Google's strict policies for 'open' Android OS revealed in newly public documents

15791011

Comments

  • Reply 121 of 206
    solipsismxsolipsismx Posts: 19,566member
    That depends on where you live. An algorithm can describe anything from a+b to an incredibly complex ten million loc function.

    I guess I'm not understanding what an API is if both algorithms and code can be protected but APIs cannot.
  • Reply 122 of 206
    Quote:

    Originally Posted by SolipsismX View Post





    I guess I'm not understanding what an API is if both algorithms and code can be protected but APIs cannot.

     

    class Person

      has Name

      has Age

      has Location

      func getName(int which) str

      func getLocation() geoip

     

    This sort of thing is pretty much what an API is. Lets say I develop a mobile location library that provides your new app with better location services. My library needs to provide you with various object types you can create, and what attributes and functions they might have (this changes depending on the programming language's features too)

     

    The actual code that implements the getLocation function or similar is unrelated to this, because Oracle is not claiming Google copied that. What they are claiming is that by using the same names and data types that the infringement is against the Structure, Sequence and Organisation of the API, ie that the API itself is sufficiently original and substantive to qualify for copyright protection.

     

    To a programmer such as myself, this is ridiculous, because as you can see by the class above, very often the names and data types are dictated simply by the obvious or limited technically (a string is a relatively fundamental datatype for any language or API)

     

    To make this an even worse case, this is Java, which until Oracle's acquisition was reimplemented seemingly with Sun's explicit blessing and was supposed to be a universal language that could be reimplemented to any other computer by simply using the same, standard API.

     

    There's obviously even more complexity when it comes down to the actual arguments in court and the finer points of the law, but from my perspective as a multi platform developer, APIs alone being copyrightable is absolute nonsense.

  • Reply 123 of 206
    Quote:
    Originally Posted by SolipsismX View Post





    Aren't algorithms copyrightable?

    The algorithm itself is an abstract mathematical formula, not something fixed in any tangible medium as required by copyright law. One can't copyright the quicksort algorithm any more than one can copyright a method for multiplying two numbers, or a method for solving a matrix equation. What one can copyright is a particular implementation of an algorithm in code.

  • Reply 124 of 206
    solipsismxsolipsismx Posts: 19,566member
    d4njvrzf wrote: »
    What one can copyright is a particular implementation of an algorithm in code. 

    That seems tricky since using cubeVolume = pow( xValue, 3); wouldn't be exactly the same as using cubeVolume = xValue * xValue * xValue;. How different would the code have to be to not be considered in violation of the copyright?
  • Reply 125 of 206
    Quote:
    Originally Posted by ItsTheInternet View Post

     

    I know that no serious developer I have ever met has held that APIs are copyrightable, 




    Even if this were true (which it obviously isn't) since when are software developers arbiters of copyright law? 

     

    Quote:


     

    even less so the SSO of APIs. It's nonsense and I speak as a professional developer.





    Calling it nonsense is only admitting you don't understand it.

     


    Quote:
    Oracle claims that their API structure itself is copyrightable. By duplicating my getTime function with identical arguments and return values, you have infringed on the structure of my API despite not copying any code (if APIs were copyrightable)

     

    Nope. You confuse functionality with expression. The functionality is a noncopyrightable element. The original expression in each of the 37 API packages is a copyrightable element. The same functionality can be achieved by a wide variety of original expression written in the code. While the functionality of the Java APIs is not protected, the precise way Sun/Oracle wrote that functionality is. It would only be noncopyrightable if there were only one or a few ways to write that functionality, but the choices were near-infinite. Google copied 7000 lines of declaring code, verbatim. They admit they copied all that code verbatim, it is not in dispute. They want that code not to be copyrightable. The Java API packages could have been written in a million/infinite different ways, but Sun/Google created their own original expression from the infinite possibilities. Google had their own infinity of options for achieving the same functionality, but chose to copy Sun/Oracle's original expression instead. 

  • Reply 126 of 206
    Quote:

    Originally Posted by ruddy View Post

     

    Calling it nonsense is only admitting you don't understand it.


     

    I explained it in detail above, so please see that reply. I can quote Judge Alsup's summation in which he goes through the legal framework for your answer and disagrees with you. I think it pretty unlikely that ultimately APIs will be covered by copyright, even the SSO. The people on the other side of this fight are Microsoft and Oracle, hardly the bastions of open and fair competition.

