Supreme Court of the United States
April 5th, 2021
Google LLC
Oracle America, Inc.
When Google was first developing Android back in 2005, they wished to incorporate libraries from Java SE, the Java Standard Edition. So, Google and Sun Microsystems[1] endeavored to come to an agreement for a license to use Java SE in Android. However, they were unable to come to an agreement, and Sun denied Google a license. While Java SE was commercial, there was an “open source” version of Java, but it didn’t have as many features as Java SE.
Here’s the important part: Google wasn’t satisfied with using the open source version with fewer features, so they wanted to develop their own custom “clean room” implementation of Java SE for Android. However, Oracle mentioned in the court case that Google’s implementation had method declarations that were identical to the real Java SE. It’s crucial to know that Google only copied the method declarations but wrote their own “unique implementations” for these method declarations in their version.
This essentially means Google copied the blueprint of Java SE for their version, but not the actual method implementations that the actual Java SE owned by Oracle used. This is what you would call a “clean room” implementation, where a new system is implemented only by understanding the behavior of an original system, rather than knowing how the original system functions with complete detail.
Oracle believed that their APIs are copyrightable, and that Google infringed on their copyright of the Java SE APIs. However, Google believed that their copying of the method declarations in their Java SE implementation was of a functional nature, rather than an expressive nature. Because Google tried not to copy the actual method implementations in their version, Google argued that they didn’t infringe on Oracle’s intellectual property that has creative expressiveness and that their incorporation of the original method declarations constituted fair use.
Google wanted to use their implementation of Java SE in Android because they believed experienced Java developers would be able to write applications for the platform in a programming language that they are familiar with. This contributed to Android’s success, as developers would not have to cumbersomely learn another programming language to develop for Android.
[1] Sun Microsystems was bought by Oracle, so they became the owner of the intellectual property rights for Java SE.6-2 ruling in favor of Google
The issue made its way through multiple courts, with some rulings in favor of Google and some in favor of Oracle, including some mixed results. The original lawsuit was filed on August 13th, 2010.
The District Court was uncertain about the validity of Google’s fair use defense, however, they held that copyright law did not cover APIs.
The Court of Appeals of the Federal Circuit ruled that Google’s reimplementation of their APIs was not fair use, and that they could have written their own method declarations with different structure and naming conventions, instead of using the original declarations in Oracle’s version. A second jury had ruled that Google’s reimplementation was fair use, but their ruling ended up being rejected.
Once Google was granted certiorari, the case made its way to the Supreme Court for the sole purpose of determining whether Google’s reimplementation was fair use. They did not acknowledge Google’s Seventh Amendment defenses, instead proceeding to analyze the statutory factors for fair use.
Google argued that their Java API reimplementation was considered fair use, while Oracle argued that their APIs are copyrightable. To begin, the four fair use factors were: nature of usage, amount and substantiality, purpose and character, and market effects. These fair use factors are the common standard when it comes to fair use rulings.
The majority opinion was in favor of Google, ruling that Google’s re-implementation was fair use. To begin, let’s look at the majority rulings for each fair use factor:
Nature of Usage: Google copied the declaring method headers from the original version owned by Oracle, but didn’t have access to their original method implementations. Google wrote their own for their version. As a result, the majority ruling stated that this type of work is different from many other types of code. They said the intent was for Java programmers to be able to access a common set of out-of-the-box methods to aid in developing applications for Android.
The ruling’s key message for this fair use factor was that “the copied lines are inherently bound together with uncopyrightable ideas”, meaning that the method declarations Oracle suggested were copyrightable were in fact not. It also stated the new method implementations independently written by Google was “the creation of new creative expression”.
In addition, the ruling said the value of the method headers copied by Google was “unlike many other computer programs”, and was derived from the investments Java developers make in learning the API. They also said their fair use ruling was unlikely to be detrimental to the usual copyright protections that are provided for computer programs.
Amount and Substantiality: One important detail which contributed to the majority ruling for this fair use factor was that Google only copied less than 1% of the total lines making up the original API, said to be only a portion of the “considerably greater whole.” This was coupled with the fact that Google only copied the lines that were necessary for a functional API for use in developing for Android.
If Google had copied a more substantial amount of declaring code, the final decision would lean more toward Oracle. If they had done so, it may have been considered more disingenuous for their purpose: to allow Java developers to bring their programming skills to the emerging Android platform.
Purpose and Character: For this fair use factor, it was said that something had to have a “further purpose” or “different character” to be considered a transformative work. As mentioned earlier, Google had copied only the method declarations that were needed to develop for the new Android platform while still being able to use the Java language.
