is too low, it is too easy to threaten people, and even if you think the patent has no merit, the risk of it being ruled valid after all are high enough that most companies roll over, and then there is the fact that you know you'll likely face the troll in courts that have historically been extremely friendly to patent trolls. East Texas and then the Federal Court. Though, luckily, the first court is unlikely to have the same clout going forward that it has historically. On unclean hands, because patent trolls register a company per patent. So odds are that the company going after you has never actually has a lawsuit go through trial. And even if you go through the effort of getting them declared a vexatious litigant, the next patent troll owned by the same umbrella company and represented by the same lawyer is unimpeded. It is theoretically possible to go after the lawyers involved for barratry, but I do not know of anyone who has done so. And as for changing core elements of the patent system and affecting other industries, I believe that the changes that I suggest would actually work out well for multiple industries. Though admittedly one of my suggestions is an entirely new type of protection to cover the needs of medical research.

by ABCLAW 8y ago

btilly 8y ago

Yes, that email address will work. I'm sorry that you do not believe that I have treated your points charitably. First on encouraging progress, it is a complex set of issues. For basic research I'm in agreement with https://issues.org/perspective-is-the-bayh-dole-act-stifling-biomedical-innovation/ https://issues.org/perspective-is-the-bayh-dole-act-stifling... and the successful approach outlined at https://www.chronicle.com/article/Michael-Milkens-Attack-on/83353 https://www.chronicle.com/article/Michael-Milkens-Attack-on/... that progress is maximized when we fund lots of ideas, make them share information quickly, and let them build on each other. The incentives of patents run exactly opposite from that. If we're paying for basic research, I do not want it locked away in patents. (Obviously private research is a different story, companies need an incentive to engage in that.) Next there is the problem of pharma. The challenge here is that proving effectiveness and safety to the FDA's satisfaction is extremely expensive. Patents are their route to recouping that. The fix that I would like requires legislation, but it is to give a patent monopoly to the company that began the FDA certification process first. That would both meet the needs of medicine, and also remove the current broken incentives that make it impossible to get anyone to pay for testing of treatments that are not covered by patents. (As an example, helminthic therapy for a variety of autoimmune disorders is extremely unlikely to go anywhere because there is no patent protection under which the costs of FDA trials could be recouped.) Now on obviousness, I agree that it is hard to judge after the fact, once you have seen the idea. Secondly what is obvious very much depends on who it is who it is obvious to. With my math background, things are often obvious to me that aren't to other programmers. But which are obvious to anyone else with a similar background. Of which many exist. What is the appropriate peer group to judge, say, a contribution that I made to an advertising system? This is why I would prefer a system where the patenter has to supply evidence that it could have been discovered, but wasn't. Contrary to the KSR test, this should make it easier to get a patent in a field that is slow moving rather than fast moving. Which I think is appropriate. A field that is already seeing innovation doesn't need patents to encourage more, and will be hurt more by typical patent periods. This is reversed in slow moving fields, where long patent terms are not such a problem and which could probably use the incentive more. On patent trolling, I've seen estimates of the cost of patent trolling to the tune of around $30 billion a year (see https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2091210 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2091210 for a citation). If you assume that there are 4 million software developers in the US with an average salary of $90k/year, that's 1/12 of the total salaries of all software developers. This is a pretty big priority. Now not all of that money would have been spent on research, but it is a pretty safe guess that patent trolling does reduce month available for research by several billion dollars a year. Why is patent trolling too easy and profitable? Well because the bar to getting bad patents is too low, it is too easy to threaten people, and even if you think the patent has no merit, the risk of it being ruled valid after all are high enough that most companies roll over, and then there is the fact that you know you'll likely face the troll in courts that have historically been extremely friendly to patent trolls. East Texas and then the Federal Court. Though, luckily, the first court is unlikely to have the same clout going forward that it has historically. On unclean hands, because patent trolls register a company per patent. So odds are that the company going after you has never actually has a lawsuit go through trial. And even if you go through the effort of getting them declared a vexatious litigant, the next patent troll owned by the same umbrella company and represented by the same lawyer is unimpeded. It is theoretically possible to go after the lawyers involved for barratry, but I do not know of anyone who has done so. And as for changing core elements of the patent system and affecting other industries, I believe that the changes that I suggest would actually work out well for multiple industries. Though admittedly one of my suggestions is an entirely new type of protection to cover the needs of medical research.