  • Reply 127 of 206

    Alsup's first epic fail is in elevating interoperability, an argument which has occasionally been used for a fair use defense of infringement, to be an excuse to deny copyrightability. He simply made this up. There is no precedent for it, and it's an absurdity that neither the CAFC or SCOTUS are going to seriously entertain. Did you not listen to the oral arguments made at the CAFC hearing? The judges basically all asked the Google lawyer, "WTF is this bullshit?"

     

    Alsup's ruling has great appeal for those of a certain open ideology, and he did his best to concoct a theory favoring that point of view. Trouble is it runs against all the established principles and precedent of copyright and isn't going to have any legs beyond the original trial. 

  • Reply 128 of 206
    gatorguygatorguy Posts: 24,176member
    ruddy wrote: »
    etc.

    Curious...
    Are you an attorney, and if so specializing in patent law? So many of your posts are painting legal questions in black and white that you leave the impression you have several years of experience with copyright and patent law. Are you claiming there's no question that Google will be found guilty of copyright infringement and subject to a possible injunction?
  • Reply 129 of 206
    Quote:

    Originally Posted by ruddy View Post

     

    Alsup's first epic fail is in elevating interoperability, an argument which has occasionally been used for a fair use defense of infringement, to be an excuse to deny copyrightability. He simply made this up. There is no precedent for it, and it's an absurdity that neither the CAFC or SCOTUS are going to seriously entertain. Did you not listen to the oral arguments made at the CAFC hearing? The judges basically all asked the Google lawyer, "WTF is this bullshit?"

     

    Alsup's ruling has great appeal for those of a certain open ideology, and he did his best to concoct a theory favoring that point of view. Trouble is it runs against all the established principles and precedent of copyright and isn't going to have any legs beyond the original trial. 


     

    I did not listen to the oral arguments and I am not a lawyer nor a US citizen so I am particularly removed from this case. I am speaking only in terms of what is reasonable to a developer. The idea that a relatively simple list of names and attributes gains copyright protection not only over the names but the organisation of the names is ridiculous to me. Such things are mere boilerplate, named as simply and directly as possible. There's certainly an element of creativity at work but it pales in comparison to the actual code that implements a method or any other typically copyrightable expression.

     

    There is little benefit to any developer of any ideology from copyright protection of this code. The only benefit is to organisations which are falling back towards patent trolling as a business model. Let's not forget that Sun originally wrote this code and their intentions for it were absolutely clear. Oracle may have gained legal rights over the code through their purchase but even if you are correct this does not mean that such a ruling would be reasonable or correct. I retain some faith in courts to be able to discern this if the law is insufficiently precise and it does seem like the courts are given leeway to rule here.

  • Reply 130 of 206
    If APIs are found to be copyrightable, wouldn't the whole concept of a software compatibility layer (http://en.wikipedia.org/wiki/Compatibility_layer) be on shaky legal ground? After all, a compatibility layer is nothing but a clean-room implementation of a foreign API. Could Linux or the BSDs be in legal trouble for reimplementing the Unix API? If the GPL copyright protections extended to APIs, would commercial implementations of Linux APIs be considered derivative works and therefore also be subjected to the GPL?
  • Reply 131 of 206
    gatorguygatorguy Posts: 24,176member
    d4njvrzf wrote: »

    Even if the Appeals Court finds API's to ultimately be available for copyright protection it doesn't mean Google or anyone else is automatically infringing by using any of them. Then the question of Fair Use has to be considered. For example there's no doubt that literary works are protected by copyright. At the same time it's perfectly legal to quote snippets, even fairly large excerpts, from a book or article in offered web search results, reviews or reports. That's what just happened with the Author's Guild, Google and their academic partners.
  • Reply 132 of 206
    Quote:

    Originally Posted by Gatorguy View Post





    Curious...

    Are you an attorney, and if so specializing in patent law? 

    Nope, this discussion is about Copyright law.

