Computer programs are constructivist mathematical proofs. Why is this annoying? Why isn't this a good argument against software patents? Should I be able to patent a mathematical theorem and its proof?