Supreme Court sides with developers in Google vs. Oracle
github.blog
github.blog
Does anyone care to weigh in on this?
I'd agree that the court was right to dodge the issue of copyrightability by going for the most narrow interpretation, it's for the lower courts to decide.
Having said that, I think they shouldn't be copyrighted really, and the door is open for expensive litigation by larger companies shaking down smaller ones with this ruling.
As far as hijacking an API goes, well anyone is free to fork an MIT-licensed project, doesn't mean that it will gain traction.
Most of computing is built upon previous work so claiming originality in an API... well many more clever people than me have left their discoveries open to the public, and I suspect that any API originality doesn't measure up well against their work :)
The hardest thing for competing chip manufacturers to overcome in the deep learning world with is the CUDA APIs / CuDNN API etc.
HN threads tend to portray this a bit as "beating an evil cash grab," which may be the case...but I don't know enough about both sides to make that judgement myself.
1.) The Court skirted the issue of whether APIs are copyrightable. If they were, it would be bad, but it's unclear if they aren't copyrightable as a matter of the law as written.
2.) The majority held on the basis of fair use and, as Thomas argues, that's arguably a stretch--and certainly a very fact-dependent one.
That said, the down side of this is the old MS "embrace, extend, extinguish" approach is up for grabs by a large company. There isn't much stopping someone from making "not java" and breaking it in subtle ways (Imagine, for example, if android's dalvik/art became a common runtime for Java apps. Or more realistically, what microsoft did with Java back in the sun days).
The good thing is that for now, just implementing interfaces likely is just fair use. The bad thing is you are allow to implement a LOT of interfaces before the copyright question comes up again.
Bear in mind, Google's internal correspondence covered knowing they needed to license Java, negotiating terms to license Java, and then when they didn't want to comply with Oracle's terms (which were literally: interoperability, because they wanted Android to run standard cross-platform Java apps), they chose to steal what they wanted instead. Andy Rubin (of "Larry Page gave me $90 million after I sexually harassed subordinates" fame) had a major bonus tied to launching Android on time, and just taking what he needed was the solution.
Also, if Android had been GPL-licensed, they could've built on OpenJDK from the get-go legitimately, but again, stealing was preferable in Google's case to obeying license terms, because they wanted to appeal to manufacturer's pushing proprietary code.
Oracle is not a hero, but Google was fighting against interoperability and open source licensing the entire way, and bought their way into public opinion.
Why is that? How is what Google did any different than what Stallman and Torvalds did when they created free reimplementations of the Unix API, or how the Wine and ReactOS projects reimplement the Windows API?
The way I see it, the ability for us to write a drop-in-compatible replacement for proprietary software is a critically important right that allows us to defeat vendor lock-in, and it has fueled the growth of the software industry for decades.
Also, Android wasn't a drop-in-compatible replacement. If it was, it might have a case, but the whole breakdown between Google and Oracle in negotiating was that Oracle had insisted on interoperability, and Google wanted to break compatibility.
This is why the idea that the Supreme Court ruled on fair use is so freaking laughable, everything Google was trying to do was an antithesis to fair use, interoperability, and open source.
The initiative that led to OpenJDK started in 2006 with the release of the JDK (IIRC, minus some components Sun lacked rights to) in 2007.
Android development started in 2003, with the first beta in 2007.
OpenJDK wasn’t, in any meaningful sense, available for the development of Android.
So what? The question is whether what they did was fair use, and the Court held that it was. The availability of alternatives under other licenses isn't part of the analysis.
Or just believed by developers who know what they are talking about you know... Nobody's really siding with Oracle in the industry because their case did not make much sense.
> they chose to steal what they wanted instead
They reimplemented everything you mean, there's no Oracle source code at all in Android and there's not even a JVM.
See lots of discussion from a few days ago on the ruling here: