I’m happy about the Supreme Court’s 6-2 decision in Google v. Oracle, affirming that Google’s reimplementation of the Java API for Android was a fair use and thus not copyright infringement. Google and Oracle’s competitions can breathe a sigh of relief, since a result that went the other way could have seriously inhibited the ability of smaller companies and new entrants to compete with established software giants.
Oracle is less happy, and frames the decision differently. In its statement, it says that the decision makes “the barriers to entry higher and the ability to compete lower.” The opposite is true.
It is correct that only companies like Google can afford to litigate a case for a decade, all the way to the Supreme Court. But now that it has done so, others can rely on the precedent it set, and its holding that reimplementing a software API to create a competing or compatible version is a fair use as a matter of law. This is a very strong fair use opinion that would seem to cover most kinds of API reimplementation, which the court correctly found is essential to competition, new market entry, and interoperability.
It is also true that even with strong precedent on your side, you might still have to litigate a fair use case, since fair use is so fact-specific. In many ways it would have been cleaner if the Court had recognized that APIs should not receive any copyright protection at all, as Public Knowledge asked the Court to do. At the same time, the Court’s reasoning is not limited to APIs and would seem to cover other forms of compatibility, cloning, and reverse engineering. Some of the Court’s reasoning might have applicability beyond software—not just its lucid description of the purpose of copyright itself (to promote the creation of new works, which includes allowing the creation of new works via fair use) but in other areas. For example, in its analysis of the fourth fair use factor—effect on the market—the Court noted that plaintiffs cannot simply assert that they have lost revenue from the fair use itself, as this is simply circular, and that courts must look to the broader public benefits that could result from a fair use holding, not just the effect on the parties to the case.
Back to Oracle’s claim: Does this decision help incumbents? It obviously helps Google avoid paying damages this one time. But if you look to history, generally smaller companies and new entrants are pro-compatibility, and incumbents are against it. For example, in Lotus v. Borland, the dominant spreadsheet company sued a new entrant for copying its menu hierarchy (and lost).
There are many similar examples from the history of the computing industry. The amicus briefs in this case are littered with them. Perhaps the most famous example of how reimplementing APIs benefits competition and new entry is how Phoenix Technologies cloned the IBM PC BIOS, which enabled the creation of the PC clone industry itself (which is now just known as the PC industry, IBM having lost its dominance and exited the market years ago).
Android itself was a new entrant operating system when it reimplemented the Java API. Now, of course, it is the world’s dominant mobile operating system. But this decision could make competing with Android much easier.
Android is an open source operating system at its core, but Google licenses several proprietary components to device manufacturers, and many popular apps don’t work unless they’re present. By limiting access to the Play Store, to Google Maps, and associated APIs, Google is able to have an “open” ecosystem that it nevertheless tightly controls.
The Supreme Court’s decision makes it clear that projects like microG, which clone Google’s proprietary Android components, are on a firm legal ground. An app might need access to Android’s proprietary location API, for instance, and can’t function without it. But an app doesn’t need to know whether it’s interacting with Google’s official location API or a compatible clone. An alternative needs to be a drop-in replacement that requires no work from application developers, who are unlikely to put in the work to function with an incompatible product that may never succeed. microG and efforts like it make it so there can be fully-featured versions of Android entirely free of Google’s control. (Companies who want to ship an official version might be contractually restricted from shipping a forked version, which is a restriction competition authorities around the world have put under scrutiny.) Google can compete with these efforts by continuing to improve the quality of its location service, but not by monopolizing methods of interacting with location services generally.
In fact, speaking of interoperable clones, in researching this blog post I came across this tutorial: a “A step by step tutorial to building a Google Maps Clone using React & MapBox API” in twenty minutes. Sounds fun! I’d also like to see someone reverse engineer the Gmail API, that the official Gmail app and third-party alternatives use to communicate with Google’s servers. It does more than the open IMAP protocol, and open successors to IMAP such as JMAP have not gotten the traction I would have liked.
Of course, cloning an API doesn’t mean that third parties can just access a cloud service of some kind for free, even though the term “API” often refers to an API for doing just that. Cloning a cloud service means that you can create a competing cloud service that works the same way, to make it easier for customers to switch. This is exactly what Oracle itself has done, sensibly cloning some Amazon AWS APIs so that it can better compete with Amazon’s dominant cloud offering.
The software industry has operated under the assumption that cloning APIs and other methods of organizing functionality for decades, despite the periodic attempts by established firms to cut off their competitors. So in some ways this decision just affirms what was thought to be the legal status quo. But this litigation has been going on for a decade, and has spurred other litigation—such as when Cisco sued Arista Networks for, among other things, copying its command structure. (Public Knowledge filed a brief in this case.) This decision might thaw out some of the chilling effects on competition that Oracle’s litigation created, and could even provide the impetus for more compatibility efforts. The technology sector needs more competition, not less, and the Supreme Court’s decision in this case will help bring that about.