Linked by Thom Holwerda on Thu 12th May 2016 23:26 UTC
Android

During hours of unrelenting cross-examination today, Andy Rubin, Google's former Android chief, was on the stand in the Oracle v. Google trial defending how he built the mobile OS.

Rubin's testimony began yesterday. He's another one of the star witnesses in this second courtroom showdown between the two software giants in which Oracle has said it will seek up to $9 billion in damages for Google's use of certain Java APIs in the Android operating system. Since an appeals court decided that APIs can be copyrighted, Google's only remaining defense in this case is that its use of those APIs constitutes "fair use."

The "API's are copyrightable"-ruling is one of those rulings we will look back on decades from now and point to as "that's where it all went wrong", much like how we now look back upon disastrous rulings like Citizens United or the slew of bad rulings that legitimised software patents.

And we have the despicable Oracle to thank for that. As I've pointed out before, it's no coincidence that the three-pronged legal attack on Android - from Apple, Microsoft, and Oracle - all started at around the same time, and that Larry Ellison was a very close friend of Steve Jobs.

When all this stuff hits the fan even harder, you know who to thank.

Thread beginning with comment 628768
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[2]: Comment by kittynipples
by kittynipples on Sun 15th May 2016 15:24 UTC in reply to "RE: Comment by kittynipples"
kittynipples
Member since:
2006-08-02

So you are taking a position that a header file is not a creative work but the corresponding source file is. Clearly the court views them to both be creative in nature, and thus covered by current copyright law. Whether or not it is inconvenient for the software industry is probably immaterial.

Reply Parent Score: 2

RE[3]: Comment by kittynipples
by Alfman on Sun 15th May 2016 16:39 in reply to "RE[2]: Comment by kittynipples"
Alfman Member since:
2011-01-28

kittynipples,

So you are taking a position that a header file is not a creative work but the corresponding source file is. Clearly the court views them to both be creative in nature, and thus covered by current copyright law. Whether or not it is inconvenient for the software industry is probably immaterial.



To the extent that you can be creative in a header file (ie with comments and whatnot), then that can be copyrighted. But to the extent that it's impossible to be creative because sharing the same function prototypes is technically necessity to be compatible, then absolutely not - it was never the purpose of copyright to block competing implementations, which is exactly what oracle is trying to do.

It's not just my position, prior to this expansion of copyright scope by the appellate court, it was such a legally unanimous position that lawyers themselves did not bother to explicitly license APIs. Not only did the courts themselves rule that APIs were not copyrightable, but even the creators of Java that this case is based on agree too. Let the ridiculousness of this situation sink in for a moment, because the success of this legal case will set into motion a lot more like it.

Reply Parent Score: 2

galvanash Member since:
2006-01-25

So you are taking a position that a header file is not a creative work but the corresponding source file is.


I personally think, after numerous discussions with different people, that this is where the fundamental misunderstanding always pops up. There is always this focus on details that simply do not matter. It has nothing to do with header files - header files are not APIs...

I struggle to hunt for an explanation that truly makes sense... People (understandably) want a simple analogy, one they can "get" immediately, for what makes an API special compared to source code when it comes to copyright protection.

I don't have one of those, sorry. ;)

This is the best I have managed so far:

Lets say you invent a new mouse trap. It is extremely effective. You decide to patent the mouse trap, so no one can copy your design.

It is rather complex and requires detailed instructions as to how it works, but you do your best to simplify its operation. You write a manual for it, explaining how to operate it. You copyright your manual so no one can make a copy of the manual without your consent.

Now someone else invents what they think is a better mousetrap. They have no idea how your mousetrap works (they never saw or read the patent), their design is quite different. It's rather complex too, but they realize that if they try and make their mousetrap's operation similar to yours, it will be to everyone's benefit because most people already know how your works.

So they make the operation of their mouse trap identical to yours (not its fundamental design, not how it catches mice, simply the way users interact with it). They of course read your documentation in order to do so, but they are not copying the documentation - copyright does not affect them at all.

They did copy something though... That "thing" they copied - that is an API. It is simply a "method of operation". It is facts contained within the documentation of the operators manual - it is neither the work itself nor the documentation and is not protected through existing IP law (at least not until this stupid appellate ruling).

No one copied the original mousetrap's design, and no one copied the original mousetrap's manual. The key point to understand is that the documentation (the copyrighted thing) does not describe the details of the mousetrap (the patented thing). How to operate the mousetrap is protected by neither copyright or patent, it is simply factual information about how to interact with it.

In the software world the source code is copyrighted. That is your implementation, your "desgin" per se. Copyright protects it from being copied, much like patents for methods. The implementation is your "mousetrap". Patents are (unfortunately) something you can get as well in some situations, but imo copyright is enough. Regardless, neither has anything to do with API protection in this case.

Header files are documentation about an API, not APIs in and of themselves. You can copyright them, but they contain facts, things which are ineligible for and do not rise to the level required to be copyrighted. Things like method names and such, parameter names, types, etc. - these are all things universally agreed upon to be exempt from copyright protection - even in the Oracle vs Google case. They have nothing at all to do with your mousetrap, they are simply facts about your mousetrap.

Oracle already lost the case concerning the copying of method names and whatnot. What they won, however, was the argument that the combination of all of these individual facts together, constitutes a creative work in and of itself.

I personally think this argument is total bunk. It is both destructive to how the software industry functions and unnecessary - software is already afforded plenty of protection though the existing understanding of copyright and patents (too much even).

Anyway, Google copied the API of Java (at least in part) - no one, not even them, is claiming otherwise. As Oracle argued, they didn't have to, but they did - because similarly to the mousetrap above, it let them leverage existing programmers familiarity with Java.

The thing is... That is a good thing. It did not hurt Sun (now Oracle). On the contrary, it helped them by promoting the use of their language. It is almost always a good thing, for both users of software products, producers of software products, and the economy in general. No one is "stealing" anyone's work when an API is copied, what they are doing is leveraging existing learned behavior, something that may seem at odds with intellectual property rights to some, but to people in the software world leveraging learned behavior is almost second nature. Software is terribly complicated stuff - if companies don't try and converge towards common interaction models it all degenerates into chaos.

That is why APIs should not be eligible for copyright, because it will actual encourage everyone to build stuff that works differently, instead of building stuff that works similarly. Familiarity is the only thing we have working in our favor as software developers - computers are getting more and more complex as time goes by, but because of convergent evolution it is (barely) manageable. Copyrighting APIs will start encouraging divergent evolution, and we will have chaos on our hands - hell on earth for developers. Everything will work differently, and for no good reason.

Its stupid. f--k Oracle.

Edited 2016-05-16 00:32 UTC

Reply Parent Score: 2