A few points.

1. It sounds like the most compelling argument for why patents are "better" than copyright has to do with the shorter term. Indeed, there is research in which people have tried to determine the optimal copyright term economically, given production and reproduction costs, and have found terms on the order of 14 years and even 2-4 years. But of course, these studies have a huge assumption -- that copyright is economically beneficial at all, an assumption that the authors of the latter work acknowledge and disagree with, and they note that absence of copyright beats any "optimal" term.

2. I have yet to see any research that shows social benefits for software patents -- or even benefits for the software industry. It appears to be redistribution + parasitic loss. I've seen some hand-waving to the effect that patents are useful because they are "assets" that a start-up company can use to attract funding... but no one I know who is really pursuing a start-up cares about this from a development standpoint. (VCs care, because if the start-up fails, they can just sell them off to trolls or companies worried about blocking litigation -- again, totally parasitic on innovation.)

3. There's a really fundamental problem here: a software patent is a monopoly on a mental process -- a thought, an idea, a mathematical formulation that just happens to use a computer to implement. There's no way to "design around" a concept -- you're simply blocked from having that thought and doing something useful for it, or else the coercive power of the state falls on you. Look at this case or this one (I'm linking to blog posts about them, but it's easy to link to the opinion themselves). Now, Rader or Newman would say that, hey, if these are "obvious" thoughts, then the patent is invalid... but in principle, there can be a coercive, state-enforced monopoly.

So, I guess, yeah, maybe patents are better because at least innovation can be taxed and hampered (if not blocked altogether) for 17 years or 20 years or whatever, as opposed to for 99 years or 150 years or however long copyrights last, but in the context of software that is only relevant for a few years (at most), it's still bad. And, in the case of that second case I linked to, Classen, it's verging on a moral abomination.

-- BahradSokhansanj - 05 Oct 2011