Opinion
Computing Profession Legally speaking

API Copyrights Revisited

Deliberating on the main arguments in recent sets of briefs filed in support of Google's U.S. Supreme Court petition.
Posted
  1. Introduction
  2. The Computer Scientists' Amicus Brief
  3. Software Company and Related Amici Briefs
  4. IP Scholar Amicus Briefs
  5. Conclusion
  6. Author
judge gavel on table

Are application program interfaces (APIs) "original expression" that copyright law protects from unauthorized implementations in computer program code? Or are they too functional to be within the scope of protection that copyright law provides to computer programs? Alternatively, should it be fair use for unauthorized persons to reimplement APIs in independently written code?

Between 1992 and 2014, federal appellate courts in the U.S. were generally in agreement that interfaces necessary to achieving compatibility among programs were unprotectable by copyright law. These rulings made it unnecessary for courts to address fair use defenses for reuses of APIs.

In a 2014 decision, the Court of Appeals for the Federal Circuit (CAFC) overturned a trial court ruling that the Java API declarations that Google used for its Android platform were not protectable by copyright law. The CAFC ruled that these declarations were indeed original expressions that copyright should and did protect because they were creative and there was more than one way to name particular Java command terms (for example, Arith. larger instead of Math.max).

In its 2014 decision, the CAFC did not decide that Google’s use of the Java API declarations in Android necessarily infringed. Because Google had raised a fair use defense to Oracle’s claim of infringement, the CAFC sent the case back to the lower court for a trial on that defense.

After a two-week jury trial in 2016, Google’s fair use defense prevailed. However, Oracle once again appealed to the CAFC, arguing that no reasonable jury could have found Google’s use of the Java declarations to be fair and non-infringing. The CAFC agreed and remanded the case to the lower court to consider Oracle’s damage claims.

To forestall a damages trial, Google has asked the U.S. Supreme Court to review both the CAFC’s copyrightability and fair use rulings. This is not its first such request. Google previously sought Supreme Court review of the CAFC’s copyrightability ruling, but the Court decided against hearing this appeal, perhaps because the fair use issue had yet to be considered.

Fifteen amicus curiae (friend of the court) briefs were filed in February 2019 in support of Google’s Supreme Court petition. This column discusses the main arguments developed in three sets of these briefs: one filed on behalf of 78 computer scientists, two filed by software companies, and one filed by 65 intellectual property (IP) scholars.

Each brief sought to provide a unique perspective that would help the Supreme Court understand the negative implications for the software industry if the Court does not overturn the CAFC’s copyrightability and/or fair use rulings. After Oracle files its opposition brief, the Supreme Court will decide whether to grant Google’s petition.

Back to Top

The Computer Scientists’ Amicus Brief

Among the 78 computer scientists who joined this brief were 12 ACM Turing Award recipients, 24 ACM Fellows, 11 IEEE Fellows, 14 American Academy of Arts and Sciences Fellows, six National Academy of Sciences Members, 24 National Academy of Engineering Members, and five National Medal of Technology recipients, as well as professors from numerous universities.

The main goal of this brief was to explain in non-technical terms the significance of the computer science distinction between interfaces and implementations. The brief asserts there is a long-standing consensus among computing professionals that interfaces were and should be free for all to use as long as programmers reimplement the interfaces in independently written code. The CAFC failed to grasp this distinction in its copyrightability ruling. Instead, the CAFC characterized the Java API declarations as source code, which it thought Google had literally copied in Android.

The brief also explained various factors that constrain design choices about API command names. "While software interface designers have some choice for naming methods and inputs, the method’s function, word length, and clarity constrain their choice. Particularly for programming language interfaces, which define the most basic commands used across programs, there are few practical options for naming declarations that satisfy these constraints." By contrast, there are many different ways to reimplement interfaces in computer program code to carry out the identified functions.

The scientists’ brief also criticized the CAFC for rebuffing Google’s argument that its reuse of the Java API declarations had fostered compatibility. In their view, Google’s use of the declarations had "empowered software developers to write Java programs that run equally well on both desktops and smartphones." The brief concluded the Court should take the appeal to correct the erroneous analysis on which the CAFC based its copyrightability ruling.

Back to Top

Software Company and Related Amici Briefs

Red Hat, Mozilla, and Python Software were lead amici in three briefs filed by software companies in support of Google’s petition on the copyrightability issue. Red Hat, for instance, characterized the CAFC’s copyrightability ruling as having upset settled expectations in the software industry and as posing a deep threat to the open source software industry, which depends on freedom to reimplement APIs to create interoperable programs.

The Red Hat brief pointed to prior legal decisions establishing a bright line rule that APIs necessary for compatibility are not protectable by copyright law. Such a rule is preferable to reliance on fair use defenses, which create legal uncertainty and thereby impede innovation.

Microsoft filed a brief that heavily criticized the CAFC’s fair use ruling. It viewed that court’s framework for engaging in fair use analysis to be rigid, unduly narrow in its focus, and misguided. The CAFC had failed to give due consideration to the highly functional nature of software in software fair use cases and to the transformative character of many software reuses.

Microsoft pointed to prior cases that had "accommodate[d] the practical need for third parties to access and reuse functional code—like the software interfaces at issue [in Oracle]—to ensure the availability of programmers and to facilitate interoperability across myriad software platforms and hardware devices." Microsoft also asserted the software industry needs a flexible, robust, and balanced framework for fair use analysis, which the CAFC had failed to provide.

Several other organizations also filed amicus briefs in support of Google’s petition. They included the Computer and Communications Industry Association, the Developers Alliance, the American Antitrust Institute, the Electronic Frontier Foundation, R Street Institute, Engine Advocacy, and a group of software innovators, startups, and investors. These briefs tended to emphasize the exceptional importance of the case to explain why the Court should hear Google’s appeal and the risks to fair competition and ongoing innovation in the software industry if the CAFC is not overruled.

Back to Top

IP Scholar Amicus Briefs

Two of the three IP scholar briefs focused on the CAFC’s copyrightability ruling, while a third focused on the CAFC’s fair use ruling.

For the most part, these briefs do not directly address the merits of the CAFC’s Oracle decisions or the analyses on which the court relied (although it doesn’t take a genius to tell which litigant’s position on the merits the briefs’ signatories generally favored).

There is an important strategic reason for eliding the merits at this stage of Google’s appeal. The Supreme Court mainly decides to review federal appellate court decisions when persuaded there is a split among federal circuit court interpretations of federal law. U.S. copyright law is supposed to be uniformly interpreted and applied throughout the nation.

So if the Third and Federal Circuits take one position on the protectability of program APIs and several other circuit courts take the opposite position, that constitutes a split among the circuits. Circuit splits make federal law into a disharmonious mess.

A second reason the Court sometimes takes appeals is that a lower court ruling is inconsistent with the Court’s previous rulings on the same or similar issues. So the IP scholar briefs also focus on the Oracle decisions’ misapplications of prior Supreme Court rulings.

These two considerations explain why an amicus brief on behalf of 65 IP scholars, of which I was a principal author, concentrates on respects in which the CAFC’s copyrightability ruling is inconsistent with a key Supreme Court precedent as well as with rulings by several other circuit courts.

Here is an excerpt from our brief to give you a sense for this type of argumentation: "Beyond merger, the Federal Circuit’s interpretation of the scope of copyright protection available to software innovations conflicts with the rulings of other circuits in four respects. First, the Federal Circuit’s interpretation of the exclusion of methods and systems from copyright’s scope under 17 U.S.C. § 102(b) is contrary to the First Circuit’s interpretation in Lotus Development Corp. v. Borland International, Inc., 49 F.3d 807 (1st Cir. 1995), aff’d by an equally divided Court, 516 U.S. 233 (1996).

"Second, several circuit courts have ruled in favor of compatibility defenses in software copyright cases. Only the Third and Federal Circuits have rejected them.

"Third, the Federal Circuit’s conception of ‘structure, sequence, and organization’ (SSO) of programs as protect-able expression as long as it embodies a modicum of creativity conflicts with the Second Circuit’s landmark decision, Computer Associates International, Inc. v. Altai, Inc., 982 F.2d 693 (2d Cir. 1992). Altai rejected the conception of SSO as determinative of protectable expression. Id. at 706.

"Fourth, the Federal Circuit’s assertion that copyright and utility patents can provide overlapping protection to program SSO is in conflict with [the Supreme Court’s] Baker [v Selden decision] as well as Tenth and Eleventh Circuit decisions."

We IP scholars will have a lot to say on the merits if we get to that. But the main goal of our brief is to persuade the Court to grant Google’s petition because of conflicts with Supreme Court precedents and splits among circuit court rulings on software copyright issues.

Back to Top

Conclusion

Google is fighting a steep uphill battle to persuade the Supreme Court to review the CAFC’s Oracle copyrightability and fair use decisions. Every year the Court selects only about 80 cases to review out of the 7,000–8,000 petitions that ask the Court to take their appeals. Thus, the odds are heavily stacked against the granting of any petition.

Two things seem to tip in favor of the Court’s grant of Google’s petition. First, there genuinely are split decisions among federal circuit courts on important scope of software copyright issues. Indeed, the Court recognized such a split in 1995 when it granted Lotus’ petition for review of the First Circuit’s Borland decision. Because the Court itself split 4-4 in that case, the First Circuit’s decision was affirmed, but without setting a precedent or resolving the split. Unsurprisingly, the circuit splits on software copyright protections have deepened since then.

Second, the specific software copyright issues presented in the Oracle case are of exceptional importance to the software industry, of which Oracle and Google are two giants. Other major software developers, including Red Hat and Microsoft, and industry associations filed supportive amicus briefs to stress the important consequences at stake in this case.

If the Court does decide to hear the Oracle case, it is, however, unlikely to review both the copyrightability and fair use rulings. Most amici, including me, who support Google’s petition would prefer the Court took the copyrightability issue rather than the fair use issue. But the fair use decision is also deeply flawed. So a grant on either issue would be much better for the software industry than a denial of review.

One other interesting data point is that since 2007, the Court has reversed whatever circuit court ruling it reviewed 70% of the time. That does not mean Google will definitely prevail before the Court, but its odds of winning improve if the Court takes the Oracle case on the merits. One bit of possibly good news for Google is that the Court has asked the Solicitor General’s opinion about whether it should take the case. That somewhat increases the chance that the Court will indeed hear Google’s appeal.

Back to Top

Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More