5 ms·

> dumb software patents their definition of "dumb" is probably quite different from yours (or most software engineers).

pclmulqdq 3y ago
To clarify, I think both of us agree that things like weird data structures, compression/encryption algorithms, and a few other things that are actually non-obvious hard tech probably should be patentable (although it's usually a shitty business strategy to patent these things), but that most other software (including things like Amazon's one click ordering patent) shouldn't be. I will admit that I don't know where his moral line on this is exactly, and I know he is very happy to draft patents beyond it for strategic reasons. I think many smart patent lawyers are ultimately legal realists here - they have their own philosophical views, but their job involves working in the real system.
- plank 3y ago
  >To clarify, … Not the parent, but: in that case you and I agree to disagree: any weird encryption algorithm is the usage of mathematics which should not be patentable in my opinion. The simplest solution in my mind is simply abolishing patents. Just as e.g. the COVID vaccines were developed using public funds, there is enough ‘protection’ in things like trade secrets (‘recipe of coca cola’) or copyright/-left (‘GPL’ or ‘iOS’) to keep us innovative enough.
  - pclmulqdq 3y ago
    I honestly don't think the patent office can draw the right line on what mathematical formulas are discovered vs invented, so that seems reasonable to me. I think you are looking at the abolition of patents from a very "software" perspective. Copyleft license terms have basically never been tested in court, and could just be unenforceable, but the tradition of software engineering seems to keep the peace there - GPL'ed and other open source hardware (which frequently has copyleft licensing) frequently gets ripped off without credit and nobody does anything about it. These would have bad facts in a court case, since the copyright doesn't cover the thing, just the schematics. Trade secrets also generally don't protect you when your product is easy to reverse-engineer and you have customers or competitors with a lot of money to do it. A drug is a prime example of this.
    - AlbertCory 3y ago
      Step by step: first remove software from the domain of patent law. That IS a heavy lift, but abolishing all patents is an impossible one. Then the question of "OK, what IP protections should apply to software, if any?" can be addressed. I don't claim to have the answer all worked out
      - pclmulqdq 3y ago
        That seems like a reasonable approach to me, and may come with a natural experiment built in. However, it's hard for me to suggest that an algorithm like PageRank, for example, is less inventive or risky to develop than a typical drug, which would suggest that some software ought to be available to patent.
        AlbertCory 3y ago
        As Ripley would say, nuking it from orbit is the only way to be sure. Whatever rules you attempt to come up with, the lawyers will nitpick them to death and find some way to keep junk patents alive. For years, everyone in the legal profession looked forward to the CLS Bank v. Alice case at the Supreme Court as the Final Word on software patents. That was 2014. My article: https://albertcory50.substack.com/p/lets-vote-on-it https://albertcory50.substack.com/p/lets-vote-on-it lists some of the patents that have been upheld since CLS Bank.
        pclmulqdq 3y ago
        Very interesting blog. I might be n = 1 as a software engineer who occasionally reads patents to learn how they work, but I also have some experience with the patent system and can read "patentese" better than most. They often leak more details than their papers. I think a key point here is that Alice is clear as dirt on its intentions. SCOTUS could have come up with a much cleaner test but didn't, and has had the option to clarify their thinking several times since then but has never taken it. I guess patents aren't "sexy" enough.
  - freejazz 3y ago
    I'm not sure how you think copyright would protect that which protects patents, but if your position is that patents are bad, so that in the alternative, corporations should keep everything secret forever (that's what you have to do to have a trade secret...) seems a bit absurd to me, to be perfectly honest.
- freejazz 3y ago
  I love how software engineers think they have a clear understanding of what should or shouldn't be patentable, yet cannot come up with words to describe the delineation more accurately than "good" or "bad" or "weird", or other ways of just facially repeating patent requirements like "non-obviousness", yet don't seem to grasp this is the exact challenge that the law has to deal with with regards to patents. It's not an easy challenge. It turns out, it's actually incredibly hard to come up with a law that describes what should or shouldn't be patentable, with regards to anything that has yet to be invented.
  - AlbertCory 3y ago
    You seem to think that, having mastered what the law IS, you're also an authority on what it should be. Wrong. The people who live under the law are equally well qualified to judge that.
    - freejazz 3y ago
      That's not what I think at all - I'm actually begging you to come up with a better delineation.. that's the challenge the law faces in regards to what it should be. We can't write a law that says "good patents are good, bad patents are bad" there needs to be a tangible, comprehensible delineation between good and bad patents that can be applied to things not yet invented... that's the challenge. I'm not sure how you came to your response at all, to be perfectly honest. I'm certainly not saying that lawyers are the only people who should be involved. I'm saying, if software engineers want to help... they need to actually help instead of just pretending like its trivial to delineate the good from the bad. It isn't. Really, the point I'm making is that, as experts in the field, if you cannot come up with a better delineation than simply stating "good" or "bad" - how is that you expect lawyers to do a better job? To further my point, that you responded the way you did is evidence that software engineers aren't actually engaging in the problem and are just sitting on the sidelines having tantrums because things aren't going the way they want.
      - AlbertCory 3y ago
        OK, so do I understand you correctly that you'd like to write a law that separates the good patents from the bad ones? Tell us: is this a "good patent": https://en.wikipedia.org/wiki/Enfish,_LLC_v._Microsoft_Corp https://en.wikipedia.org/wiki/Enfish,_LLC_v._Microsoft_Corp. How about this? https://patents.google.com/patent/US7818399B1/en?oq=us7%2c818%2c399+ https://patents.google.com/patent/US7818399B1/en?oq=us7%2c81... Both of these are considered valid, post-Alice.