     

    Quote:


    Are you claiming there's no question that Google will be found guilty of copyright infringement


    There is no question that Google has _already_ been found guilty of copyright infringement for the 37 Java APIs. Judge Alsup let Google off the hook by saying, post-trial, that APIs are not copyrightable. If the appeals court overrules him, as it seems will happen given their reactions to oral arguments, then Google will be back on the hook for infringing Java. That is all fact.

     

    Once the appeals court finds that APIs are copyrightable (for the same reasons all other original expression is copyrightable), then the CAFC will either rule on fair use themselves, or send the case back to District Court for a new trial to determine the fair use question, and if necessary, damages.

     

    Google still has a chance to be found not guilty by reason of fair use. But if they lose on fair use too (either by the CAFC or a new jury), then yes, Oracle will have all the grounds they need for an injunction against Android, first in the US and then round the globe by international copyright treaties. It is a very real possibility.

  • Reply 133 of 206
    Quote:

    Originally Posted by ItsTheInternet View Post

     

     

    I did not listen to the oral arguments and I am not a lawyer nor a US citizen so I am particularly removed from this case. I am speaking only in terms of what is reasonable to a developer. 


    What you really mean is developers who share your ideology about restricting copyright's relationship to software. I'm sure the developers responsible for .NET or iOS APIs, as well as many smaller and independent developers have a very different idea of what's reasonable for them. 

     


    Quote:

    The idea that a relatively simple list of names and attributes gains copyright protection not only over the names but the organisation of the names is ridiculous to me. Such things are mere boilerplate, named as simply and directly as possible. There's certainly an element of creativity at work but it pales in comparison to the actual code that implements a method or any other typically copyrightable expression.

    The 37 Java API packages are nothing like a simple list of names and attributes. If they were it would have been trivial for Google to come up with their own. Instead they copied the Java declaring code verbatim for the sole purpose of hijacking Java developers. You believe declaring code and implementing code should be treated differently by copyright law, which is nonsense since they must work together and require each other. In any case copyright law says all original expression is protected, it doesn't value one form of expression more than another. Original expression _is_ the creative element that's protected, copyright doesn't protect the methods or functions. Pretending this is about extending copyright protection to methods or functions is intellectually dishonest.

     

    Quote:

    There is little benefit to any developer of any ideology from copyright protection of this code. 

    Nonsense. you are merely spouting anticopyright dogma.

     

    Quote:


     The only benefit is to organisations which are falling back towards patent trolling as a business model. Let's not forget that Sun originally wrote this code and their intentions for it were absolutely clear. 


     


    Sun's business model was licensing their APIs. It's how they made their money. Funny you should forget that.

     



    Oracle may have gained legal rights over the code through their purchase but even if you are correct this does not mean that such a ruling would be reasonable or correct. I retain some faith in courts to be able to discern this if the law is insufficiently precise and it does seem like the courts are given leeway to rule here.


    You're amazingly optimistic if you think the US Courts share your anticopyright ideology.

  • Reply 134 of 206
    Quote:
    Originally Posted by ruddy View Post

     

    I'm sure the developers responsible for .NET or iOS APIs, as well as many smaller and independent developers have a very different idea of what's reasonable for them. 


    I haven't heard anyone up in arms about the existence of Mono, the clean-room implementation of the .NET APIs which has been around for many years. Microsoft hasn't made any noise. I think the notion that one can prevent others from reimplementing an API, even if they write their own source code, is quite foreign to the software community. Back when SCO was suing IBM and Novell over Linux, the copyright lawsuit was over the actual source code. SCO didn't try to claim appropriation of the Unix API.

  • Reply 135 of 206
    Quote:
    Originally Posted by d4NjvRzf View Post

     

    I haven't heard anyone up in arms about the existence of Mono, the clean-room implementation of the .NET APIs which has been around for many years. Microsoft hasn't made any noise. 


    Why would they? Microsoft made CLI an open standard in order to make projects like Mono possible. If MS hadn't wanted to do that, there would be no Mono today. Different companies employ different business models around their APIs. 

  • Reply 136 of 206
    gatorguygatorguy Posts: 24,176member
    ruddy wrote: »
    Nope, this discussion is about Copyright law.

    There is no question that Google has _already_ been found guilty of copyright infringement for the 37 Java APIs. Judge Alsup let Google off the hook by saying, post-trial, that APIs are not copyrightable. If the appeals court overrules him, as it seems will happen given their reactions to oral arguments, then Google will be back on the hook for infringing Java. That is all fact.

