I think using an API for the purposes of interoperability should be allowed under fair use.

However! My controversial opinion is that Google's use of the Java API wasn't for interoperability. They didn't use the Java API so that Java programs could run on Android, they used it so that Java developers could transfer their skills to a new platform which is precisely the value of a cleanly designed API -- value that Oracle doesn't want to provide for free.

You're missing an argument in there that joins your two assertions.

The fact that something is creative or requires skill to produce doesn't mean that it should be copyrighted.

Copyright is a government enforced monopoly. In order for society to enforce that monopoly on behalf of private enterprise, there has to be a quid pro quo. The usual one is to promote creation and distribution of new works.

The question then is: would no APIs exist if they weren't copyrightable?

Unfortunately courts generally can't write laws for the legislature.

A creative work that takes skill isn't the definition of what should get copyright. For instance sewing patterns and recipies aren't granted copyright.

> My controversial opinion is that Google's use of the Java API wasn't for interoperability.

Following that logic, then any even slightly incompatible implementation wouldn't be legal. For instance WINE isn't a perfect recreation of Win32 by any means.

It's not the API documentation that's at dispute here. It's literally the abstract "structure, sequence, and organization" of the API. And there's tons of ninth circuit case law that sso isn't copyrightable.

> I never said it had to be perfect, just that the reason for using the API is for interoperability -- WINE is exactly that. It allows software that was designed for Windows to run on Linux.

What's the test here that a court can use to make that distinction? And just letting you know at one of my former jobs, we modified our codebase to better run on Wine a few times.

This brings to mind something I heard from my legal friends in the early days of this case. Their position was that this wasn't a dispute about APIs, copyright, etc. It was a licensing dispute -- Sun/Oracle wanted to charge for Java on Mobile, and Google didn't want to pay. Google then executed an engineering 'hack' whereby Java JVM bytecode (.class files) was converted to Google's Dalvik VM format (.dex) via a tool (dx), thus avoiding Sun's licensing fees and still being able to take advantage of the Java ecosystem.

Their argument would have been this was a fairly mechanical effort to avoid licensing fees, analogous to reprinting a copyrighted book in Piglatin, and therefore Google was liable to pay licensing fees to Oracle.

This was from ages back though; the case has of course evolved since then. I do wonder however if Google's effort to paint this as a case of straightforward re-implementation of an API may not backfire upon them.

We need a system that rewards those who experiment & innovate, if anything by Oracle winning this case it further highlights the value in Oracle's IP (promoting innovation and competition).