'Declaring code' is fair game 4 min read
In a recent decision, the US Supreme Court held that Google did not infringe Oracle's copyright when it took 11,500 lines of functional code to build the Android operating system. The decision upholds the longstanding industry practice of copying this type of code.
- A 6-2 majority of the US Supreme Court ruled that Google was entitled to take 'declaring code' under the principles of 'fair use'.
- The court left open the question whether this type of code is protected by copyright. Developers operating under US law need to interpret the factors of fair use to see if their actions are legitimate.
- Australia does not have a general 'fair use' doctrine so the same question would not be decided on the same basis here.
- General counsel
- IP/IT legal teams
- Software development/IT teams
In 2005, Google began developing its Android software platform for mobile devices. Java was already a popular programming language, created by Sun Microsystem (now owned by Oracle). Google copied 11,500 lines of Java code, used in application programming interfaces (APIs). APIs are software components used to enable one piece of software to communicate with another: eg a Twitter API allows a user to post an article from a web browser onto Twitter's software platform. The Java code that Google took was 'declaring code', which is the part of the API that organises and 'declares', or identifies, pre-written functions and tasks, so that the API can communicate with other software. Other code in the API, called 'implementing code', performs the tasks. Google took the 'declaring code' so that its programmers, who were familiar with the Java language, could use prewritten codes to build certain functions, rather than starting from scratch.
It is well established in many countries that copyright protects software code, but not the function the code performs. In the context of APIs, their functional nature has led some commentators to argue that copyright does not protect APIs, or that copying them is not copyright infringement. This argument is consistent with the widespread industry practice – for decades – where APIs are freely used and copied. When Oracle made its claim against Google in 2010, this attracted significant attention. Many of the copyright associations, as well as the US Government, supported Oracle's view; while Microsoft, IBM and other software development organisations were on Google's side.
At first instance in 2012, the US District Court for the Northern District of California acknowledged the existing industry practice and held that APIs were not copyrightable. On appeal to the Federal Circuit, APIs were found to be copyrightable, but the case was remitted on the question of whether Google's use was justified as 'fair use'. After two further iterations, Google petitioned the US Supreme Court to review the decisions on both copyrightability and fair use.
The US Supreme Court majority sidestepped the question of whether the declaring code was in fact copyrightable, but found in favour of Google on 'fair use'. On the key elements of this exception, they said:
- on the purpose and character of the use, Google had 'transformed' the code into a new platform (the Android smartphone platform);
- on the nature of the copyright work, the declaring code was a distinct type of code serving an organisational function which was 'inherently bound' with uncopyrightable ideas;
- on the amount and substantiality of the portion used, only 0.4% of the total API code was taken, and Google only took the part needed for functionality. Google copied that part because programmers had learnt to work with the Java language, not because of the 'creativity' or 'beauty' of that code; and
- on the market effects of the use, Google's smartphone platform was not a market substitute for the Java platform. 'Creativity-related harms' to the public could eventuate if only Oracle could use these APIs.
The decision puts an end to the decade-long dispute over a $9 billion claim by Oracle. Two judges (Justice Thomas and Justice Alito) dissented. They criticised the majority's avoidance of the question of copyrightability. They also said the majority's interpretation of fair use was too wide and 'eviscerates copyright'.
Like many other common law countries, under Australian law 'fair dealing' exceptions to copyright infringement are limited to specific categories of activities such as research or study, parody or satire, or reporting news. There is no US-style general 'fair use' doctrine, and none of the heads of 'fair dealing' would have applied in this case.
Section 47D of the Copyright Act 1968 (Cth) specifically addresses interoperability: it allows copying of computer programs to make software interoperable. However, this provision is only available to an owner or licensee of the copy of the program being copied, and it is questionable whether it would have assisted Google in this case.
Under Australian law, a further source of uncertainty in such a situation is that copyright is only infringed if a substantial part of a 'work' has been copied. Whether an infringement would be found in this situation would depend on whether the declaring code copied, or the API, or a wider package or suite, amounts to the 'work'.
Developers of proprietary software that have operations in the US can continue the widespread practice of using functional code from APIs, if such use falls within the US Supreme Court's analysis of fair use. The type of code being copied will be critical to this question.