The majority opinion noted how using APIs can make software development more productive, and that Google’s use of the API matched up well with “creative progress”, which is the basic intent of copyright.
Market Effects: This fair use factor takes into account the “effect” that the copyrighted work could suffer from as a result of the copying of this work. The majority ruling states that Google’s reimplementation is not a sufficient market substitute for Oracle’s Java SE and that Oracle would have been able to benefit from their reimplementation, as the copyright holder.
It was also mentioned in this ruling that the enforcement of copyright here would adversely affect the creative landscape, or stifle innovation.
To summarize the majority ruling, Oracle’s API implementation was not copyrightable and Google’s implementation was fair use. It said that Google used only the method declarations that were necessary for Android development, that their reimplementation was considered a transformative work that aligns with the values of copyright, and that this did not bring about any market consequences that would negatively affect Oracle. Justice Breyer wrote the majority opinion of the court in favor of Google, leading to the notion that the declaring code is “far from the core of copyright” and becomes valuable only when third parties i.e. software developers value it, in addition to being “bound together by uncopyrightable ideas”.
Justice Clarence Thomas wrote a dissenting opinion in favor of Oracle, mentioning that their copyright intends to protect the declaring code more than the implementing code. He reasoned that developers interact with the APIs by making use of the declarations, but they never see the underlying implementations behind the methods that they use. In addition, he stated that the market effects of Google’s implementation “decimated Oracle’s market” and that the declaring code attracted developers in a way that positioned Google’s implementation for Android as a market substitute for Oracle’s Java SE.
He stated that the majority had conflated the transformative use of a copyrighted work with derivative use. It was said that the ability to produce derivative works was exclusive to the copyright holder.
The final decision in favor of Google was considered “a big win” for innovation and interoperability among many computer scientists, computer programmers, and other tech workers.
In a 2014 Electronic Frontier Foundation press release in light of the decisions by the Federal Circuit Courts, they noted that excluding APIs from copyright protection helped bolster development of modern digital infrastructure, including the Internet, by way of allowing existing APIs to be reimplemented or extended. In addition, they implied that allowing APIs to be copyrighted could potentially be considered an anti-competitive practice.
The article also includes Judge William Alsup’s ruling from 2012, who said that Java APIs are not copyrightable and that Oracle would have been allowed to “tie up” a set of utilitarian methods that many Java programmers rely on. To clarify, these methods perform actions that are considered “generic” and standardized by many different APIs for many different programming languages in the computing world, as opposed to methods that perform actions specifically related to Sun’s or Oracle’s ecosystem.
Rather than setting a new precedent, the Supreme Court ruling solidified the consensus which has been held by computer scientists and other workers in the tech industry for decades: that APIs are not entitled to copyright protection, and by this rule, developers as well as companies would be allowed to build on others’ work and advance the field of computing.
This included the possibility of promoting competition in the tech industry, which is often regarded as a key to innovation. Extending copyright protection to APIs has been said in the software industry to end up stifling innovation in addition to promoting a sense of leniency with anti-competitive practices.
Caballar, Rina. “Google v. Oracle Explained: The Fight for Interoperable Software.” IEEE Spectrum, 19 Feb. 2020, https://spectrum.ieee.org/google-v-oracle-explained-supreme-court-news-apis-software
Graif, Michael. “Google v. Oracle: a closer look at the dissent.” World Intellectual Property Review, 16 Apr. 2021, https://www.mintz.com/sites/default/files/media/documents/2021-04-22/Google%20v%20Oracle%20%20a%20closer%20look%20at%20the%20dissent.pdf
Samuels, Julie. “Computer Scientists Urge Court to Block Copyright Claims in Oracle v. Google API Fight.” Electronic Frontier Foundation, 30 May 2013, https://www.eff.org/press/releases/computer-scientists-urge-court-block-copyright-claims-oracle-v-google-api-fight
McSherry, Corynne. “Computer Scientists Ask Supreme Court to Rule APIs Can’t Be Copyrighted.” Electronic Frontier Foundation, 7 Nov. 2014, https://www.eff.org/press/releases/computer-scientists-ask-supreme-court-rule-apis-cant-be-copyrighted
Stim, Rich. “Measuring Fair Use: The Four Factors.” Stanford Copyright & Fair Use, Oct. 2019, https://fairuse.stanford.edu/overview/fair-use/four-factors/
“GOOGLE LLC v. ORACLE AMERICA, INC.” Supreme Court of the United States, 5 Apr. 2021, https://www.supremecourt.gov/opinions/20pdf/18-956_d18f.pdf