10 ms·

It's weird how uncommon the knowledge that Google lost the lawsuit is. They lost it in appeal at the Federal Circuit. There is a remaining "fair use" decision f

by guelo 11y ago

It's weird how uncommon the knowledge that Google lost the lawsuit is. They lost it in appeal at the Federal Circuit. There is a remaining "fair use" decision from the lower courts but Google doesn't have much hope there. Google's current approach using Harmony's clean-room API implementation has been basically deemed illegal. Google will likely have to pay millions if not billions to Oracle for copyright infringement. And they will continue to be infringing as long as they are distributing their current Android systems. P.S. The law of the land is now that when you create an API it is copyrighted.

curt15 11y ago
>P.S. The law of the land is now that when you create an API it is copyrighted. No, it's just the law of the Federal Circuit and only when interpreting 9th circuit law. Their decision is not binding precedent on any of the usual circuits, and their usual domain is patent law, not copyright law.
- bitwize 11y ago
  No. The Federal Circuit has subject matter jurisdiction, not geographic jurisdiction. Their decisions are binding precedent throughout the United States.
  - rayiner 11y ago
    They don't have subject matter jurisdiction over copyright. Their decision on a copyright case isn't binding in other circuits.
    - josaka 11y ago
      I think that is correct. This case is a little misleading because there were both patent and copyright claims at the district court level. As a result of the patent claims, the Federal Circuit had jurisdiction over the entire case, even though the copyright claims were to be adjudicated under 9th Circuit precedent. If they had only brought copyright claims, this appeal would have been heard at the 9th Circuit Court of Appeals.
Buge 11y ago
But this article seems to be saying that Oracle's version is GPLed. If that's the case, then how can Oracle sue Google for copyright infringement? Is it for violating the GPL?
- nl 11y ago
  The lawsuit has (mostly) nothing to do with the license the implementation is under. Oracle's argument is that an API is copyrightable, and cloning an API is illegal. The only known way to obtain a license to use those APIs is to use the GPLed Java implementation of those APIs.
  - superuser2 11y ago
    >use the GPLed Java implementation of those APIs Google did a clean-room implementation. But what about derivatives? Obviously you have to redistribute your source. Does using your patched version magically become legal iff your changes are merged into the upstream? This seems to defeat the GPL entirely, since it doesn't matter that you have a license to modify the source code if doing so means you are violating copyright on the APIs.
    - nl 11y ago
      This seems to defeat the GPL entirely, since it doesn't matter that you have a license to modify the source code if doing so means you are violating copyright on the APIs. Indeed. IANAL, but that is one (of many) problems with this legal outcome. Although it's possible that the explicit permission to change the code in the GPL would be interpreted as a license the the API as well.
      - icebraining 11y ago
        Although it's possible that the explicit permission to change the code in the GPL would be interpreted as a license the the API as well. I don't see how it can be any other way; since the infringement was based on the copyright of the API, by having a legal license to the copyright of the code that implements the API, I don't see how can one lack a license to the API. It'd be different if the API was somehow patented, but copyright explicitly covers the actual implementation.
rodgerd 11y ago
> P.S. The law of the land is now that when you create an API it is copyrighted. I look forward to the extinction WINE and PostgreSQL, and the resurrection of the SCO corpse for another tilt at Linux vendors, he said sarcastically.
- riffraff 11y ago
  > PostgreSQL why, which API does that infringe?
  - dtech 11y ago
    SQL. Time for IBM to earn some well-earned royalties from all the current implementers.
    - jtome 11y ago
      Starting with oracle hopefully. (Un)fortunately (and IANAL) I don't think copyrights are enforceable if you don't enforce them everytime you know of a violation.
      - lftl 11y ago
        Nope, that's trademarks.
      - rms_returns 11y ago
        That won't happen because IBM also uses a lot of Java stuff including the great eclipse infrastructure. They will end up suing each other for a lot of components if that happens. If rumors are to be believed, Oracle and IBM have some sort of "agreement" that they won't sue each other on these matters.
        bcg1 11y ago
        I would venture a guess that IBM is a paid up licensee of Java, they've been releasing their own proprietary JVM for quite a while and never been sued by Sun/Oracle for it AFAIK
pron 11y ago
To make a few more things clear: that APIs are not copyrighted has never been settled law either, so the chances of getting sued over copyright violation has always existed. The only difference is that before (as Google did) you could make two arguments: 1) that APIs aren't copyrightable to begin with, and 2) if they are, then you're making fair use of them. Now only option 2 exists (with the caveat of the ruling serving as a precedent, as other commenters say). Also, copyrighted APIs are only "language-level" APIs (or just APIs as we've always called them) -- not web APIs (or protocols, as we used to call them), simply because web APIs lack other prerequisites to be copyrightable. To the question of fair use. There are several fair use defenses, but two apply to the industry in cases like this: 1/ non-profit use, 2/ use for the sake of compatibility (to prevent vendor lock-in). AFAIK, all third-party implementations of APIs happen to be one or the other (or both). Android is neither -- it is for profit and it is not compatible with Java programs. More than that: it wasn't even done to prevent vendor lock-in. So even if Google loses on the fair-use case, chances are that the defense still holds for all other cases of third-party implementation. Unlike the claims made by Google's PR, this ruling's effect on the industry are at worst tiny, and most likely nil.
- magicalist 11y ago
  > with the caveat of the ruling serving as a precedent, as other commenters say No, this is incorrect. Because it was a copyright case decided by the Federal Circuit, it is precedent nowhere. Literally no court will need to treat it as precedent. Please educate yourself further before making these lengthy posts in threads like this; you're spreading misinformation. > Also, copyrighted APIs are only "language-level" APIs... not web APIs (or protocols, as we used to call them) From this and your post below, this appears to be a distinction you have made up (to appeal to the audience here? I have no clue). "Structure, sequence and organization" are certainly aspects of web APIs that could be copyrightable under the Federal Circuit's ruling. Also, no one called web APIs "protocols" anymore than they called system calls "protocols". Unless you're talking about actual protocols, in which case I'd just stick to calling them "protocols" because that's what we still call them. > There are several fair use defenses, but two apply to the industry in cases like this: 1/ non-profit use, 2/ use for the sake of compatibility (to prevent vendor lock-in). Non-profit use is neither necessary or sufficient for fair use (see Google's recent book scanning ruling, for instance), and the compatibility argument is not with existing JVMs but with existing Java source code. As for the rest of you post on fair use, I would suggest other readers to take a look at the actual four factors of fair use in the US, as your summary is a poor one. One good source: http://fairuse.stanford.edu/overview/fair-use/four-factors/ http://fairuse.stanford.edu/overview/fair-use/four-factors/
rms_returns 11y ago
> The law of the land is now that when you create an API it is copyrighted. The US is the only strange land where such a law could be even comprehended. Asserting API copyright to someone on the basis of accusation of copying nine lines of code[1] indicates that a jury isn't in its right mind. 1. https://news.ycombinator.com/item?id=3940683 https://news.ycombinator.com/item?id=3940683
- _asummers 11y ago
  Nine lines of extremely trivial code; one would be hard pressed to write it any other way. The judge from the initial trial basically yelled at the Oracle attorneys saying that it was code that a highschooler could have written.
- throwawaykf05 11y ago
  HN is a strange forum, where ostensibly intelligent people will post uninformed, sensationalistic comments despite the presence of multiple earlier comments in the same thread by better-informed people explaining things clearly. To wit: https://news.ycombinator.com/item?id=10848716 https://news.ycombinator.com/item?id=10848716 (While I am not sure it nails all the nuances exactly, especially around the distinction between web and language APIs, it is an infinitely better account of the situation than what I've seen in any tech media so far. Golly, I wonder why.) Of course, it doesn't help that the better informed comment is also down-voted, making it harder to find.
  - rms_returns 11y ago
    No doubt, the link you have provided has some merit and explains a lot about the Oracle's side of coin. But after reading a lot about this story from all different angles, my final impression is that no organization is all too saintly or 100% honest, and neither are any altogether guilty. In fact, the thing about organizations is that they are multi-faced and do a lot of complex things and its just impossible to keep track of all those details. Considering all this, I feel that the only important thing in technology is that concentration of power shouldn't be in the hands of one/few tech companies, be it Oracle or Google. Rather, there should be checks and balances everywhere. Ideally, all software should be produced by communities where all coders are chosen strictly on merit every commit passes through the scrutiny of all members.
rms_returns 11y ago
> P.S. The law of the land is now that when you create an API it is copyrighted. Oracle has stolen an "r" from the API specification of my full name. How/where can I sue them?
- imtringued 11y ago
  You can sue anyone for anything. That doesn't mean you will be successful.
  - rms_returns 11y ago
    Then how come Oracle did it successfully?
    - pjmlp 11y ago
      Because Google was playing dirty after not wanting to pay Sun what they were asking for.
      - rms_returns 11y ago
        Why should they pay for using something that is free and open source?
        pjmlp 11y ago
        Because back then it was only free when targeting desktop and servers, not embedded devices. Also the license requires that all of Java is taken, not cutting down pieces as they feel like.
invalidname 11y ago
It won most counts and lost on the API's aren't copyrightable stance. The thing is that even if API's are copyrightable what is the damage if someone does a clean room implementation... Since Oracle/Sun is effectively giving OpenJDK for free I'd argue the damages are effectively 0 USD so Oracle would win a 0 USD compensation. Since the lower court justice mostly sided with Google I'm guessing compensation won't be much for Oracle.

