Interest Groups

Google v. Oracle: A Quick Overview of the Court’s Fair Use Analysis - Solo Small Firm Practice News

Get the news you want the way you want it: click the RSS button in the right corner to add this feed to your RSS reader, or click here to subscribe to this content. By subscribing, you’ll find this news on your Member Account page, and the latest articles will be emailed to you in your customized IndyBar E-Bulletin e-newsletter.

Solo Small Firm Practice News

Posted on: May 6, 2021

By Travis Baxter, Maginot Moore & Beck LLP

In the recently decided Supreme Court case Google LLC v. Oracle America, Inc., 593 US ___ (2021), the Court took up the case of whether Google’s copying of Oracle’s API (Application Programming Interface) software constituted a fair use. In a 6-2 decision authored by Justice Breyer, with Justice Barrett taking no part in the case and Justices Thomas and Alito dissenting, the Court determined that, indeed, Google’s use of Oracle’s API was a fair use.

To set the stage, Oracle owns a copyright in the Java SE computer program that includes the Java programming language. Google copied portions of the API in the Android platform software, specifically, portions of the API’s declaring code of the Java language. Put simply, the declaring code involves the simple code that programmers write to instruct the platform to perform more sophisticated functions and operations. For instance, in the example used by the court, to determine the maximum of two numbers, instead of writing the code that will determine the maximum of two numbers from scratch, the programmer can simply use the declaring code “java.lang.Math.max(x, y),” which directs the platform to access the implementing code that determines the maximum of x and y. Google independently wrote all the implementing code for the Android platform, but copied approximately 11,500 lines of the widely-known declaring code, and the organizational structure thereof, from Java’s API so that programmers would be able to work in the Android operating system without having to learn an entirely new set of declaring code commands.

There were two issues before the court. First, the court was asked to determine whether the APIs were even copyrightable subject matter. The majority punted on this question, however, and assumed for the sake of argument that the APIs are copyrightable.

The second issue, and the issue on which the court decided the case, was whether Google’s use of the declaring code was a fair use under 17 U.S.C. § 107. To step back for a second, the procedural posture of the case is initially relevant to the fair use issue. The district court held a jury trial in which the jury determined that Google’s use was a fair use as a matter of fact. The Federal Circuit then reversed, holding that fair use is a mixed issue of law and fact and, using the jury’s factual determinations, Google’s use of the APIs was not a fair use as a matter of law. In the instant decision, the Supreme Court agreed with the Federal Circuit that fair use is indeed a mixed issue of law and fact, but disagreed on the ultimate conclusion that Google’s use was not a fair use.

Next, the court moved on to analyzing the four fair use factors set forth in §107, beginning with the nature of the copyrighted work. According to the court, the nature of the declaring code is basically a system and nomenclature that would be intuitively easy to remember so that programmers could easily program using the Java language and would therefore be more likely to use Java over other programming languages. The court determined that the declaring code is further from the core of copyright than other computer programming (i.e. the implementing code) since it is inherently bound with uncopyrightable ideas (i.e. general task division and organization) and other creative expression (i.e. Android’s implementing code). Additionally, a good amount of the declaring code’s value derives from programmers’ investment of their time in learning the declaring code, rather than the creative expression embodied in the code. As such, the court decided that this factor weighs in favor of fair use.

The court then moved on to an analysis of the purpose and character of the use factor. The court’s analysis generally dismissed the commerciality and potential bad faith considerations of Google’s use, instead focusing on whether Google’s use of the APIs was transformative. The court found that Google’s use was designed to create new products, which furthers the “progress” goal set forth as the primary objective of copyright in the IP Clause of the U.S. Constitution. Google’s use was determined to be a “reimplementation,” in building the Android platform using the same words and syntaxes of the existing Java system, which, according to the Court, may even benefit Oracle due to the Java language attaining more widespread use. Consequently, the court decided that Google’s use was transformative of the initial work, and this factor therefore weighs in favor of fair use.

