The Court of Appeals for the Federal Circuit’s second decision in the long-running Oracle v. Google copyright case is astonishing, since it is the first time a federal appeals court has reversed a jury verdict on copyright fair use. But, it’s not surprising – the CAFC telegraphed its views on Google’s fair use defense in its first decision, which held that Oracle’s Java declaring code was copyright-protected (“Google overstates what activities can be deemed transformative under a correct application of the law”).
Like its 2014 decision, the 2018 decision (decided by the same 3-judge panel) rejecting Google’s fair use defense has triggered a flood of articles analyzing, supporting or criticizing the decision.
Rather than rehash what other commentators have said about this case, here are what I see as the practical take-aways.
First, and most importantly: it ain’t over until it’s over. Google is almost certain to seek Supreme Court review (it did, unsuccessfully, after the 2014 decision – all the more reason to try again, now that it’s facing a trial on damages).
While the Supreme Court takes only a small percentage of cases, there’s a reasonable chance it will take this one. A significant factor will be whether the Solicitor General’s office argues in favor of appeal. It didn’t in 2014, arguing that the CAFC’s decision on copyrightability was correct, that the Java “declaring code” fell within the definition of a computer program —“a set of statements or instructions to be used directly or indirectly in a computer in order to bring about a certain result” — and not the exclusion for a “method of operation” or “system.”
However, the Solicitor General was more receptive to Google’s fair use argument: “legitimate concerns with interoperability and lock-in effects are better addressed through the fair use doctrine ….”
What position it will take on fair use on an appeal from the current decision remains to be seen. However, if I were Google I’d rather be seeking the Solicitor General’s support from an Obama Department of Justice than a Trump DOJ.
That said, the Supreme Court hasn’t decided a copyright fair use case in almost 25 years. It has never decided a software copyright case (although it tried in Lotus v. Borland, where it tied 4-4). This case is a good candidate for Supreme Court review, with or without the Solicitor General’s support, either on the copyrightability issue (the 2014 decision), fair use (the 2018 decision) or both. If nothing else, the Supreme Court could reverse the CAFC on the narrow ground that there were facts on which the jury could have found fair use, and therefore the CAFC went too far in usurping the role of the jury.
Second, this case was decided by the Court of Appeals for the Federal Circuit, supposedly applying Ninth Circuit copyright law. The case is not binding on any other circuit, not even the Ninth Circuit, from which it arose. It’s no secret that the CAFC is not viewed as an influential court on matters of copyright law. In fact, it gets very few copyright appeals (it’s primary role is that of an appeals court for patent cases, and even its performance in that context is controversial). Therefore, the case may have less practical impact than would have been the case if it were decided by the Ninth Circuit or the Second Circuit, the two most influential federal appeal courts on copyright matters.
This is not to say it is unimportant – at least for now, it will have persuasive force for courts adjudging the copyrightability of APIs and fair use in the context of API reproduction.
Third, every fair use case stands on its own set of facts, and therefore it’s difficult to apply precedent in this area of copyright law. This is particularly true with respect to software copyright law. This is both a negative and a positive for companies in a position similar to Google – negative in that uncertainty makes it difficult to chart a liability-free course through the shoals of fair use; positive in that future litigants can distinguish this case, which is based on facts that are highly specific. However, few companies can afford the enormous costs that Google has incurred in the defense of this case, so on balance the decision is a negative for companies seeking to copy an API without the copyright holder’s permission.
The bottom line: if a client asked me whether it could use an API such as Java without permission, I would tell them they could not, since they face a double legal hurdle: first, whatever federal court they find themselves in may follow the CAFC’s 2014 decision, and hold that they’ve violated the copyright owner’s exclusive rights of reproduction and distribution, at the very least. Even the First Circuit, which decided Lotus v. Borland in favor of Lotus on the grounds that the user commands in Lotus 1-2-3 were a “method of operation” is not necessarily safe for defendants. It’s been too long (over 25 years) and an API like Java is too different from the command structure in Lotus for a defendant like Google to feel secure in the First Circuit.
Second, the outcome of a fair use defense is always uncertain – that is baked into the very nature of copyright fair use. If the client is copying the API to achieve interoperability or to move an API to a new hardware platform, it may have a better argument than Google had with Java and Android (Google’s defense failed on both of these issues), but fair use is so unpredictable that in most cases it would be borderline negligence for an attorney to tell a client to rely on it as a defense.
On many occasions I have said that I would love to hear what Google’s copyright counsel advised Google when Google said it wanted to copy parts of the Java API and include it in Google’s Android smart phone operating system. It would be an understatement to say that this would have been very risky advice. Perhaps Google was advised that this course of action involved a big legal risk, and it concluded that the risk was justified. Regardless, unless Google can persuade the Supreme Court to reverse the CAFC, it faces a jury trial that could exceed $20 billion in damages.