The fact that the Supreme Court has asked the Obama Administration (via the Office of the Solicitor General) to comment on Google’s application for certiorari in Oracle v. Google* has focused renewed interest on this case – not that it needs it. The case, if the Supreme Court accepts it, could be a replay of Lotus v. Borland, an important software copyright case the Supreme Court tried but failed to decide in 1996, when the Court deadlocked 4-4 (one justice abstaining).
*Note: For detailed procedural and substantive back ground on this case see these earlier posts: How Google Could Lose on Appeal; Oral Argument in Oracle v. Google: A Setback for Google?; CAFC Reverses Judge Alsup – Java API Declaring Code Held Copyrightable; Google Rolls the Dice, Files Cert Petition in Oracle Copyright Case. I also made a presentation to the Boston Bar Association on this case before the CAFC decision, slides here.
The issue is this: computer software — both source code and object code — is protected by copyright law so long as it meets copyright’s statutory requirements, the most important of which, for purposes of the case discussed in this post, is originality. At issue are the 7,000 lines of “declaring code” of Oracle’s Java API software. This software was copied by Google when it implemented the Android smartphone operating system. The Java API declaring code clearly satisfies copyright law’s requirement of “originality.” The issue, therefore, is why this code should not be protected by copyright.
*Note: APIs are specifications that allow programs to communicate with each other. A computer program designed to be compatible with another program must conform to the API of the first program, which establishes rules about how other programs must communicate information so that the two programs can work together to execute specific tasks.
The argument Google has made to the Supreme Court is that the software a “method of operation”* that should be denied copyright protection?
*Note: The Copyright Act states that “[i]n no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work.” 17 U.S.C. § 102(b)
This issue has bedeviled the courts in this case. A federal district court judge in California held that Oracle’s Java API was an unprotectible method of operation. The Court of Appeals for the Federal Circuit reversed, holding that copyright protects the expression of a method of operation, and therefore the Java API is protected by copyright.
Now the software industry and copyright lawyers are eager to learn whether the Supreme Court is likely to accept review of this case, and the Court’s request that the Obama Administration weigh in thickens the plot by suggesting that the Court is giving Google’s appeal serious consideration.
The Supreme Court takes about 1% of the petitions filed with it, and despite its request to the Office of the Solicitor General I don’t think this case is likely to make the cut. Why is that so?
Setting aside the question of copyright fair use (which remains pending and will be retried on remand if the Supreme Court denies cert.), the core question is whether the Java API — 7,000 lines of programming code — is protected by copyright law.
The question is, why would the 7,000 lines Google copied from the Java API not be copyrightable? Google argues that the API code is not copyrightable because it is a “method of operation,” a category the Copyright Act expressly excludes from copyright protection. Google is inviting the Court to pick up, once again, an issue it proved unable to decide in 1996. In that case, Lotus v. Borland, decided by the First Circuit Court of Appeals based in Boston, the First Circuit held that a menu command hierarchy in a user interface (“File,” “Print,” “Copy” ….) was a “method of operation,” and therefore excluded from copyright protection, regardless of whether it is original. The menu commands were the “means by which a person operates something” — in that case a computerized spreadsheet — and therefore the commands were not protected by copyright.
The Supreme Court accepted review of that case, but one justice recused himself, and the remaining eight judges voted 4-4, leaving the case undecided at the Supreme Court and affirming (by default) the decision of the First Circuit, where it has stood as controlling law ever since (but only in the First Circuit).
No other federal circuit court has adopted Lotus v. Borland wholesale, and several courts have finessed the issue. In addition to the Federal Circuit’s decision in the current case, at least one circuit court has expressed disagreement, but that was in 1997, eighteen years ago. In other words, this is not an issue that has been burning up the federal courts since 1996. Google does its best to elevate this to a clear circuit split (one sure way to attract the attention of the Supreme Court is to present a circuit split), but frankly its effort is not entirely convincing given the infrequency with which the the issue has reached the appeals courts.
The amicus brief filed by the Electronic Frontier Foundation and a host of computer scientists (77 scientists) is no more convincing. The scientists based their argument on the need for compatibility. They argue that “copyrightable APIs would discourage innovation by creating potential liability for the mere act of writing a compatible program.” The problem with the scientists’ “compatibility/interoperability” argument is that it goes to fair use, an issue that neither the district court nor the Federal Circuit decided. At the trial of this case the jury deadlocked on the issue, and therefore absent intervention by the Supreme Court the case will be remanded for a retrial on this defense. While the scientists argue that the need for compatibility should lead the court to take the case and hold that APIs are uncopyrightable, their argument does not clearly address the “method of operation” argument made by Google.
Moreover, the scientists ignore the fact that the Google Android-Java API is not compatible with the Oracle API, a fact that Google has conceded.
If the Supreme Court is looking for the Office of the Solicitor General to help them decide whether this case is review-worthy, Google is likely to be disappointed. First, there is no reason to believe this is an issue of interest to the Obama administration – it hardly falls within any of the social action categories the administration is focused on. Second, the case can be appealed to the Court in the future by whichever party loses the retrial on fair use. Perhaps the case will settle, or the losing party will not appeal, saving the Court (and the Solicitor General) the effort. Lastly, as noted, while there may be a split of opinion among the federal circuit courts, the decisions that create the split are spread out over two decades. This hardly amounts to a cry for help from divided federal courts. Rather, this seems to be an obscure issue which could lie dormant for another ten or twenty years. If events prove otherwise, the Supreme Court will have another opportunity to take up the issue in a future case.
The parties’ briefs filed with the Supreme Court:
Update: The Supreme Court did not accept this case for review
Second Update: On re-appeal, following a trial and second appeal on fair use, the Supreme Court accepted Google’s appeal of this case. See my Oracle v. Google Resource Page here.