By refusing to hear this appeal, the highest court in the US has effectively declared that copyright may subsist in APIs - overruling Judge William Alsup's first instance decision.
What are APIs?
APIs are a set of commands, functions or procedures that a developer can use when building software for a particular operating system, application or service. They provide a developmental shortcut, in that they allow developers to build software that can talk to other software. For example, APIs are used to enable software running on Microsoft Windows to 'talk' to the underlying operating system to instruct it to print a document or send an email.
APIs are particularly useful in the context of cloud services, such as software-as-a-service (SaaS). Through APIs, developers can use the commands, functions or procedures of existing applications, obviating the need to replicate that functionality themselves. Put another way, instead of having to develop payment processing or billing functionality, a developer can use APIs to talk to, and utilise, other applications or platforms that already do these things.
The Oracle v Google case
Google pressed the Supreme Court to revisit the Federal Circuit Court's decision on copyright applying to the APIs, but without success. It argued that the Federal Circuit Court had made the mistake of characterising APIs as blended or commingled with the underlying application's software code. By viewing these two parts as a single piece of software, Google contended that the Federal Circuit Court incorrectly classified the APIs as original works deserving copyright protection, whereas they are in fact functional routines and protocols – akin to street names on street signs – that are separate and distinct from the underlying code which makes up the software application. Consequently, Google argued they should not attract copyright. This was also the view expressed by Judge Alsup at first instance, where His Honour stated:
That a system or method of operation has thousands of commands arranged in a creative taxonomy does not change its character as a method of operation. Yes, it is creative. Yes, it is original. Yes, it resembles a taxonomy. But it is nevertheless a command structure, a system or method of operation — a long hierarchy of over six thousand commands to carry out pre-assigned functions. For that reason, it cannot receive copyright protection — patent protection perhaps — but not copyright protection."
The Federal Circuit Court disagreed with Judge Alsup, and held that, even though Google had engaged in writing its own code to implement the APIs, copyright subsisted in the 'declaring code' which was made up of Oracle's headers and other basic directional signals to implement Oracle's software. The Federal Circuit Court observed that:
... although an element of a work may be characterized as a method of operation, that element may nevertheless contain expression that is eligible for copyright protection."
The decision (or, more specifically, the absence of decision by the Supreme Court) creates a potential legal minefield for software developers. Many developers have operated under the presumption that APIs are not protected by copyright. At least under US copyright law, in order to mitigate the risk of copyright infringement developers may have to license a third party's APIs, or redevelop their own API libraries with identical functionality. Either way, this may increase development costs.
The Australian position
It is not entirely clear how an Australian court would treat APIs from a subsistence of copyright perspective. A computer program, or a compilation of computer programs, is expressly referred to as a literary work under section 10 of the Copyright Act 1968 (Cth) (Copyright Act). 'Computer program' is, in turn, defined under the Copyright Act as 'a set of statements or instructions to be used directly or indirectly in a computer in order to bring about a certain result'.
In Data Access Corporation v Powerflex Services Pty Ltd, the High Court held that copyright did not subsist in 'commands' in the Data Access computer language contained in a system of computer programs known as 'Dataflex'. Data Access claimed copyright over 169 words in the computer language, many of which were ordinary English words. The third respondent, Dr David Bennett created an application development system which was compatible with the Dataflex language, using the process of reverse engineering. One issue was whether copyright subsisted in the 'Reserve Words' of the Dataflex language.
The High Court found that none of the Reserve Words could be considered individual computer programs. First, the Court found that, as each Reserve Word was only a single word, none of them could constitute a 'set of instructions' in the Dataflex language. Secondly, the Court found that the Reserve Words never intended to express a logical relationship between the function to be performed and the physical capabilities of a 'device having digital information processing capabilities'. The Court considered that other words or other strings of characters could have replaced the words of the Dataflex language and they would still perform the same function.
The High Court also found that the 'collocation' of the Reserve Words could not be a 'computer program' for the purposes of section 10(1) as the simple listing together of the Reserve Words did not cause a computer to perform an identifiable function. The words performed a function for the author (to provide familiar prompts) rather than perform a function for the computer itself.
