For me, the problem here is that there is no logically consistent way to rule in favor of Google without undermining the concept of software copyright itself. APIs are creative expression: at a basic level, it is possible to change the names, the order, and to some extent the type of parameters and achieve the same result. It is also possible to change the structure of the API itself at a large scale in such a way that the programmer still can do what the API purports to provide, in a completely different way. Therefore, if one were to call APIs merely functional "declaring code" that cannot be copyrighted, then that would put
all software code at risk of being called such. The idea that, say, a certain algorithm can be expressed in many different ways lies at the heart of how we justify making such expressions copyrightable.
With this ruling, either software will be found to be entirely (or mostly) uncopyrightable, which is unlikely, or software copyright will turn into an even bigger legal morass that requires a team of top-tier lawyers just to understand which parts of your software are effectively copyrightable (or potentially infringing) and which aren't.
The Supreme Court should have found in favor of Oracle, and told Google to bring their case before Congress if they're so worried that API copyrightability would destroy the industry.