To clarify, I think both of us agree that things like weird data structures, compression/encryption algorithms, and a few other things that are actually non-obvious hard tech probably should be patentable (although it's usually a shitty business strategy to patent these things), but that most other software (including things like Amazon's one click ordering patent) shouldn't be. I will admit that I don't know where his moral line on this is exactly, and I know he is very happy to draft patents beyond it for strategic reasons. I think many smart patent lawyers are ultimately legal realists here - they have their own philosophical views, but their job involves working in the real system.

plank 3y ago
>To clarify, … Not the parent, but: in that case you and I agree to disagree: any weird encryption algorithm is the usage of mathematics which should not be patentable in my opinion. The simplest solution in my mind is simply abolishing patents. Just as e.g. the COVID vaccines were developed using public funds, there is enough ‘protection’ in things like trade secrets (‘recipe of coca cola’) or copyright/-left (‘GPL’ or ‘iOS’) to keep us innovative enough.
- pclmulqdq 3y ago
  I honestly don't think the patent office can draw the right line on what mathematical formulas are discovered vs invented, so that seems reasonable to me. I think you are looking at the abolition of patents from a very "software" perspective. Copyleft license terms have basically never been tested in court, and could just be unenforceable, but the tradition of software engineering seems to keep the peace there - GPL'ed and other open source hardware (which frequently has copyleft licensing) frequently gets ripped off without credit and nobody does anything about it. These would have bad facts in a court case, since the copyright doesn't cover the thing, just the schematics. Trade secrets also generally don't protect you when your product is easy to reverse-engineer and you have customers or competitors with a lot of money to do it. A drug is a prime example of this.
  - AlbertCory 3y ago
    Step by step: first remove software from the domain of patent law. That IS a heavy lift, but abolishing all patents is an impossible one. Then the question of "OK, what IP protections should apply to software, if any?" can be addressed. I don't claim to have the answer all worked out
    - pclmulqdq 3y ago
      That seems like a reasonable approach to me, and may come with a natural experiment built in. However, it's hard for me to suggest that an algorithm like PageRank, for example, is less inventive or risky to develop than a typical drug, which would suggest that some software ought to be available to patent.
      - AlbertCory 3y ago
        As Ripley would say, nuking it from orbit is the only way to be sure. Whatever rules you attempt to come up with, the lawyers will nitpick them to death and find some way to keep junk patents alive. For years, everyone in the legal profession looked forward to the CLS Bank v. Alice case at the Supreme Court as the Final Word on software patents. That was 2014. My article: https://albertcory50.substack.com/p/lets-vote-on-it https://albertcory50.substack.com/p/lets-vote-on-it lists some of the patents that have been upheld since CLS Bank.
        pclmulqdq 3y ago
        Very interesting blog. I might be n = 1 as a software engineer who occasionally reads patents to learn how they work, but I also have some experience with the patent system and can read "patentese" better than most. They often leak more details than their papers. I think a key point here is that Alice is clear as dirt on its intentions. SCOTUS could have come up with a much cleaner test but didn't, and has had the option to clarify their thinking several times since then but has never taken it. I guess patents aren't "sexy" enough.
- freejazz 3y ago
  I'm not sure how you think copyright would protect that which protects patents, but if your position is that patents are bad, so that in the alternative, corporations should keep everything secret forever (that's what you have to do to have a trade secret...) seems a bit absurd to me, to be perfectly honest.
freejazz 3y ago
I love how software engineers think they have a clear understanding of what should or shouldn't be patentable, yet cannot come up with words to describe the delineation more accurately than "good" or "bad" or "weird", or other ways of just facially repeating patent requirements like "non-obviousness", yet don't seem to grasp this is the exact challenge that the law has to deal with with regards to patents. It's not an easy challenge. It turns out, it's actually incredibly hard to come up with a law that describes what should or shouldn't be patentable, with regards to anything that has yet to be invented.
- AlbertCory 3y ago
  You seem to think that, having mastered what the law IS, you're also an authority on what it should be. Wrong. The people who live under the law are equally well qualified to judge that.
  - freejazz 3y ago
    That's not what I think at all - I'm actually begging you to come up with a better delineation.. that's the challenge the law faces in regards to what it should be. We can't write a law that says "good patents are good, bad patents are bad" there needs to be a tangible, comprehensible delineation between good and bad patents that can be applied to things not yet invented... that's the challenge. I'm not sure how you came to your response at all, to be perfectly honest. I'm certainly not saying that lawyers are the only people who should be involved. I'm saying, if software engineers want to help... they need to actually help instead of just pretending like its trivial to delineate the good from the bad. It isn't. Really, the point I'm making is that, as experts in the field, if you cannot come up with a better delineation than simply stating "good" or "bad" - how is that you expect lawyers to do a better job? To further my point, that you responded the way you did is evidence that software engineers aren't actually engaging in the problem and are just sitting on the sidelines having tantrums because things aren't going the way they want.
    - AlbertCory 3y ago
      OK, so do I understand you correctly that you'd like to write a law that separates the good patents from the bad ones? Tell us: is this a "good patent": https://en.wikipedia.org/wiki/Enfish,_LLC_v._Microsoft_Corp https://en.wikipedia.org/wiki/Enfish,_LLC_v._Microsoft_Corp. How about this? https://patents.google.com/patent/US7818399B1/en?oq=us7%2c818%2c399+ https://patents.google.com/patent/US7818399B1/en?oq=us7%2c81... Both of these are considered valid, post-Alice.