No, this is incorrect. Because it was a copyright case decided by the Federal Circuit, it is precedent nowhere. Literally no court will need to treat it as precedent. Please educate yourself further before making these lengthy posts in threads like this; you're spreading misinformation.
> Also, copyrighted APIs are only "language-level" APIs... not web APIs (or protocols, as we used to call them)
From this and your post below, this appears to be a distinction you have made up (to appeal to the audience here? I have no clue). "Structure, sequence and organization" are certainly aspects of web APIs that could be copyrightable under the Federal Circuit's ruling.
Also, no one called web APIs "protocols" anymore than they called system calls "protocols". Unless you're talking about actual protocols, in which case I'd just stick to calling them "protocols" because that's what we still call them.
> There are several fair use defenses, but two apply to the industry in cases like this: 1/ non-profit use, 2/ use for the sake of compatibility (to prevent vendor lock-in).
Non-profit use is neither necessary or sufficient for fair use (see Google's recent book scanning ruling, for instance), and the compatibility argument is not with existing JVMs but with existing Java source code.
As for the rest of you post on fair use, I would suggest other readers to take a look at the actual four factors of fair use in the US, as your summary is a poor one. One good source: