Supreme Court May Decide the Fate of APIs (But Also Klingonese and Dothraki)

New submitter nerdpocalypse writes: In a larger battle than even Godzilla v. Mothra, Google v. Oracle threatens not only Japan but the entire nerd world. What is at stake is how a language can be [copyrighted]. This affects not just programming languages, APIs, and everything that runs ... well ... everything, but also the copyright status of new languages such as Klingon and Dothraki.

Comments

Re:SFLC's brief explains parts of this well

[phantomfive]

What I gather is the linked article at Slate seems more ominous than this sounds.

The Slate article is sensationalistic, like most news these days, designed to make people feel outrage.

From the brief, I understand that Google took code and changed it (which may or may not be against the terms) and then redistributed it (via the SDK? the Android OS?) for developers to use.

Java was designed with the intention of every implementation to be 100% compatible.....the same APIs, the same runtime, etc. When Google made Android, they used an incompatible version of Java (mainly changing UI code, but also the runtime and some other things). They used Java to make things easier for developers.

So now the question is whether it is allowed for Google to do that, or if Oracle has the right to stop them. This isn't about whether you are violating copyright if you use a language, and that's one area the Slate article seems confused.

It is also not about whether you can use an API for interoperability purposes. That is covered by fair use, and will still be covered by fair use, no matter how this case is decided. But interoperability not why Android used Java.......Android used Java to make it easier for developers to write programs for their system.

I had to laugh when I read this...

[mark-t]

Oracle, the appeals court, and the Department of Justice think that owning a copyright on a language isnâ(TM)t a big deal, because others can always invent new words rather than copying

Of course, the enormous flaw in this notion is that if you always have to invent new words simply to avoid copying anyone else, then nobody else is actually going to understand you. This is particularly true for things like spoken or written languages, but carries an element of truth to it for computer languages as well. While inventing any language can arguably already be challenging, inventing one that other people will actually use or adopt is usually either a function of the notoriety of the inventor, a matter of blind luck, or some combination of the two.

Since it can be argued that copyright provides a mechanism for otherwise possibly unknown artists to publish their works on the same relatively level playing field as those who may have already gained some notoriety, it seems I think that suggesting that such things should somehow be copyrightable is even at best wholly counter-productive.

Re:I had to laugh when I read this...

[silentcoder]

Wrong.
  The GPL prevents linking via API to existing GPL'd libraries. It does not stop you from writing your own library with the exact same method declarations but your OWN implementation.
What google did is specifically NOT prevented by the GPL either.

The GPL focusses on linking because that means MY implementation is used by your code. If you write your own library with the same declarations and your own implementation - then even though your application code is unchanged, I no longer hold a claim.

Indeed most of GNU's libraries (both those under the GPL like readline and those under the LGPL like glibc) could not have existed if Oracle is right - since they were mostly re-implementations of long-existing APIs that every Unix OS ever developed also included. LibC in particular - EVERY unix since the very first Bell Labs one has had a generic C library - and they all implemented essentially the same core set of functions. Their APIs are all virtually identical yet they were all deemed legal and all under their own distinct copyrights. Some were proprietory, some were BSD licensed (i.e. the libc's in every BSD today) and GNU made theirs LGPLd.

A better example would be the wine project. Wine reimplemented just about the entire windows API - all the calls are identical - so identical that you can run windows programs and games with Wine - but every implementation written from scratch as a clean-room reverse-engineer process. That has been legal for many, many years -this case threatens that. It would make it possible for Microsoft to get Wine declared illegal.

If this is illegal Wine would actually be MORE illegal since it is much more compatible with the original API than Android is with Java. In fact that is Oracle's entire PR about this matter: that they are trying to sue google for NOT making the API ENTIRELY compatible with theirs !

Disclaimer: I am a former Oracle engineer, I quite my job because I could not in good conscience keep working for the company that filed this suit !

If I had to guess

[MikeRT]

At least Scalia, Thomas and Alito will hammer Oracle. They tend to be very antagonistic to arguments like this. In Kelo v. New London, which was a similar abuse of intent in the law (5th amendment there), they wrote scathing dissents. Allowing APIs to be copyrighted is like allowing technical jargon (that's not trademarked) to be copyrighted. They fall dangerously close to the list of things the Copyright Office says are not covered by law.

Part of this makes me wonder if this isn't a "heads we win, tails you lose" scenario for Oracle. If they win, they get to badly hurt Google. If they lose, there's a Supreme Court precedent that allows them to clone any small competitor's products (patent considerations notwithstanding) at a 100% API compatible level and use Oracle integration and consulting to ram them out of business. It smells like a Larry Ellison strategy.