APIs, however, operate somewhat differently from the way the Reserve Words operated in the Powerflex case. Whilst the declaring code of an API might resemble Reserve Words, executing these lines of declaring code causes a computer application to perform an identifiable function and interact with another application as a result of the underlying implementing code. Further, API declaring code expresses a logical relationship between the function to be performed and the physical capabilities of the application it relates to.
As a result, APIs could attract copyright in two ways – either in the individual header lines and declarations, or in the library of APIs that are created and collated together by the developer. Each of these is considered below.
Individual APIs
On the basis of the definition of a computer program under the Copyright Act and the Powerflex case, it is possible that APIs are 'computer programs' (and therefore 'literary works'). However, in order to attract protection under the Copyright Act, the work in question must also be an 'original' work. This requires that the work be expressed in an original manner – and not merely constitute an idea, concept or fact. In the oft-cited words of the English Court of Appeal in Hollinrake v Truswell [1894] 3 Ch 420:
[copyright] does not extend to ideas, or schemes, or systems, or methods; it is confined to their expression; and if their expression is not copied the copyright is not infringed."
Moreover, Australian courts have found that, where the ability to select and arrange particular words, figures or symbols is, by their very nature, limited, it is even less likely that copyright will subsist, since to confer copyright protection on these things would effectively grant a monopoly on an idea: State of Victoria v Pacific Technologies (Australia) Pty Ltd (No 2) (2009) 117 FCR 61.
On the basis of this case law, we think it more likely than not that an Australian court would agree with Judge William Alsup's first instance decision in Oracle v Google, in which His Honour held:
When there is only one way to express an idea or function, then everyone is free to do so and no one can monopolize that expression. ... [C]opyright protection never extends to names or short phrases as a matter of law."
This is also consistent with recent observations of the European Court of Justice that:
... [to] accept that the functionality of a computer program can be protected by copyright would amount to making it possible to monopolise ideas, to the detriment of technological progress and industrial development."
APIs as a compilation
In Ice TV Pty Limited v Nine Network Australia Pty Limited, the High Court held that copyright could subsist in a compilation where 'sufficient effort of a literary nature' is expended in expressing the information in that compilation (at [28]). The compilation in question in IceTV was an electronic program guide.
The High Court made it clear, however, that the 'background' research, skill or know-how of the compilers of the guide was not determinative as to whether a compilation was 'original'. Rather, it is the effort directed by the compilers to expressing that compilation in a particular way. In the words of the High Court:
[T]he critical question is whether skill and labour was directed to the particular form of expression of the time and title information [in the electronic program guide], including its chronological arrangement. The skill and labour devoted by Nine's employees to programming decisions was not directed to the originality of the particular form of expression of the time and title information. The level of skill and labour required to express the time and title information was minimal (at para 54). [emphasis added]"
Accordingly, in order to obtain copyright protection in Australia for a compilation of APIs, the developer would need to establish that there was originality in the mode of expression of the APIs (as compiled into the API package). We think it possible, though unlikely, that the effort of producing a compilation, hierarchy or taxonomy of APIs would be considered as being directed towards the expression of the APIs (and therefore deserving of copyright protection).
Conclusion
From an Australian copyright law perspective, the requirement for originality may well be the hurdle that denies copyright protection to APIs. This may all seem like an obscure technical point for copyright and programming boffins. However, consider this – if APIs are protected by copyright, what other common languages, operations, protocols or methodologies could be treated in this manner?
One commentator posits that human language itself – the process by which humans use common protocols (vocabulary and grammar) to interoperate with each other – may fall within this expanded ambit of US copyright protection, and that:
... the standardization of languages and protocols that are the foundation of the Internet could become balkanized by claims of ownership and intellectual property. That the mere speaking of a language, be it Klingon or code, could subject one to violation of [US] federal law."
Oracle v Google will now be remitted back to the US District Court who will consider, amongst other things, whether Google's use of the APIs falls within the defence of 'fair use' under the US Copyright Act.
