3 ms·

Invalidating a patent to me is an admission that the patent shouldn't have been granted in the first place. I appreciate that there is a long legal battle to ge

Invalidating a patent to me is an admission that the patent shouldn't have been granted in the first place. I appreciate that there is a long legal battle to get there but the patent authority then charging a fee feels like them getting money for their own failings. If anything they should be paying those people who's business was in legal limbo compensation for their error.

pbhjpbhj 5y ago
So then, instead of accepting that patent offices can't find all relevant prior art/prior use (disclosed anywhere in the World) you require perfection in the discovery of relevant 'art' (closely related technical documents). That shifts the problem from "occasionally a patent is granted that isn't valid" to "it will cost $billions for even a single patent just to do a complete search" (more realistically this isn't possible you'd have to exhaustively search all prior publications). IMO if you want a patent system you have to make it work without expecting perfection from patent offices.
- AlbertCory 5y ago
  > "occasionally a patent is granted that isn't valid" ROTFL. It's more like "with software, the patent is virtually always invalid." I'm sure there are people better than I am, but I can invalidate practically any software patent. And I did it for Google, including in Germany in the Google Maps case of 2014. The average time an examiner is given for a patent is a day and a half (old data). That is not nearly enough time to find the "relevant prior art." So what the patent offices are doing is handing out worthless government "monopolies" and making tons of money in fees while they're doing it.
  - pbhjpbhj 5y ago
    >ROTFL. It's more like "with software, the patent is virtually always invalid." // Yes, the establishment of prior art is particularly hard for software IMO, but that's mainly not an EPO issue (per the OP) as software patents in Europe are much more restricted compared to before the USPTO. Your timing is broadly right IME, probably a bit generous for less complex areas (though it varies and EPO spend a bit longer). Note that in part it's an industry issue - if software companies wanted to help them they can point out prior art before parents are granted (it's part of the process in EPO and USPTO jurisdictions). >I'm sure there are people better than I am, but I can invalidate practically any software patent. // EPO _granted_ patents? I'm highly doubtful you can do that in the normal search period (0.5-1.5 days). If you want to argue that far more resources should be devoted to patent examination; that's possible but in EPO countries the balance seems about right -- countries have a hard job arguing for thousands more civil servants to save only a few (!?) court cases. In USPTO you need an overhaul of the court system, massive increase in examining staff I don't think we'll fix it. All just my opinion of course.
    - AlbertCory 5y ago
      > "EPO _granted_ patents? I'm highly doubtful you can do that in the normal search period (0.5-1.5 days)." "In the normal search period" is the key phrase here. If you're granting a government monopoly which is extremely costly to challenge (as the article demonstrates), then you owe us more than a cursory look at the prior art. In any case, this is probably an argument for just making software nearly impossible to patent, even more than the EPO already does. As (I think) Joel Spolsky said, no more than 5 or 6 per year sounds about right.
      - pbhjpbhj 5y ago
        Yeah, with product lifetimes and development in general being on a much shorter cycle parents should probably have reduced in term to reflect that.
- VBprogrammer 5y ago
  The remark about paying compensation was slightly in jest. I do think it's unreasonable for there to be a fee for issuing a patent as well as invalidated one though. Expecting perfection is a long way from expecting their incentives to be aligned with the desired outcomes.
AlbertCory 5y ago
Correct. The PTO (speaking of the US here) takes in way more money in fees than it spends, and doesn't even get to keep that money to improve its searching.

pbhjpbhj 5y ago

So then, instead of accepting that patent offices can't find all relevant prior art/prior use (disclosed anywhere in the World) you require perfection in the discovery of relevant 'art' (closely related technical documents). That shifts the problem from "occasionally a patent is granted that isn't valid" to "it will cost $billions for even a single patent just to do a complete search" (more realistically this isn't possible you'd have to exhaustively search all prior publications). IMO if you want a patent system you have to make it work without expecting perfection from patent offices.

AlbertCory 5y ago
> "occasionally a patent is granted that isn't valid" ROTFL. It's more like "with software, the patent is virtually always invalid." I'm sure there are people better than I am, but I can invalidate practically any software patent. And I did it for Google, including in Germany in the Google Maps case of 2014. The average time an examiner is given for a patent is a day and a half (old data). That is not nearly enough time to find the "relevant prior art." So what the patent offices are doing is handing out worthless government "monopolies" and making tons of money in fees while they're doing it.
- pbhjpbhj 5y ago
  >ROTFL. It's more like "with software, the patent is virtually always invalid." // Yes, the establishment of prior art is particularly hard for software IMO, but that's mainly not an EPO issue (per the OP) as software patents in Europe are much more restricted compared to before the USPTO. Your timing is broadly right IME, probably a bit generous for less complex areas (though it varies and EPO spend a bit longer). Note that in part it's an industry issue - if software companies wanted to help them they can point out prior art before parents are granted (it's part of the process in EPO and USPTO jurisdictions). >I'm sure there are people better than I am, but I can invalidate practically any software patent. // EPO _granted_ patents? I'm highly doubtful you can do that in the normal search period (0.5-1.5 days). If you want to argue that far more resources should be devoted to patent examination; that's possible but in EPO countries the balance seems about right -- countries have a hard job arguing for thousands more civil servants to save only a few (!?) court cases. In USPTO you need an overhaul of the court system, massive increase in examining staff I don't think we'll fix it. All just my opinion of course.
  - AlbertCory 5y ago
    > "EPO _granted_ patents? I'm highly doubtful you can do that in the normal search period (0.5-1.5 days)." "In the normal search period" is the key phrase here. If you're granting a government monopoly which is extremely costly to challenge (as the article demonstrates), then you owe us more than a cursory look at the prior art. In any case, this is probably an argument for just making software nearly impossible to patent, even more than the EPO already does. As (I think) Joel Spolsky said, no more than 5 or 6 per year sounds about right.
    - pbhjpbhj 5y ago
      Yeah, with product lifetimes and development in general being on a much shorter cycle parents should probably have reduced in term to reflect that.
VBprogrammer 5y ago
The remark about paying compensation was slightly in jest. I do think it's unreasonable for there to be a fee for issuing a patent as well as invalidated one though. Expecting perfection is a long way from expecting their incentives to be aligned with the desired outcomes.

AlbertCory 5y ago

Correct. The PTO (speaking of the US here) takes in way more money in fees than it spends, and doesn't even get to keep that money to improve its searching.