How the Supreme Court can get it right for software innovation in Google v Oracle
As the US Supreme Court considers the oral arguments in the key copyright fight between two Silicon Valley giants, Mozilla’s senior IP and product counsel contends that a ruling for Oracle would be “another brick in the wall of software’s walled gardens”
More than 10 years have passed since Oracle sued Google for reimplementing Java’s application programming interface (API) packages. It has now been eight years since Judge William Alsup, who taught himself Java to better understand the case, ruled that APIs were a method of operation not subject to copyright protection. At the time, Santa Clara University Professor Tyler Ochoa wrote that the opinion “is so well-written, and so carefully reasoned, and demonstrates such a strong understanding of the technology involved, that I will be astonished if it is not upheld on appeal”. This prediction did not fare well on two trips to the Court of Appeals for the Federal Circuit. However, now the Supreme Court will give Ochoa and Alsup one final chance to be proved right.
Mozilla Corporation, which makes the open source browser Firefox, filed an amicus brief with the court, arguing that APIs should not be copyrightable or, alternatively, reimplementation of APIs should be covered by fair use. Mozilla was joined by Etsy, Mapbox, Medium Corporation, Patreon and the Wikimedia Foundation. The brief explained that software developers have long assumed that re-implementation of APIs would not lead to copyright liability. This has allowed developers to build software that is interoperable with other companies' products. But, equally important, it allows developers to build competing software that uses familiar commands and protocols. This promotes software competition in the same way that allowing rival car manufacturers to place the accelerator to the right of the brake promotes competition in the auto industry.
Oracle and its supporters argue that a ruling for Google would threaten software innovation. Narrowing the scope of copyright protection for software, the argument goes, would reduce the incentive to write new software. But this simplistic formula fails to engage with the balance at the heart of copyright law. The dichotomy between ideas and their expression and fair use ensure that follow-on innovation can also thrive – this is especially important in a field such as software, which is characterised by iterative development and interconnections.
Free and open source software depends on the ability to reimplement protocols that ensure interoperability between free, open or proprietary systems. The long-running Samba project provides an example – since 1992, this project has developed software to allow open source operating systems to communicate with Windows systems. Samba implements its own file-sharing system, but uses reimplemented Windows commands to "speak to Windows clients like a native", benefitting both the Windows and open source ecosystems. If Oracle’s vision had been reflected in the law in 1992, this key part of open source infrastructure would likely never have been built. And it is likely that Linux, which now underlies most enterprise computing, would be nowhere near as successful.
A ruling for Oracle would be another brick in the wall of software’s walled gardens. New companies will have an even harder time in the future if they cannot develop software that uses familiar APIs, forcing anyone who wants to compete to build their own technology silo. Admittedly, the negative impact of a ruling for Oracle will be mitigated somewhat by the fact that many modern APIs are available under permissive licences. But that is not true of all software and the scope of licences can raise complex questions. Indeed, serious issues have come to the fore about whether Oracle’s reimplementation of Amazon’s S3 API is licensed. Moreover, giving established players a new tool to control competitors is the last thing that we should be doing in a marketplace that is already characterised by a few dominant players. Given this, it is not surprising that so much of the software industry supports Google’s position in this case. Those on Google’s side have included non-profits, such as the Open Source Initiative and the Software Freedom Conservancy, medium-sized companies such as Mozilla, Mapbox and Patreon and giants such as Microsoft and IBM. And while "never read the comments" is usually good advice, a visit to the comment sections on sites popular with developers such as Ars Technica would give you a sense of how strongly the programming community supports the freedom to reimplement APIs.
The Supreme Court will consider both whether the Java API packages are copyrightable and Google’s defence under fair use. A finding of fair use would help promote interoperability and innovation. But this might come at the cost of complex and expensive litigation, given that fair use is a multi-factor test and typically a deeply factual analysis. The best outcome is that the court rules that APIs are not copyrightable under Section 102, which will provide the most clarity for developers. Ultimately, copyright law should serve the creators that it is supposed to incentivise.