Some Thoughts on Google v. Oracle, from Prof. Rebecca Tushnet (Harvard)

"Categories, microworks, and market circularity."


From Rebecca Tushnet's 43(B)log:

[1.] The majority clearly says that, as with other categories of protected works, distinctions can be made within the categories, drawing lines "among" computer programs, books, and films. Not all literary works are the same; Infinite Jest gets a different kind of copyright protection than my emails do. Likewise, while the recent Warhol case at times seems to imply that the derivative works right overrides fair use, the same GvO passage says that copyright provides both reproduction and derivative works right, but also subjects all works to fair use.

[2.] One of my minor obsessions is "courts that reproduce the entire works in suit in the opinion"—whether they find for the plaintiff or the defendant, and whether they rule on substantial similarity or fair use, they never even consider whether it's ok to do so. It's obviously a good idea for purposes of understanding what the law is—a description of a song or picture will never allow a subsequent reader to understand what the protectable expression in the song or picture was—and I think obviously fair, but it's amusing to me that it happens without anyone pointing out that this must be in reliance on fair use.

Anyway, in GvO, Justice Breyer instead reproduces an entire short story, which was just minding its own business and had nothing to do with the case, in two different languages no less. And he does so in the course of suggesting that the scope of fair use would be more limited with respect to that short story than to a sentence of the same length in a longer novel. I think that's a troubling conclusion—Justin Hughes has written very well about the problem of "microworks" and the right result would probably be to say that the book of stories from which that story comes should be the proper unit of analysis for factor three.

But now I have questions: Can casebook authors use this portion of the case without fear? Could the copyright owner bring a CASE Act claim against the US as a result? I don't actually know whether claims against the US can go to the Copyright Office or have to go to the Court of Claims.

[3.] On factor four [of the fair use analysis], it was nice to see acknowledgement that (1) the licensing package Oracle offered was very different from what Google ultimately copied, and thus didn't show market harm from what Google actually copied and (2) this was a circularity problem, which should be avoided.

Also, relevant to the "mixed question of law and fact" issue, the majority says: "the jury's fair use determination means that neither Sun's effort to obtain a license [so in the Court's opinion, but likely refers to Google's effort to obtain a license -EV] nor Oracle's conflicting evidence can overcome evidence indicating that, at a minimum, it would have been difficult for Sun to enter the smartphone market, even had Google not used portions of the Sun Java API." I think that means that the jury verdict must be interpreted to have favored Google on factor four, resolving the factual part of factor four in its favor. I am not sure what that means for summary judgment in future cases, especially if factor one remains more of a legal question.

NEXT: More Speculation about the Cert Denial in Small v. Memphis Light, Gas & Water

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. The discussion of length was … interesting, I mean, one of the first things you learn in CS is that just because you wrote a lot of lines of code doesn’t mean much. Just because you wrote a complex problem in a few lines doesn’t mean much. It matters what you are doing.

    I dont know much about fair use, so I dont know if the description of Suns and Google business practices resolve in favor of Google wrt fair use. Tbh I kinda thought Oracles strongest argument was that Google behaved improperly.

    But ultimately, it came down to the fact that these are declaring codes and you need to invoke them in order to do anything, and you cannot really distinguish what Google did from what ordinary software engineers do every day. I mean I suppose you can, but there would be no legal basis for that distinction other than, the jury said this!, and, well, the jury voted in favor of Google. Which, echoed in the opinion, seemed broadly correct.

    1. >what ordinary software engineers do every day.

      But generally under some sort of license. It’s worth remembering that various groups use e.g., GPL header files to ensure compatibility.

      IIRC, that was the case here. One of the main things Sun/Oracle wanted was restrictions on “embracing and extending” Java.

      1. > But generally under some sort of license. It’s worth remembering that various groups use e.g., GPL header files to ensure compatibility.

