Google "uses" Java in the toolchain, for some definition of "use." But They don't sell anyone any Java. It will be tough sledding for anyone trying to show there is Java IP inside Android.
... but all apps are "written in Java", in practice, and one could argue that Android would not have obtained such a large ecosystem without this play. This is all but admitted in Google emails from the Sun days. Whether this means Google has to somehow credit Oracle/Sun, it's for the courts to decide, and as of now, the score is 1-1, so not really clear-cut.
> It will be tough sledding for anyone trying to show there is Java IP inside Android.
This has already been considered in the original trial and the appeal. IIRC, a few secondary files were reused or copied verbatim at some point; regardless, Oracle's position is that the API itself is copyrightable, and as such any implementation has to acknowledge the original rights.
(Note: I don't think Oracle should win, I just acknowledge that neither party really has any higher ground, moral or otherwise.)
That's not strictly true. And if it is or not, it can't be relevant. If you take a JVM language like Scala, compile it to Java bytecode, and run it on a Java(tm) runtime you downloaded from oracle.com, does Oracle have any IP rights in Scala? In the toolchain? In your code?
What if you wrote it in Java? Same answers.
What if you translate that code to Dalvik bytecodes? Same answers.
When I say "There is no Java in Android" I mean the runtime in Android is not Java. Everything else is code that was written in Java and translated to Dalvik and runs in those runtimes that are themselves not Java.
So either Oracle has IP rights all over the place, in everything that is in source code that a Java compiler compiles, or that at some point in the toolchain is in Java bytecode, including stuff like Scala, or not.
Google, and Android, make and sell none of the things in which Orcale has plausible IP rights.
That's what the courts have been asked to decide. You can argue with me until we're both dead and this will not change.
> If you take a JVM language like Scala, compile it to Java bytecode, and run it on a Java(tm) runtime you downloaded from oracle.com, does Oracle have any IP rights in Scala?
No. But if you build a like-for-like JVM exposing the same API calls as the official runtime, Oracle's position is that you are infringing their IP rights (unless you do stuff for binary compatibility and yadda-yadda-Sega), since your are "copying" the "intellectual work" that the API design is.
At least one court agreed with them, and the US Supreme Court "refused to disagree", so it can't be so out-of-this-world as a concept.
Say you design a language called Foobar, and build a great runtime for it. It gets wildly popular and you even release it as GPL. Google decides that it really likes the language, and really needs all those developers who use it, but "screw you and screw the GPL": they implement a brand new runtime that will run your Foobar programs mostly unchanged. They refuse to acknowledge any of your work and will not work with you on anything -- it's not Foobar anymore, it's "Googlebaz".
Wouldn't you be a bit pissed off? Isn't that actually copying stuff?
Unfortunately, this sort of "emulation" is often necessary in our industry, and for the best of reasons; now, just because Google acted like total assholes (note how they made their choice when Sun was still alive -- I strongly doubt they would have done it with Ellison on the other side), we risk losing the right to it.
How did Google act like assholes? Android has the same Apache license as the base class and other library implementations they used, which had been open source for many years at that time.
I am not assuming the answer to what's being litigated. I am pointing out that if Oracle has an IP interest in something inside an Android device, it also has exceptionally dangerous and disruptive IP claims all over the place, including things like Scala and all code written in the Java language.
Regarding your example about runtime environments: No, I don't see how inventing a language, which is generally not protectable, and someone inventing a different runtime for a language should make anyone think their IP has been ripped off.
2. As was noted in various trials (Oracle, MS antitrust) "Java" is not one thing, it's multiple. It can refer to the language, the dev environment, the runtime or the API. This case is specifically about the API. Similarly, Android is not just the software that runs on the handset, it also includes the dev environment needed to make that software.
So Oracle should go after Eclipse and other IDEs? Or Apache Harmony?
> *" This completely undermines some of Google's legal arguments as a defense for API copyrightability, because the cases they cited (Sega etc) explicitly held that copying certain things was OK for maintaining binary interoperability."
That binary compatibility is required by javac, which is in the toolchain. Then the Java bytecode gets translated to Dalvik bytecode and packaged as an APK.
One way translation of byte code is not interoperability. Android apps cannot interoperate with "real" Java apps.
In order to do that, in pure Ellison style, they will use any tactic, no matter how ruthless or disparaging. Ellison does not care about style or manners, he only cares about winning and cashing in; in many ways he's very similar to his great friend Jobs (who did not care about anything but his products, basically).
(Not that Google comes out squeaky-clean, from all this. They clearly piggybacked on the java ecosystem while explicitly avoiding to grant any attribution...)