Copyright and patents for software in general make perfect sense as long as you have a patent-awarding body with more smarts than the USPTO.
If I write some software, why should anyone else be able to copy it buckshee if I don't want them to? If it were a novel it would be copyrightable, and software is only the same thing - text (or at least representable as text). Software patents make sense too if handled sensibly, I should know as I've got a half dozen or so of the things. If I make a hardware widget that performs a certain action, I can patent that, so why not a software widget too? Where it all goes wrong is the USPTO granting massively overbroad patents or patents for things that (patently) fail the test of non-obviousness.
Anyway, that's not the case in point here, what is being argued over is APIs, a very different thing from software and a lot more slippery to pin down. In a way it's more like the old spat between Intel and Zilog over the Intel 8080 instruction set, which is sort of like the API for a processor's internal workings. Zilog reimplemented the entire 8080 instruction set from scratch (and some) and then the Intel lawyers turned up. In the end, it was ruled that Intel couldn't stop other companies reimplementing the same instruction set, but they could copyright the instruction mnemonics and so Zilog had to give all the instructions different mnemonics in their assembler. A bit of a pain if you needed to work on both Intel and Zilog, but liveable.
Edit for crossposting: there is only one indefinite copyright that I'm aware of, and that's Peter Pan. Everything else is lifetime plus a certain number of years which varies from jurisdiction to jurisdiction.
Biting the hand that feeds IT
