Which ones?
I want all useful algorithms to become publically known. Patents help accomplish that.
I don't think that's true. If you invent an algorithm that is genuinely novel and you get a competitive advantage from it, patenting it is the last thing you want to do. Keep it a trade secret.
Why? Because as a legitimate company you can't enforce the patent. If you try, you'll just be counter-sued for infringing someone else's patents.
The one that I was thinking of was http://en.wikipedia.org/wiki/Karmarkars_algorithm .
> If you invent an algorithm that is genuinely novel and you get a competitive advantage from it, patenting it is the last thing you want to do. Keep it a trade secret.
You're assuming that the competitive advantage from using the algorithm exceeds the value that one can get from licensing. That's not necessarily true.
> Because as a legitimate company you can't enforce the patent. If you try, you'll just be counter-sued for infringing someone else's patents.
That's not necessarily or always true. Take drugs - the majors don't clone one another - they honor each other's patents.
There are some interesting optimization methods that probably wouldn't have been made public if it weren't for patents.
I want all useful algorithms to become publically known. Patents help accomplish that.
How do you propose to get people to forego trade secret protection?