It is impossible to know for certainty if innovation would be faster or slower if patents were not given for software, but I feel it is a key question. You see

It is impossible to know for certainty if innovation would be faster or slower if patents were not given for software, but I feel it is a key question. You see well known problems discussed, such as prior art being buried in compiled code and the lack of awareness in general of the prior art the examiners have (usually they just do searches on the patent archive). However, I haven't seen as much discussion of the problem that software patents are just not needed to spur innovation in the software arena in the first place, since rapid innovation would occur anyway. (Not to mention granting a 20-year monopoly in regular years is like granting a 100-year monopoly in software years.)