3 ms·

As I read more about software patents, and with what knowledge I've gained working with software, I remain opposed to the idea of software patents in general.

As I read more about software patents, and with what knowledge I've gained working with software, I remain opposed to the idea of software patents in general. My problem, though, is that I have no concept of what the arguments are that support software patents. Can anyone list any of the more compelling arguments supporting software patents, or any blogs/articles that discuss why they're not bad?

throwaway89 13y ago
My two cents: First of all, what is a software patent? Software, per se, is not eligible for patent protection under US patent law. But claiming a method which uses a computer processor is legal. Whats the difference? transformation of a machine... software == bad, software executing a on a processor ==good. When you see some article that says Country X is not allowing software patents...It is not allowing patents on purely software. Case for "software" patents- Why not? If 90 years ago, someone can invent a telecommunications system with hardware, then why can't I do it today with software? If I come up with an "idea" using Twillio's API and it is novel, uesful, and meets the 4 categories of patents - I am entitled to a patent. However, it should be obvious to everyone that if i'm using Twillio's API, there should be prior art available to my idea. However, if there is not any prior art, then I am entitled to my patent. In other words, my "idea" should be easy to reject because there are obviously tons of prior art/teachings available. However, if these are granted as patent...Then this is a quality issue argument and not whether software should be legal. Also, if we can get rid of trolls out of the system, we'd me far better off.
- cma 13y ago
  Should math ideas be patentable? Pay up to use a technique used in the proof of the Poincare Conjecture? Make novel mathematics have mysterious 20 year gaps between bursts of new discoveries? One time I searched for an API that would let you send a unicast UDP message to multiple receivers with just one system call. I just thought why not reduce copying and lower the number of system calls? Nope, passing a N addresses/ports instead of 1 address/port is a patented technique when used in a networking system call. By reading this you've automatically doubled your patent-breach liability. Patents were meant to spread ideas, but even hearing or reading about one doubles your penalty if a jury doesn't accept your clever work-around and finds you liable.

throwaway89 13y ago

My two cents: First of all, what is a software patent? Software, per se, is not eligible for patent protection under US patent law. But claiming a method which uses a computer processor is legal. Whats the difference? transformation of a machine... software == bad, software executing a on a processor ==good. When you see some article that says Country X is not allowing software patents...It is not allowing patents on purely software. Case for "software" patents- Why not? If 90 years ago, someone can invent a telecommunications system with hardware, then why can't I do it today with software? If I come up with an "idea" using Twillio's API and it is novel, uesful, and meets the 4 categories of patents - I am entitled to a patent. However, it should be obvious to everyone that if i'm using Twillio's API, there should be prior art available to my idea. However, if there is not any prior art, then I am entitled to my patent. In other words, my "idea" should be easy to reject because there are obviously tons of prior art/teachings available. However, if these are granted as patent...Then this is a quality issue argument and not whether software should be legal. Also, if we can get rid of trolls out of the system, we'd me far better off.

cma 13y ago
Should math ideas be patentable? Pay up to use a technique used in the proof of the Poincare Conjecture? Make novel mathematics have mysterious 20 year gaps between bursts of new discoveries? One time I searched for an API that would let you send a unicast UDP message to multiple receivers with just one system call. I just thought why not reduce copying and lower the number of system calls? Nope, passing a N addresses/ports instead of 1 address/port is a patented technique when used in a networking system call. By reading this you've automatically doubled your patent-breach liability. Patents were meant to spread ideas, but even hearing or reading about one doubles your penalty if a jury doesn't accept your clever work-around and finds you liable.