Patents don't cover algorithms. They can cover the use of algorithms to accomplish a particular task but that doesn't affect others much.
In fact, the general patents rely on the doctrine of equivalents and are thus algorithm independent.
The closest you can get to an algorithm patent is to claim the use of an algorithm to produce algorithm-specific output. For example, one could patent the use of a new sorting algorithm to produce sorted output. However, such patents are almost all extremely weak because there are typically multiple algorithms to accomplish the same result.
Yes, a patent on an algorithm to solve a specific NP-complete problem in polynomial time would affect others, but isn't that exactly the sort of thing that patents should cover?
Patents don't cover algorithms. They can cover the use of algorithms to accomplish a particular task but that doesn't affect others much.
In fact, the general patents rely on the doctrine of equivalents and are thus algorithm independent.
The closest you can get to an algorithm patent is to claim the use of an algorithm to produce algorithm-specific output. For example, one could patent the use of a new sorting algorithm to produce sorted output. However, such patents are almost all extremely weak because there are typically multiple algorithms to accomplish the same result.
Yes, a patent on an algorithm to solve a specific NP-complete problem in polynomial time would affect others, but isn't that exactly the sort of thing that patents should cover?