        You don’t need header files. In this case, the declarative code could be reconstructed from the API definition and is commonly done with certain protocols, such as a SOAP WDSL. This would have resulted in exactly the same code. The structural code is a byproduct of the API. They cannot be separated.

        Oracle knows this and was using it to attempt bilk google out of cash.

        1. >In this case, the declarative code could be reconstructed from the API definition…This would have resulted in exactly the same code.

          Why wouldn’t that be a derivative work?

    2. “and you cannot really distinguish what Google did from what ordinary software engineers do every day”

      Actually, you can quite easily. Ordinary software engineers every day LICENSE the code.

  2. I’m puzzled by this “Sun’s effort to obtain a license”. Sun originally wrote Java and Oracle acquired it by buying Sun.

    The history of the case is long and complicated. I’m not sure the appeals court nor the Supreme Court understood what an API actually is.

    1. Appeals, no. Breyer, Yes. SCOTUS deferred to him.

  3. As a practicing software engineer of more than three decades experience, the author of about two million lines of production code including a 3500-entry API with a 135,000-line implementation, and as a follower of this case since the beginng, I find Oracle’s behavior despicable and entirely in the wrong.

    The original judgement by Judge Alsup was absolutely right to begin with. And in keeping with a century-long set of industry standards (precedents from the “real world”), as Microsoft and others pointed out.

    An API is a “small” language that gives the names and specifications of the operations provided by a system; it does not say how the system does these operations (the “implementation”, which is the creative work). An API is purely functional in nature. It has been established for more than a century that languages are not copyrightable. Nor are purely functional works.

    As Breyer notes (less precisely than he should), there are two kinds of facts: (1) that a system provides a “maximum for integer pairs” operation, together with the name “max” for that operation (the API item); and (2) how the system computes that max, whether by normal integer arithmetic or by the (unreasonable) means of translating integers into Roman numerals and then performing an arcane calculation with those Roman numerals (the implementation). The former is one of the items in the definition of a (non-copyrightable!) language; the latter is potentially creative and copyrightable. Or maybe not, assuming the “normal integer arithmetic” implementation, which is generally a single hardware instruction already provided by the hardware vendor (Intel, or AMD, or IBM, or …)

    1. Your entire argument seems to be misplaced considering that Breyer assumed that what Oracle’s code WAS copyrightable.

      1. As I said, Judge ALSOP

        1. As I said, Judge ALSOP was absolutely right to begin with.

      2. Was Amazon’s S3 cloud management API copyrightable? I note that Oracle has copked it: see

  4. When I first heard about the Google v. Oracle and the USSC ruling it seemed wrong to me. However, after reading large portions of the ruling, and remembering Sega v. Accolade (I worked in the games industry during this time), I think it was a foregone conclusion that Google v. Oracle would be decided this way.

    All the discussion of fair use pales in comparison with what was decided in Sega v. Accolade. In it the USSC ruled that decompiling security code to reverse engineer it to get around the security code and write software for the Sega game system was ‘fair use’. In a world when performing such actions on private information that was unpublished, it is a foregone conclusion that using public information for basically the same thing (to work with java libraries and API’s) would be deemed ‘fair use’.

    In the end, people should go back and complain about the Sega decision and not the Google decision, as it is the parent that led to the current ruling.

    Interestingly, the Sega decision is only mentioned twice in the Google decision, however those references are significant:

    1) Page 17 – “Sega Enterprises Ltd. v. Accolade, Inc., 977 F. 2d 1510, 1521–1527 (CA9 1992) (holding that wholesale copying of copyrighted code as a preliminary step to develop a competing product was a fair use)”

    2) Page 34 – Sega Enterprises, 977 F. 2d, at 1523–1524 (“An attempt to monopolize the market by making it impossible for others to compete runs counter to the statutory purpose of promoting creative expression”)

    These seem like at least two legs of the stool that holds up fair use and promoting creative expression. Hard to rule for Oracle with the Sega decision sitting there and acting as binding precedence.

Please to post comments