I wasn't making the case for software patents -- just pointing out that the logic doesn't work. Just because the constituent parts of a whole don't satisfy some legal criteria doesn't mean that the "whole" itself doesn't.
My issue is you restated his analogy changing the legal categories and the subjects and then disproved your interpretation. Patent law and copyright law are not the same and math's relation to software is not the same as the relation of individual words to a book.
His argument is weak, but you are attacking something else.