The third factor analyzed in the decision was the amount and substantiality of the portion of the (presumably) copyrightable material used. The court decided that the 2.86 million lines of implementing code that Google did not copy were much more important to the copyrighted nature of the overall work than the 11,500 lines of declaring code that were copied. Additionally, the purpose of the declaring code is to call up the implementing code, and the copying was done because the code was known, not because of the creativity, beauty, or even purpose of the declaring code. Moreover, the copied portion was tethered to the transformative nature of Google’s use, i.e. to create a different platform for a different environment (i.e. smartphones). In particular, the copied declaring code was “the key that [Google] needed to unlock the programmers’ creative energies” and enable programmers to use the Java programming language on the Android platform. Thus, according to the court, since Google’s objective was to permit programmers to make use of their knowledge of the Java API, rather than the simply make Java usable on the Android platform, the “substantiality” factor also weighs in favor of fair use.

Finally, the court looked at the market effects factor. The court analyzed evidence that Sun’s (Sun Microsystems was the previous owner of the Java code) amount of loss was likely low because they were in the computer market, and not well-positioned to produce a smartphone product. Moreover, evidence was presented at trial that Sun’s computer products would benefit from the broader use of Java in the Android smartphone platform. On the other hand, Google has made a substantial profit from the Android platform. Nonetheless, when further factoring in the risk of harm to the public of allowing Oracle to monopolize the market and stifle creative expression, the court determined that the market effects factor weighs in favor of fair use.

Thus, according to the majority, since all four fair use factors weigh in favor of fair use, Google’s copying of the portions of Oracle’s Java code is a fair use as a matter of law.

Justice Thomas wrote a strongly worded dissent in which he argued that the API is indeed copyrightable, and Google’s use was “anything but fair.” Justice Thomas explained that, even though declaring code serves a function, it is not a “method of operation” excluded from copyright protection by 17 U.S.C. § 102(b) because the declaring code is inextricably bound to the implementing code, and therefore the declaring code has no function on its own. Moreover, the language of §101 specifically allows copyright protection for computer code that “indirectly … bring[s] about a certain result.” Justice Thomas noted that Congress therefore intended for declaring code to be protectable.

In the fair use analysis, Justice Thomas found that three of the four fair use factors lean toward the finding that Google’s use was not a fair use. Justice Thomas disagreed with the majority’s determination that the declaring code was further from the core of copyright than the implementing code, but nonetheless found that because the code is predominantly functional, nature of the copyrighted work factor likely favors Google anyway. Justice Thomas further argued that the market effects were the most important factor, and that Google cost Oracle significant business by eliminating the reason Google’s competitors would pay for the Java platform and, moreover, interfered with Oracle’s ability to license the Java platform to developers of smartphone operating systems. In the analysis of the purpose and character of use factor, the dissent found that Google’s use was overwhelmingly commercial and non-transformative because it did not add any new expression, meaning or message, heavily favoring Oracle. Finally, Justice Thomas disagreed with the majority’s conclusion on the amount and substantiality factor, finding that because Google copied the heart or focal points of the declaring code verbatim, and the declaring code serves as a market substitute for the original work or licensed derivatives thereof, Google’s use was substantial.

Thus, according to Justice Thomas, three of the four fair use factors, including the two most important ones, favored Oracle. He would therefore have held that Google’s use of the Java declaring code was not a fair use.

Overall, the court’s decision is very fact-specific, which is typical for fair use cases. Nonetheless, the focus on the transformative nature of Google’s verbatim copying of the declaring code is intriguing, as is the application of the fair use factors to the functional nature of the declaring code. Finally, the court’s discussion of the creativity of the declaring code weighed against the policy considerations of facilitating additional creation by programmers using the declaring code could find applicability in areas of copyright outside computer programs and APIs. 

If you would like to submit content or write an article for the Intellectual Property Section, please email Kara Sikorski at

Subscribe to Intellectual Property Section news here!


Indianapolis Bar Association (IndyBar) est. 1878 | 4,536 Members (as of 2.11.21)