For at least a couple of decades we've been copyrighting software, and now I realize that I don't understand why we do. Talking with Doc Searls last night, who is a master of analogy, we kind of agreed that perhaps there is no suitable analogy for software, that software is a unique thing unto itself.
First: Yes, software is unique. I don't want to go too deeply into that though as that's a very, very long story to do it truly correctly.
Second: Rewriting the question a bit (I believe fairly) to be "Why should we copyright software?", I would observe that there is more to copyright then most people think, some of which is just implicit in the copyright.
Holding a copyright on something means you can enforce the integrity of the work, whatever it may be. It means that nobody can download Radio Userland, rename it, write Userland (and all other attribution) right out of the codebase, and sell it for $20 a pop. "Moral rights" like these are so importent in Europe that you actually can't sell the right of attribution. No matter what else you sell to some company, you retain the credit.
Holding the copyright to the software also ensures that Dave & Userland retain control over Frontier et al. Nobody can pick it up and force it in a direction contrary to the wishes of Userland. Dave has written somewhere about how importent this is to him, but I can't find a reference. (IIRC, he mentioned this in the context of open source and possibly open-sourcing Frontier.)
When you don't treat copyright as a stick to beat people with, but instead as ethics codified into law about how we treat each other's work, there's nothing objectionable about it, and no reason not to apply it to software.
That said, it must be applied differently, and again, explaining that is a long task. (In fact, I've got a currently 70-page "essay" that I've resumed work on that still has only barely explained why this is true; I still need to propose my solution. I wish I could write as quickly as I could think.)