4 ms·

It isn’t a software patent. Compression method is patentable as an algorithm.

speedgoose 3y ago
I would assume that such algorithms falls in the "schemes, rules and methods for performing mental acts, playing games or doing business, and programs for computers". https://www.epo.org/en/legal/epc/2020/a52.html https://www.epo.org/en/legal/epc/2020/a52.html
- 4death4 3y ago
  No, it doesn’t. The ban against software patents is to prevent things like “X, but using software” from being patented. A good example is Sonos’ patent on controlling the volume of multiple speakers from an app. That’s just a dumb patent. Video encoding is very complex and non-obvious. Personally, I see the argument for patent protection.
  - speedgoose 3y ago
    I don't know. The sonos patent is stupid indeed, but I could see how it got accepted as it's about hardware speakers with some software control, which can be patented. But h265 looks purely algorithmic to me. Perhaps the patent is about using dedicated hardware decoders, but I'm not sure how that would impact Netflix. Netflix is software on the consumer side.
    - thfuran 3y ago
      >But h265 looks purely algorithmic to me So are patents of physical devices. Those algorithm just uses levers or whatever.
      - GoblinSlayer 3y ago
        Physical devices don't work according to made up rules, they work on their own rules.
  - orra 3y ago
    Video encoding and decoding is complex, but can in principle be done by hand. An electronic computer is simply faster and more accurate. Don't forget that the first computers were teams of humans.
    - nickelpro 3y ago
      Yes, and the algorithm itself, whether by hand or on a computer, is patentable. What is not patentable is taking a previously existing algorithm, prior-art, and claiming patentable novelty by performing that algorithm on a computer.
  - panzi 3y ago
    So you're saying an algorithm is a mathematical method? > The European Patent Convention (EPC), Article 52, paragraph 2, excludes from patentability, in particular > 1. discoveries, scientific theories and mathematical methods;
    - duckmysick 3y ago
      It's more complicated than that. From the European Patent Convention guidelines: https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_3.html https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_3.ht... > The exclusion applies if a claim is directed to a purely abstract mathematical method and the claim does not require any technical means. For instance, a method for performing a Fast Fourier Transform on abstract data which does not specify the use of any technical means is a mathematical method as such. [...] > If a claim is directed either to a method involving the use of technical means (e.g. a computer) or to a device, its subject-matter has a technical character as a whole and is thus not excluded from patentability under Art. 52(2) and (3). [...] > When assessing the contribution made by a mathematical method to the technical character of an invention, it must be taken into account whether the method, in the context of the invention, produces a technical effect serving a technical purpose. > Examples of technical contributions of a mathematical method are: [...] > - encoding data for reliable and/or efficient transmission or storage (and corresponding decoding), e.g. error-correction coding of data for transmission over a noisy channel, compression of audio, image, video or sensor data;
cubefox 3y ago
algorithm = program = software The computer science definition is all the same.
- Izkata 3y ago
  Kinda like that "In theory, theory and practice are the same. In practice, not." Algorithm = documentation or specification, not software.
- jillesvangurp 3y ago
  There's no such thing as "the computer science definition" as far as patent law is concerned. Lawyers aren't engineers or scientists. Patents are basically lists of claims that have a lot of complicated wording that is designed to get interpreted by other patent laywyers in very specific ways. What is and isn't a software patent is extremely murky water. As soon as you involve algorithms, it's not all that different from an actual machine or engineering system from a patent law point of view. Which is why you find the words like "method" or "apparatus" in a lot of patent claims. That's the little legal loophole that makes something an actual patent vs. just a "software patent". Whether it's enforceable or gets past the patent office review is a second thing of course. In the case of the infamous 1 click shopping patent, the answers turned out to be that there was a lot of prior art and the EU declined to accept the patent for being too obvious. In the US it did get accepted but then a later ruling invalidated most of the claims. The patent expired a few years ago. Having the patent is one thing; enforcing it another. The US being a bit more eager to accept all sorts of obvious patents has been used by patent trolls to harass small companies into payments. And it's also uses as a defensive tool by all sorts of big companies as a defensive tool and as a way to get other big companies to pay royalties for stuff that they did. Which, in this case, Netflix thought they could get out of.
  - shmerl 3y ago
    > As soon as you involve algorithms, it's not all that different from an actual machine or engineering system from a patent law point of view. It should be different. The more abstract it gets, the less patentable it should be. Algorithms are more abstract than machines.
phh 3y ago
Could you expand more? I have a bit of a hard time seeing how one can patent algorithms but not software. I kinda see it, but I would appreciate a broader explanation, if you could share more? Also, you can patent the global systems and its algorithms, but you can't patent the software that is in-between? (a quick google doesn't raise much)
- duckmysick 3y ago
  The Guidelines for Examination from the European Patent Office explain this in detail: https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_6.html https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_6.ht... The gist is a computer program must produce a "further technical effect". It can't be just running code on a generic computer. It would have to be specific to the internals of the computer, like a method of securing a bootloader. There's a section with more examples: https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_6_1.html https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_6_1.... I imagine it can lead to more questions than answers. It's complicated.
sekai 3y ago
> Compression method is patentable as an algorithm So math is… patentable?
- gargablegar 3y ago
  Yes if it has a marked impact on physical data storage/size. I have a patent which involves a formulaic approach to measuring and classifying certain “stuff” It is filed as data compression and being able to represent the same thing in a compressed way is why it can be protected.
  - mort96 3y ago
    Where in the law is the condition that "math is patentable if it has a marked impact on data size"?
shmerl 3y ago
How is it any different? Sounds like demagoguery to say that software isn't patentable, but algorithms are. Algorithms shouldn't be patentable even stronger, since it's more abstract than software. It's function claiming otherwise.

