I think Thomas makes a mistake in disregarding the reason for introducing the distinction between declaring and implementing code. The law (as I understand it -

I think Thomas makes a mistake in disregarding the reason for introducing the distinction between declaring and implementing code. The law (as I understand it - IANAL) doesn't exhaustively enumerate what does and does not constitute fair use. Therefore, any specific instance of fair use is bound to introduce new distinctions that are not specifically mentioned in the law itself. In this particular case, Google's copying of interfaces was ruled to be fair use because its purpose was to allow Java developers to reuse their skills. This reasoning only holds for the declaring parts of the copyrighted work though. And that's why the distinction becomes necessary even if Congress rejects it as a primary distinction between copyrightable and non-copyrightable work. That's also why I think the ruling is perhaps not quite as broadly applicable as some (including Thomas) are thinking right now. It may not cover code that isn't used by a significant number of developers. I haven't read the entire ruling, so I'm not sure whether it covers non-public interfaces that are nevertheless useful for interoperability. Such interfaces were at the center of past disputes (I think Microsoft's SMB protocol was one such case).