    <span style="line-height:1.4em;">Once the appeals court finds that APIs are copyrightable (for the same reasons all other original expression is copyrightable), then the CAFC will either rule on fair use themselves, or send the case back to District Court for a new trial to determine the fair use question, and if necessary, damages.</span>

    Google still has a chance to be found not guilty by reason of fair use. But if they lose on fair use too (either by the CAFC or a new jury), then yes, Oracle will have all the grounds they need for an injunction against Android, first in the US and then round the globe by international copyright treaties. It is a very real possibility.

    So you don't have a law degree then and no special insight to add? Then you must be depending on the same Mueller article I and many others read for your view of how things will proceed. Your comments seem to mirror his for the most part. IMO Fair Use seems to have a good chance of succeeding if it comes to that as most of the original jurors reportedly found in Google favor on that question in the first trial.
  • Reply 137 of 206
    Quote:

    Originally Posted by ruddy View Post

     

    Why would they? Microsoft made CLI an open standard in order to make projects like Mono possible. If MS hadn't wanted to do that, there would be no Mono today. Different companies employ different business models around their APIs. 


     

    Right but you misrepresent Sun's business model. They did not license the APIs. In fact they operated very similarly to Mono in this case. Please, tell me who benefits from having the API structure copyrightable? No small developer is going to benefit here as there need only be one prior example and it need not be registered as copyrighted. Therefore literally any developer using any organisation of an API (so most developers) could be liable with absolutely no way to even research it beforehand.

     

    I get the feeling you are not a developer as you think this is ideological rather than pragmatic.

  • Reply 138 of 206
    Quote:
    Originally Posted by ruddy View Post

     

    Why would they? Microsoft made CLI an open standard in order to make projects like Mono possible. If MS hadn't wanted to do that, there would be no Mono today. Different companies employ different business models around their APIs. 


    From what I gather, the CLI only describes the C# language and runtime environment, not the .NET libraries that happen to be written using the CLI:

     

    "The Mono project provides an open source C# compiler, a complete open source implementation of the Common Language Infrastructure including the required framework libraries as they appear in the ECMA specification, and a nearly complete implementation of the Microsoft proprietary .NET class libraries up to .NET 3.5. " (http://en.wikipedia.org/wiki/C_Sharp_(programming_language)#Implementations)

     

    If you don't like the Mono example, what about Wine, the reimplementation of the Windows API which has certainly not been endorsed by MS? It has been around for twenty years and I'm not aware of any legal action against it. In fact, it's used in several major software projects, such as Oracle's own Virtualbox (https://www.virtualbox.org/ticket/2940) as well as Parallels (http://wiki.winehq.org/Parallels). My point is that until this Oracle case the software community has always considered copyright to protect source code, and has treated an API as merely a set of specifications. If the specifications themselves were considered copyright protected, all those alternative implementations of Unix, such as GNU/Linux and the BSDs, probably wouldn't exist today.

  • Reply 139 of 206
    jungmarkjungmark Posts: 6,926member
    gatorguy wrote: »
    So you don't have a law degree then and no special insight to add? Then you must be depending on the same Mueller article I and many others read for your view of how things will proceed. Your comments seem to mirror his for the most part. IMO Fair Use seems to have a good chance of succeeding if it comes to that as most of the original jurors reportedly found in Google favor on that question in the first trial.

    I don't think you can lift copywrited material verbatim and imbed it into your works. It's not a review or critique.
  • Reply 140 of 206
    Quote:

    Originally Posted by jungmark View Post





    I don't think you can lift copywrited material verbatim and imbed it into your works. It's not a review or critique.

     

    It's a misleading view. Let's say I wrote a book on cats. I label each chapter with the name of the major cat species I am dealing with, include some brief details on their geographic distribution and add a few major categories in to distinguish one cat group from another.

     

    Now, another person produces a book that is completely different from my book other than chapter names and organisation.

     

    Is this really copyright infringement? There is clearly some creativity that goes into putting together a list of cats, but it's pretty damn limited. Do you think that once I have organised cats in a certain way, that I should gain copyright protection over that particular organisation as well as the names? That is especially ludicrous to me.

Sign In or Register to comment.