I would assume that such algorithms falls in the "schemes, rules and methods for performing mental acts, playing games or doing business, and programs for computers". https://www.epo.org/en/legal/epc/2020/a52.html https://www.epo.org/en/legal/epc/2020/a52.html

4death4 3y ago
No, it doesn’t. The ban against software patents is to prevent things like “X, but using software” from being patented. A good example is Sonos’ patent on controlling the volume of multiple speakers from an app. That’s just a dumb patent. Video encoding is very complex and non-obvious. Personally, I see the argument for patent protection.
- speedgoose 3y ago
  I don't know. The sonos patent is stupid indeed, but I could see how it got accepted as it's about hardware speakers with some software control, which can be patented. But h265 looks purely algorithmic to me. Perhaps the patent is about using dedicated hardware decoders, but I'm not sure how that would impact Netflix. Netflix is software on the consumer side.
  - thfuran 3y ago
    >But h265 looks purely algorithmic to me So are patents of physical devices. Those algorithm just uses levers or whatever.
    - GoblinSlayer 3y ago
      Physical devices don't work according to made up rules, they work on their own rules.
- orra 3y ago
  Video encoding and decoding is complex, but can in principle be done by hand. An electronic computer is simply faster and more accurate. Don't forget that the first computers were teams of humans.
  - nickelpro 3y ago
    Yes, and the algorithm itself, whether by hand or on a computer, is patentable. What is not patentable is taking a previously existing algorithm, prior-art, and claiming patentable novelty by performing that algorithm on a computer.
- panzi 3y ago
  So you're saying an algorithm is a mathematical method? > The European Patent Convention (EPC), Article 52, paragraph 2, excludes from patentability, in particular > 1. discoveries, scientific theories and mathematical methods;
  - duckmysick 3y ago
    It's more complicated than that. From the European Patent Convention guidelines: https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_3.html https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_3.ht... > The exclusion applies if a claim is directed to a purely abstract mathematical method and the claim does not require any technical means. For instance, a method for performing a Fast Fourier Transform on abstract data which does not specify the use of any technical means is a mathematical method as such. [...] > If a claim is directed either to a method involving the use of technical means (e.g. a computer) or to a device, its subject-matter has a technical character as a whole and is thus not excluded from patentability under Art. 52(2) and (3). [...] > When assessing the contribution made by a mathematical method to the technical character of an invention, it must be taken into account whether the method, in the context of the invention, produces a technical effect serving a technical purpose. > Examples of technical contributions of a mathematical method are: [...] > - encoding data for reliable and/or efficient transmission or storage (and corresponding decoding), e.g. error-correction coding of data for transmission over a noisy channel, compression of audio, image, video or sensor data;

cubefox 3y ago

algorithm = program = software The computer science definition is all the same.

Izkata 3y ago
Kinda like that "In theory, theory and practice are the same. In practice, not." Algorithm = documentation or specification, not software.
jillesvangurp 3y ago
There's no such thing as "the computer science definition" as far as patent law is concerned. Lawyers aren't engineers or scientists. Patents are basically lists of claims that have a lot of complicated wording that is designed to get interpreted by other patent laywyers in very specific ways. What is and isn't a software patent is extremely murky water. As soon as you involve algorithms, it's not all that different from an actual machine or engineering system from a patent law point of view. Which is why you find the words like "method" or "apparatus" in a lot of patent claims. That's the little legal loophole that makes something an actual patent vs. just a "software patent". Whether it's enforceable or gets past the patent office review is a second thing of course. In the case of the infamous 1 click shopping patent, the answers turned out to be that there was a lot of prior art and the EU declined to accept the patent for being too obvious. In the US it did get accepted but then a later ruling invalidated most of the claims. The patent expired a few years ago. Having the patent is one thing; enforcing it another. The US being a bit more eager to accept all sorts of obvious patents has been used by patent trolls to harass small companies into payments. And it's also uses as a defensive tool by all sorts of big companies as a defensive tool and as a way to get other big companies to pay royalties for stuff that they did. Which, in this case, Netflix thought they could get out of.
- shmerl 3y ago
  > As soon as you involve algorithms, it's not all that different from an actual machine or engineering system from a patent law point of view. It should be different. The more abstract it gets, the less patentable it should be. Algorithms are more abstract than machines.

phh 3y ago

Could you expand more? I have a bit of a hard time seeing how one can patent algorithms but not software. I kinda see it, but I would appreciate a broader explanation, if you could share more? Also, you can patent the global systems and its algorithms, but you can't patent the software that is in-between? (a quick google doesn't raise much)

duckmysick 3y ago
The Guidelines for Examination from the European Patent Office explain this in detail: https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_6.html https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_6.ht... The gist is a computer program must produce a "further technical effect". It can't be just running code on a generic computer. It would have to be specific to the internals of the computer, like a method of securing a bootloader. There's a section with more examples: https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_6_1.html https://www.epo.org/en/legal/guidelines-epc/2023/g_ii_3_6_1.... I imagine it can lead to more questions than answers. It's complicated.

sekai 3y ago

> Compression method is patentable as an algorithm So math is… patentable?

gargablegar 3y ago
Yes if it has a marked impact on physical data storage/size. I have a patent which involves a formulaic approach to measuring and classifying certain “stuff” It is filed as data compression and being able to represent the same thing in a compressed way is why it can be protected.
- mort96 3y ago
  Where in the law is the condition that "math is patentable if it has a marked impact on data size"?

shmerl 3y ago

How is it any different? Sounds like demagoguery to say that software isn't patentable, but algorithms are. Algorithms shouldn't be patentable even stronger, since it's more abstract than software. It's function claiming otherwise.