You see where I'm going with this? Ideas are easy. Implementation is 99 44/100 % of the work -- always has been, and probably always will be. And yet now, we're at the point where not only are companies patenting specific implementations, which I would have a problem with in and of itself (copyright vs. patent) but ideas before an implementation even exists. It's not like they've done the work and want to protect the fruits of their labor; they're saying, in effect, "If anyone ever does this work, we own their ass."
Are you saying, for example, that a company could go out and patent a polynomial solution to an NP-complete problem without actually having an algorithm that implements the solution? Could you provide some specific examples of companies patenting an idea before a solution exists? I'm not saying that they don't exist, I'm genuinely curious to know if this is happening.
Also I'm curious as to why you think a specific implementation shouldn't be patentable. If I did write an algorithm to solve an NP-complete problem in polynomial time then is that algorithm copyrighted by me? If so, is your argument that there's no needs for patents because copyright provides sufficient protection? If that's true then what were the actual arguments for software patents in the first place?