by guelo 11y ago

curt15 11y ago

>P.S. The law of the land is now that when you create an API it is copyrighted. No, it's just the law of the Federal Circuit and only when interpreting 9th circuit law. Their decision is not binding precedent on any of the usual circuits, and their usual domain is patent law, not copyright law.

bitwize 11y ago
No. The Federal Circuit has subject matter jurisdiction, not geographic jurisdiction. Their decisions are binding precedent throughout the United States.
- rayiner 11y ago
  They don't have subject matter jurisdiction over copyright. Their decision on a copyright case isn't binding in other circuits.
  - josaka 11y ago
    I think that is correct. This case is a little misleading because there were both patent and copyright claims at the district court level. As a result of the patent claims, the Federal Circuit had jurisdiction over the entire case, even though the copyright claims were to be adjudicated under 9th Circuit precedent. If they had only brought copyright claims, this appeal would have been heard at the 9th Circuit Court of Appeals.

Buge 11y ago

But this article seems to be saying that Oracle's version is GPLed. If that's the case, then how can Oracle sue Google for copyright infringement? Is it for violating the GPL?

nl 11y ago
The lawsuit has (mostly) nothing to do with the license the implementation is under. Oracle's argument is that an API is copyrightable, and cloning an API is illegal. The only known way to obtain a license to use those APIs is to use the GPLed Java implementation of those APIs.
- superuser2 11y ago
  >use the GPLed Java implementation of those APIs Google did a clean-room implementation. But what about derivatives? Obviously you have to redistribute your source. Does using your patched version magically become legal iff your changes are merged into the upstream? This seems to defeat the GPL entirely, since it doesn't matter that you have a license to modify the source code if doing so means you are violating copyright on the APIs.
  - nl 11y ago
    This seems to defeat the GPL entirely, since it doesn't matter that you have a license to modify the source code if doing so means you are violating copyright on the APIs. Indeed. IANAL, but that is one (of many) problems with this legal outcome. Although it's possible that the explicit permission to change the code in the GPL would be interpreted as a license the the API as well.
    - icebraining 11y ago
      Although it's possible that the explicit permission to change the code in the GPL would be interpreted as a license the the API as well. I don't see how it can be any other way; since the infringement was based on the copyright of the API, by having a legal license to the copyright of the code that implements the API, I don't see how can one lack a license to the API. It'd be different if the API was somehow patented, but copyright explicitly covers the actual implementation.

rodgerd 11y ago

> P.S. The law of the land is now that when you create an API it is copyrighted. I look forward to the extinction WINE and PostgreSQL, and the resurrection of the SCO corpse for another tilt at Linux vendors, he said sarcastically.

riffraff 11y ago
> PostgreSQL why, which API does that infringe?
- dtech 11y ago
  SQL. Time for IBM to earn some well-earned royalties from all the current implementers.
  - jtome 11y ago
    Starting with oracle hopefully. (Un)fortunately (and IANAL) I don't think copyrights are enforceable if you don't enforce them everytime you know of a violation.
    - lftl 11y ago
      Nope, that's trademarks.
    - rms_returns 11y ago
      That won't happen because IBM also uses a lot of Java stuff including the great eclipse infrastructure. They will end up suing each other for a lot of components if that happens. If rumors are to be believed, Oracle and IBM have some sort of "agreement" that they won't sue each other on these matters.
      - bcg1 11y ago
        I would venture a guess that IBM is a paid up licensee of Java, they've been releasing their own proprietary JVM for quite a while and never been sued by Sun/Oracle for it AFAIK

pron 11y ago

To make a few more things clear: that APIs are not copyrighted has never been settled law either, so the chances of getting sued over copyright violation has always existed. The only difference is that before (as Google did) you could make two arguments: 1) that APIs aren't copyrightable to begin with, and 2) if they are, then you're making fair use of them. Now only option 2 exists (with the caveat of the ruling serving as a precedent, as other commenters say). Also, copyrighted APIs are only "language-level" APIs (or just APIs as we've always called them) -- not web APIs (or protocols, as we used to call them), simply because web APIs lack other prerequisites to be copyrightable. To the question of fair use. There are several fair use defenses, but two apply to the industry in cases like this: 1/ non-profit use, 2/ use for the sake of compatibility (to prevent vendor lock-in). AFAIK, all third-party implementations of APIs happen to be one or the other (or both). Android is neither -- it is for profit and it is not compatible with Java programs. More than that: it wasn't even done to prevent vendor lock-in. So even if Google loses on the fair-use case, chances are that the defense still holds for all other cases of third-party implementation. Unlike the claims made by Google's PR, this ruling's effect on the industry are at worst tiny, and most likely nil.

