6 ms·

That same argument could be levied against all patents more generally, and as a result, I don't think it really helps the anti-software-patent cause. Physical p

That same argument could be levied against all patents more generally, and as a result, I don't think it really helps the anti-software-patent cause. Physical patents have been licensed for this reason (to avoid the r&d of alternative methods) since nearly the advent of patents.

konschubert 8y ago
The hard thing in software is not idea, but execution. Patents that patent trivial ideas are nothing but roadblocks. Edit: There are some ideas in software that were non-obvious. Algorithms and protocols. Those might make sense to be patent-able.
- monocasa 8y ago
  Algorithms should be the least patenable as math is explicitly not granted patent protections.
  - vetinari 8y ago
    Neither should be protocols; it is easy to come up with a new one, all that patenting one does is to prevent interoperability.
  - tfha 8y ago
    I thought math was explicitly not patentable
    - zeroxfe 8y ago
      Yep, that's what monocasa means.
- rayiner 8y ago
  > The hard thing in software is not idea, but execution. It depends entirely on the software. I spent much of my career as an engineer developing software for SDRs. E.g. I spent the better part of a summer doing Matlab simulations to figure out how you can get a network of frequency agile SDRs to quickly converge onto the same radio channel when they start up, where the spectrum environment is unknown a priori and the radios cannot communicate with each other without being on the same channel. Figuring out what code to write (which involves a lot of theory and math) is the hard part. Once you figure that out, the rest is just coding. Another area is protocols, where the "idea" is more important than any given implementation. Inventing TCP/IP (and verifying it works) is harder than implementing TCP/IP. Of course, just because you give away the idea doesn't mean nobody has to pay for it. In the case of TCP/IP, the U.S. military paid for it. In the case of its conceptual successors (QUIC), advertising is paying for it. Without software patents those are your two main avenues for bankrolling protocols. I think at least the second one is inferior compared to a model based on patent protection, like MPEG or h.264.
  - schrodinger 8y ago
    > how you can get a network of frequency agile SDRs to quickly converge onto the same radio channel when they start up Can you share the outcome?
- Zak 8y ago
  The thing that's legally meant to be patentable is the execution, not the idea. The thing is in a lot of cases, the execution isn't the hard part: Given an e-commerse website having a database full of user accounts that includes their shipping address and credit card information, implement one-click ordering. Given a capacitive touchscreen and UI framework for interacting with it, implement slide-to-unlock. A programmer competent in the use of the relevant technologies can easily implement these requests, and will probably come up with a solution similar to those employed by Amazon and Apple, respectively. That one-click ordering on an e-commerce site was a good idea was actually novel, but there isn't a form of IP protection that's intended to cover mere ideas that are demonstrated publicly.
TheSpiceIsLife 8y ago
Are patent holder obliged to license a patent?
- Retric 8y ago
  In some situations yes, but generally no.
lemoncucumber 8y ago
Software is unique though in that it can be covered by both patents and copyright. Maybe copyright alone would suffice.

konschubert 8y ago

The hard thing in software is not idea, but execution. Patents that patent trivial ideas are nothing but roadblocks. Edit: There are some ideas in software that were non-obvious. Algorithms and protocols. Those might make sense to be patent-able.

monocasa 8y ago
Algorithms should be the least patenable as math is explicitly not granted patent protections.
- vetinari 8y ago
  Neither should be protocols; it is easy to come up with a new one, all that patenting one does is to prevent interoperability.
- tfha 8y ago
  I thought math was explicitly not patentable
  - zeroxfe 8y ago
    Yep, that's what monocasa means.
rayiner 8y ago
> The hard thing in software is not idea, but execution. It depends entirely on the software. I spent much of my career as an engineer developing software for SDRs. E.g. I spent the better part of a summer doing Matlab simulations to figure out how you can get a network of frequency agile SDRs to quickly converge onto the same radio channel when they start up, where the spectrum environment is unknown a priori and the radios cannot communicate with each other without being on the same channel. Figuring out what code to write (which involves a lot of theory and math) is the hard part. Once you figure that out, the rest is just coding. Another area is protocols, where the "idea" is more important than any given implementation. Inventing TCP/IP (and verifying it works) is harder than implementing TCP/IP. Of course, just because you give away the idea doesn't mean nobody has to pay for it. In the case of TCP/IP, the U.S. military paid for it. In the case of its conceptual successors (QUIC), advertising is paying for it. Without software patents those are your two main avenues for bankrolling protocols. I think at least the second one is inferior compared to a model based on patent protection, like MPEG or h.264.
- schrodinger 8y ago
  > how you can get a network of frequency agile SDRs to quickly converge onto the same radio channel when they start up Can you share the outcome?
Zak 8y ago
The thing that's legally meant to be patentable is the execution, not the idea. The thing is in a lot of cases, the execution isn't the hard part: Given an e-commerse website having a database full of user accounts that includes their shipping address and credit card information, implement one-click ordering. Given a capacitive touchscreen and UI framework for interacting with it, implement slide-to-unlock. A programmer competent in the use of the relevant technologies can easily implement these requests, and will probably come up with a solution similar to those employed by Amazon and Apple, respectively. That one-click ordering on an e-commerce site was a good idea was actually novel, but there isn't a form of IP protection that's intended to cover mere ideas that are demonstrated publicly.

TheSpiceIsLife 8y ago

Are patent holder obliged to license a patent?

Retric 8y ago
In some situations yes, but generally no.

lemoncucumber 8y ago

Software is unique though in that it can be covered by both patents and copyright. Maybe copyright alone would suffice.