http://qs321.pair.com?node_id=968511


in reply to Copyright on languages

I believe there's no past decision on this because everyone in the industry operates with the understanding that computer languages (e.g. Perl, not perl) aren't Copyrightable.
"Everyone"? Considering the kind of lawsuits and patents we have seen in the past, I don't want to put it past some companies that assume languages are copyrightable. AFAIK, it has not been tested in a *US* court (but see the remark elsewhere about a recent European decision).

I also don't think the fact that APIs aren't copyrightable is a clear cut case. After all, many APIs aren't so much designed, as that they are the outcome of the first thing the programmer wrote down (many of my APIs certainly are). As such, they would be more expressions than ideas. I don't see an API as a "fact".

Replies are listed 'Best First'.
Re^2: Copyright on languages
by BrowserUk (Patriarch) on May 02, 2012 at 19:28 UTC
    I also don't think the fact that APIs aren't copyrightable is a clear cut case.

    Can you imagine if int add( int, int ) were (or suddenly became) copyrightable?

    Perhaps that can be seen as too trivial; but consider something like:

    void qsort ( void * base, size_t num, size_t size, int ( * comparator +) ( const void *, const void * ) );

    If, every time a programmer sat down to write a function; he had to perform a copyright search to check that he wasn't duplicating a function signature that someone else had used somewhere at some time in the past, this industry would grind to a complete halt.

    That is proof by reductio ad absurdum, but the principle holds. And with the advent of IP Trolls, and the general rise and rise of companies like Apple and Oracle using litigation as the first resort rather than the last; it is not so absurd.

    Api definitions involve little more than the serial juxtaposition of a half a dozen or so keywords, chosen from a very limited subset. The permutations are finite.

    Imagine if writers were able to copyright phrases -- short combinations of common words. It would very soon lead to the situation where it became impossible to communicate in writing at all.

    buk_The buk_only buk_solution buk_would buk_be buk_for buk_each buk_individual buk_or buk_company buk_to buk_'register' buk_their buk_own buk_unique buk_prefix; buk_and buk_then buk_you buk_end buk_up buk_with buk_communications buk_of buk_this buk_sort buk_and buk_life buk_just buk_gets buk_king buk_ridiculous.


    With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
    Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
    "Science is about questioning the status quo. Questioning authority".
    In the absence of evidence, opinion is indistinguishable from prejudice.

    The start of some sanity?

      Of course, BrowerUK, you are absolutely correct. I just don't understand why the judge did not use that reasoning.

      Imagine if writers were able to copyright phrases -- short combinations of common words. It would very soon lead to the situation where it became impossible to communicate in writing at all.
      First of all, you don't "copyright things", other than by creating it. And copyright law doesn't say "well, a work has to be at least N words large". Yet, noone has to worry that if they write down a short phrase they are infringing a copyright, even if someone else has written down the same phrase before them. Yet, they aren't free to copy an entire book.

      Now, if APIs were copyrightable (and let me make it clear, I do not have the opinion they should), why should they be treated differently? That is, one shouldn't have to worry about int add( int, int ), but from that, it doesn't imply you can copy a large API like Java.

      My point is no more than "I don't think it's obvious that APIs aren't copyrightable". And the mere fact there was a court case, and it wasn't thrown out as frivolous makes me think I wasn't the only one thinking it wasn't obvious.

        First of all, you don't "copyright things", other than by creating it.

        That's not true. You can register a Copyright, and it really limits your legal options if you don't register that Copyright (under US law).

        It's even one of the issues to be resolved in this case: Can Oracle sue for infringement of "37 APIs" given that Copyright it registered is not specifically for those "37 APIs".

        Of course, BrowerUK, you are absolutely correct.

        I was simply expressing my opinion -- just as you were expressing yours. I see nothing wrong in that. Unless you consider that your opinion is unassailable?

        That dripping sarcasm might stand up if your follow-on discussion was logical. Or even if you demonstrated a deep(er) understanding of the issues involved.

        And copyright law doesn't say "well, a work has to be at least N words large". Yet, noone has to worry that if they write down a short phrase they are infringing a copyright, even if someone else has written down the same phrase before them. Yet, they aren't free to copy an entire book.

        But would the law be broken if I wrote a book that had the same chapter titles and ordering as a previously published work?

        How about if the book contains similar plot lines and characters; albeit with different names?

        Now, if APIs were copyrightable (...), why should they be treated differently? That is, one shouldn't have to worry about int add( int, int ), but from that, it doesn't imply you can copy a large API like Java.

        Where do you draw that line? What constistutes "the Java API"?

