> Why Oracle didn't sue Apache?
Simple: because they didn't have the motivation (see point (I)). Their goal wasn't to copyright APIs (that just was one of the many tactics used in court, and the one that just happened to win), but to get their share of Android. Suing Apache wouldn't have helped their goal in the least.
Also, while not relevant to the question, I will note that even if APIs are copyrighted, Harmony has a much stronger fair-use defense than Android: 1. It was done for the purpose of interoperability and 2. it was not intended to make a profit. Each of these is a fair use defense alone, and certainly together, but neither is true for Android (although the first might be partially true).
> Sun didn't sue Google in the 5 years Google used Android
Because they didn't have the resources and the inclination, as the company was already in a downward spiral and such a lawsuit could only have hurt them at the time. Google, however, fully expected them to.
> The ruling talks about API and API's has nothing to do with being a language level API or a "web API".
That's only because the term "API" -- until very recently -- was only used to describe language-level API, while what today is known as a web API has always been called a protocol, so don't let the recent overloading of the term confuse you.
More to the point, the reason copyright may apply to APIs but certainly not to protocols is because copyright law is not only concerned with the intent, abstraction, or use of the work, but also with its form. This why other IP protections (like patents) exist, to cover works whose form does not match those covered by copyright.
In particular, copyright requires a fixed representation (in text, image, recording etc.). So, for example, if I tell you a story, the story isn't copyrighted. But if I write it down -- or even record it -- then the text of the story or the recording is copyrighted. It doesn't matter that the purpose in all three cases was the same: to convey a story (once that fixed form is copyrighted, the protection may extend to non-exact reproductions, but that original fixed-form must exist). While a protocol may be described in a text and that text is copyrighted, the described protocol itself cannot be, because it does not have a fixed representation in text (a completely different text could describe the same protocol). A language API does have a fixed form.
> the appeal court ruled that the fair use question must be decided in the forthcoming case, nothing more. As any other copyright case
Yes, but that was the point. Google made some arguments that they believed would make APIs non-copyrightable. The court ruled that those arguments are fair-use arguments which are not relevant to the question of copyrightability, but may be used as a defense once APIs are copyrighted.
> No, it has to go to court so yes, any player must be concerned
Quite the opposite. It is the copyright holder who has to sue you. Any company would sue you if they believe they have a case, and you would still need to defend yourself in court when they do. The question of API copyrightability had never been decided, so even before the ruling there was a good chance you'd be sued, and you'd still need to defend yourself. Now, it is true that a ruling against API copyrightability would have removed this threat (which had existed even prior to this case), but now your position is not much different than before, only that except making two arguments (as Google did) -- that API are not copyrightable and if they are, you've made fair use of them -- you only have one: fair use.