magicalist 11y ago
> with the caveat of the ruling serving as a precedent, as other commenters say No, this is incorrect. Because it was a copyright case decided by the Federal Circuit, it is precedent nowhere. Literally no court will need to treat it as precedent. Please educate yourself further before making these lengthy posts in threads like this; you're spreading misinformation. > Also, copyrighted APIs are only "language-level" APIs... not web APIs (or protocols, as we used to call them) From this and your post below, this appears to be a distinction you have made up (to appeal to the audience here? I have no clue). "Structure, sequence and organization" are certainly aspects of web APIs that could be copyrightable under the Federal Circuit's ruling. Also, no one called web APIs "protocols" anymore than they called system calls "protocols". Unless you're talking about actual protocols, in which case I'd just stick to calling them "protocols" because that's what we still call them. > There are several fair use defenses, but two apply to the industry in cases like this: 1/ non-profit use, 2/ use for the sake of compatibility (to prevent vendor lock-in). Non-profit use is neither necessary or sufficient for fair use (see Google's recent book scanning ruling, for instance), and the compatibility argument is not with existing JVMs but with existing Java source code. As for the rest of you post on fair use, I would suggest other readers to take a look at the actual four factors of fair use in the US, as your summary is a poor one. One good source: http://fairuse.stanford.edu/overview/fair-use/four-factors/ http://fairuse.stanford.edu/overview/fair-use/four-factors/

rms_returns 11y ago

> The law of the land is now that when you create an API it is copyrighted. The US is the only strange land where such a law could be even comprehended. Asserting API copyright to someone on the basis of accusation of copying nine lines of code[1] indicates that a jury isn't in its right mind. 1. https://news.ycombinator.com/item?id=3940683 https://news.ycombinator.com/item?id=3940683

_asummers 11y ago
Nine lines of extremely trivial code; one would be hard pressed to write it any other way. The judge from the initial trial basically yelled at the Oracle attorneys saying that it was code that a highschooler could have written.
throwawaykf05 11y ago
HN is a strange forum, where ostensibly intelligent people will post uninformed, sensationalistic comments despite the presence of multiple earlier comments in the same thread by better-informed people explaining things clearly. To wit: https://news.ycombinator.com/item?id=10848716 https://news.ycombinator.com/item?id=10848716 (While I am not sure it nails all the nuances exactly, especially around the distinction between web and language APIs, it is an infinitely better account of the situation than what I've seen in any tech media so far. Golly, I wonder why.) Of course, it doesn't help that the better informed comment is also down-voted, making it harder to find.
- rms_returns 11y ago
  No doubt, the link you have provided has some merit and explains a lot about the Oracle's side of coin. But after reading a lot about this story from all different angles, my final impression is that no organization is all too saintly or 100% honest, and neither are any altogether guilty. In fact, the thing about organizations is that they are multi-faced and do a lot of complex things and its just impossible to keep track of all those details. Considering all this, I feel that the only important thing in technology is that concentration of power shouldn't be in the hands of one/few tech companies, be it Oracle or Google. Rather, there should be checks and balances everywhere. Ideally, all software should be produced by communities where all coders are chosen strictly on merit every commit passes through the scrutiny of all members.

rms_returns 11y ago

> P.S. The law of the land is now that when you create an API it is copyrighted. Oracle has stolen an "r" from the API specification of my full name. How/where can I sue them?

imtringued 11y ago
You can sue anyone for anything. That doesn't mean you will be successful.
- rms_returns 11y ago
  Then how come Oracle did it successfully?
  - pjmlp 11y ago
    Because Google was playing dirty after not wanting to pay Sun what they were asking for.
    - rms_returns 11y ago
      Why should they pay for using something that is free and open source?
      - pjmlp 11y ago
        Because back then it was only free when targeting desktop and servers, not embedded devices. Also the license requires that all of Java is taken, not cutting down pieces as they feel like.

invalidname 11y ago

It won most counts and lost on the API's aren't copyrightable stance. The thing is that even if API's are copyrightable what is the damage if someone does a clean room implementation... Since Oracle/Sun is effectively giving OpenJDK for free I'd argue the damages are effectively 0 USD so Oracle would win a 0 USD compensation. Since the lower court justice mostly sided with Google I'm guessing compensation won't be much for Oracle.