The code underlying Google’s Android operating system includes 11,000 lines of code copied directly from Java, an Oracle software platform that helps developers’ software run on multiple operating systems. This Fall, the Supreme Court will determine if such seemingly blatant copying is Copyright infringement. The question gets more interesting when one considers that this code is actually less than 0.5% of the entire Java code and, according to Google and a large portion of the software community, purely functional. If the Court sides with Oracle, this new view of copyright protection for software could upend an entire industry.
Java is a programming language initially developed by Sun Microsystems and later purchased by Oracle, and is at the center of a decade-long lawsuit between Google and Oracle that some say will have far-reaching consequences for the future of software innovation. Java is, in simplistic terms, a programming language that allows a developer’s code to be executable across platforms, and operating systems, such as Apple or Microsoft. This unique functionality has made it incredibly popular among software developers as a foundation for starting their code.
Java has always been freely available to developers writing apps. However, if a company wanted to use it in a platform or a new device, then a license was required. If the developer did not want to pay the licensing fee, they had to develop their own set of interoperable code. This is the route that Apple and Microsoft chose back in the early 2000s.
Around the same time, Google also avoided these licensing fees and developed its own operating system, Android. Yet, Android based its code on Java–it wanted developers to be able to utilize the cross-platform benefits of Java, but without paying the licensing fees that Sun demanded.
Unlike Apple and Microsoft, Google did not start completely from scratch. Instead, it made use of a small but essential portion of Java, API’s to be precise, to develop the Android OS and ensure that developers of Android apps could make use of Java.
Where It All Went Wrong
While Oracle was victorious at the Federal Circuit, successfully arguing that Copyright protection could apply when a mere 0.5% of the Java code was copied, when Google first approached Sun Microsystems over a decade ago, Sun saw Google’s use of Java as a business opportunity rather than a Copyright risk.
According to a 2012 article in WIRED, Jonathan Schwartz – the former CEO of Sun Microsystems said that Java had always been free to use and that although Sun Microsystems did not necessarily like the way Android used Java, it had no intention of stopping it.
However, times change. In 2010, Sun Microsystems was purchased by Oracle. Oracle saw how useful Java was and started monetizing it, clamping down and enforcing copyrights. Oracle realized that Google should have taken a license for Java when it began developing the Android platform and brought a lawsuit for copyright infringement and patent infringement against the software giant. The patent side of the lawsuit failed when the trial kicked off in 2012, and that should have been the end of the battle on the functional side of protecting Java. Yet the copyright side of the lawsuit focused on Java APIs, something that Google has long argued are also functional and thus shouldn’t be subject to Copyright protection.
Google won at the trial court, the Federal Circuit then overturned the judgment and ruled in favor of Oracle, and sent it back down to the trial court, who again ruled in favor of Google on a subsidiary point. I thought it interesting that the district judge that initially ruled Google’s favor had been a coder–if ever a trial judge was well-qualified to digest the facts of a complex case, this was it. And yet, the Federal Circuit overturned the trial court twice! But I digress.
That ruling then went back up to the appellate court, and the appellate court again reversed and held in favor of Oracle. The lawsuit is now at the Supreme Court with Google appealing that decision and arguing that the Federal Circuit was incorrect.
At the moment, most industry players and academics are on Google’s side. There are only a handful of entities, including the government, who are pushing Oracle’s side. Which brings us to the question of what the Supreme Court is looking for?
Related: Software Patents – “That’s Where All The Action Is”
So, What is So Complicated About This Question?
Patents cover the utilitarian aspects of software; copyrights focus on artistic expression, i.e., what is free speech, what is an expression, what is creative? If we look to the Court generally, not just in IP cases, we see that expression reaches pretty far–look at Citizens United, where the Supreme Court found that even financial contributions of a company to a political campaign are considered expressive and thus afforded greater protections. The boundaries of what is considered speech or expression or creativity are pretty broad.
Given that context, is it unreasonable to consider software APIs to be expressive? When we look at software, there is a fascinating question: Where do expression, creativity, and art end in something so functional? Every line of code is accomplishing a purpose. You are not painting a picture. It is not art in the traditional sense so, is code an expression? Is there creativity involved?
Absolutely, yes.
Just as photographers can take the same picture of the Eiffel Tower in different ways, there are various ways to achieve the same goal in coding, and even in creating the APIs. There are a hundred different ways to perform a lookup of an image in a database, some of those are more elegant than others, some are faster than others, some are more accurate than others. So, there is creativity involved. The broad legal definition of creativity or expression that the courts have created seems to cover any instance where a person chooses one option from among many to achieve a goal in art or software.
That leads us back to the Supreme Court’s question about whether certain types of software are not to be afforded copyright protection. What happens when you have only one way to write code? Does that mean that copyright law does not apply? The courts generally say yes. Copyright should only apply if you have options and if there is creativity in the development of the code. This whole case all comes down to whether this 0.5% of the entire code, these 11,000 lines that Google copied – involved expression or functionality. Were there different ways to write this code?
Copyrighting APIs
If the Supreme Court decides that companies can copyright APIs, there could be a knock-on effect on innovation. If developers can start protecting and monopolizing the ways that different computer systems talk to each other (i.e., APIs), popular opinion is that innovation will be stifled. The software industry may have to pay for previously free APIs and all the legal hoops that a developer would have to jump through would greatly impede the speed of coding.
Yet, these fears are a bit theoretical. Just about every other technical field has long been weighed down by the ‘oppressive’ patent system and its granting of monopolies to companies left and right. Yet, we have seen incredible innovation across the board in the past four decades despite a very strong and litigious patent system. Even software developers have had to deal with patent licensing regimes. But they simply look at a patent, and either decide to pay a small licensing fee and pass that cost along to the buyer, or they design their own code. That is innovation.
Given the innovations that many fields, including software, have seen despite the burdens of patent protection, there is little evidence that the Supreme Court’s decision to allow Copyright protection of APIs would have any significant stifling effect on software innovation. Even if fees began to be charged by copyright holders, those fees would likely be relatively small–the market won’t allow excessive licensing fees–developers would just choose to innovate and go in another direction rather than pay to copy existing code. Cross-licensing will also take place, where the Oracles of the world are saying, ‘well, we’ll give you 100 licenses Google if you give us 100 of your licenses’, and in the end, no money is exchanged. Licensing and payments do not significantly stymie innovation, and our experience with patent protection shows that they do not prevent innovators from creating the functionality they need. The market won’t allow it.
My Judgment
If I were judging this case, I would side with Oracle because of the way that we have defined copyright as protecting expressions, as compared to underlying functionality. The Java API that Google made use of was not solely an overarching idea; it was one of many ways to implement an overarching idea. That makes it creative and expressive. Therefore if that is the way we have decided to define copyright protection, then this should be protectable.
Policy discussions are not supposed to factor into Supreme Court decisions, and too often, people, and even some of the justices, go down that route. Policy is dictated by Congress and the laws they pass. The Supremes merely interpret those laws. So, the Supremes will hopefully not be influenced by the policy-driven arguments of Google and its ilk, and instead focus on interpretation of the law, which I think falls squarely in Oracle’s favor. If that result leads to negative ramifications for innovation, then Google, Amazon, Microsoft, etc. should turn to Congress, not the courts, to whittle away at copyright protections.
Related: Taylor Swift, Big Machine and Audible: The Battle over Copyright Control