        Besides which, the case does not revolve around "the API", but rather about "bits of the API". Ie. A certain, defined, specific, named, subset of, the entrypoints within "The API".

        Informed opinion has it that specific implementation ("the source code") of a program or other software entity is copyrightable; but the step-wise description ("algorithm"); and the interface specification ("API") are not.

        My opinion is founded upon my understanding of these and other references going back over the long term of my particular interest in this subject.

        Disagree by all means; counter-argue please; but do save up the distraction of 'dismissal through ridicule' for those occasions where you can back it up with some ammunition worthy of the name.


        With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
        Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
        "Science is about questioning the status quo. Questioning authority".
        In the absence of evidence, opinion is indistinguishable from prejudice.

        The start of some sanity?

Re^2: Copyright on languages
by tilly (Archbishop) on May 03, 2012 at 00:07 UTC
    APIs are in fact copyrightable. However there is a limit to their copyrightability, and the bstraction-Filtration-Comparison test is the standard legal test in the USA to try to distinguish between copyrightable elements and non-copyrightable elements.

    Of course the line between artistic expression and the effect of the code is drawing lines in the sand that don't really exist. But these imaginary lines are supported by lawyers and judges, which can cost you a lot of money.

      APIs are in fact copyrightable.

      If that is "in fact" so, then you should be able to easily cite unassailable proof for it?

      A specific implementation of an api -- Ie. the source code -- can be copyrighted. And copyrights on parts of that source code -- as determined by the BFC test -- can be be legally enforced.

      But no one has yet succeeded -- though few have tried -- in enforcing copyright on the specification of an API when re-produced under clean-room conditions.


      With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
      Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
      "Science is about questioning the status quo. Questioning authority".
      In the absence of evidence, opinion is indistinguishable from prejudice.

      The start of some sanity?

        Proof of copyrightability is easy to come by. Just see Johnson Controls v. Phoenix Control Systems where it says:

        (12) Source and object code, the literal components of a program, are consistently held protected by a copyright on the program. See, e.g. CMS Software Design Sys., Inc. v. Info Designs, Inc., 785 F.2d 1246, 1249 (5th Cir.1986) (source code); Apple, 725 F.2d 521 (object code). Whether the non-literal components of a program, including the structure, sequence and organization and user interface, are protected depends on whether, on the particular facts of each case, the component in question qualifies as an expression of an idea, or an idea itself.

        (13) Here, the district court found that the structure, sequence and organization of the JC-5000S was expression, and thus subject to protection. The JC-5000S is a very sophisticated program, and each individual application is customized to the needs of the purchaser. This practice of adaptation is one indication that there may be room for individualized expression in the accomplishment of common functions. The district court's finding of expression is also supported by the special master's report. Regarding one particular point type, the master indicated that, although it is common for process control software packages to include provisions for collecting historical data, and using various integration and averaging schemes to do so, it is unusual to implement this function as a point type, as Johnson did. This finding also supports the view that some discretion and opportunity for creativity exist in the structure, and that the structure of the JC-5000S is expression, rather than an idea in itself. This issue will no doubt be revisited at trial, but at this stage of the proceedings we cannot say that the district court clearly erred.

        And there you have it. The API is part of the structure, sequence and organization and user interface and therefore is potentially copyrightable. However actually achieving copyright protection requires a higher bar than for just copying literal text.

        In Oracle's case the bar is going to be even higher. In their case even if the court rules that it was copyrightable, Google can argue that Oracle has years of telling people that they have no copyright interest on code that has copied elements of those APIs in their own code (by using them), and therefore by the doctrine of unclean hands can't now turn around and sue someone over that copyright.

        As always I am not a lawyer and this is not legal advice. It is merely my lay opinion that there are court rulings that support the claim that APIs are potentially copyrightable, but the bar to winning copyright is very high, and even if Oracle is ruled to have copyright, they still are likely to lose their case

Re^2: Copyright on languages
by ikegami (Patriarch) on May 02, 2012 at 19:15 UTC

    As such, they would be more expressions than ideas.

    I believe the output of mechanical processes isn't covered by Copyright either.

    I don't see an API as a "fact".

    An API is the knowledge of what methods are available, how to call them and what they do. How is that not fact?

      An API is the knowledge of what methods are available, how to call them and what they do. How is that not fact?
      To quote yourself
      While I recognise that creating a good API is indeed a difficult task
      If you're creating it, it's not a fact, is it?

      Unless you want to argue that everything ever created becomes a fact, in which case we can just ditch copyright law, and software licenses right away.

        Definitions, phone numbers, addresses and algorithms are all invented, yet are all facts or ideas, and none can be Copyrighted.