A yellow copyright symbol on a digital background

Many people were disappointed when the most-watched copyright case of the past 10 years, Oracle’s lawsuit against Google over Google’s copying Java application programming interface (API) code, failed to yield better guidance on the scope of copyright protection for computer programs. The case was probably a bad test case anyway, as it involved some very functional aspects of some very functional code – the declaring code (which acts as a set of signposts pointing the software to where certain computing assets are located) for an API (which enables interoperability between different computer programs). This rather prosaic aspect of rather prosaic code lacked the dazzle of even, say, a bookkeeping program, where the arrangement of screens and user experience might involve some obvious and welcome creativity. So when the Federal Circuit found that the declaring code was protectable because there was more than one way to word the signposts (and revived interest in the notion of code protecting nonliteral elements like sequence, structure or organization), stakeholders throughout the software industry were waiting for a Supreme Court epiphany either endorsing this rather aggressive interpretation of expressiveness in code or reeling it in.

Of course, the Court did neither. It punted. Presumably afraid that a decision one way or the other would move financial markets and have unforeseen consequences, the Court assumed the declaring code was protected by copyright and decided the case on fair use. Many Internet pundits have interpreted Oracle v. Google to mean that APIs cannot be protected by copyright, but they’re wrong – Justice Stephen Breyer went to considerable pains to point out that the declaring code (which made up about 11,500 lines of code) aspect of the API is quite different from the vast bulk of the rest of the API code, known as the “implementing code” (which made up 2.86 million lines of code). Breyer found the implementing code to be close to the “core of copyright.” A fair use of declaring code might not be a fair use of implementing code. And, of course, the Federal Circuit’s decision finding the declaring code expressive and protectable as a nonliteral expression of source code remains good law.

So it was that another case emerged shortly after the Oracle v. Google decision that might force a reckoning on copyrightability of code, reining in the scope of copyright protection for software programs and taking it out of the fickle, highly fact-dependent doctrine of fair use. The software company SAS Institute Inc. has been suing a rival, World Programming Limited (WPL), over WPL’s unabashed, blatant copying of SAS’ computer programs, filing multiple lawsuits for over a decade now, most recently in the Eastern District of Texas. SAS created a suite of software programs, known as the SAS System, used in data access, management, analysis and presentation. WPL copied various aspects of the SAS System, though interestingly, SAS did not allege that WPL had access to or copied the actual source code itself. “SAS alleges WPL copied its ‘Input Formats,’ which are the fundamentals (vocabulary and syntax) used in the SAS System. … SAS also claims WPL copied its ‘Output Designs,’ which are the result of applying Input Formats to user data.” That is, SAS’ claim, Oracle-style, was that WPL copied nonliteral elements of its code. WPL, for its part, appears to admit as much, gambling completely on the ineligibility of copyright protection for the subject nonliteral elements and/or fair use. Also, SAS’ most recent lawsuit initially included a count of patent infringement. The patent count, though, was rather quickly voluntarily dismissed, suggesting that SAS included the claim as a kind of forum shopping to get any appellate review of the case in front of the Federal Circuit and its expansive interpretation of copyright protection for source code.

You probably know that the Eastern District of Texas is a well-known magnet for patent infringement cases. You probably also know that the architect of the Eastern District’s erstwhile venue dominance is the Hon. Chief Judge J. Rodney Gilstrap, and further know that patent cases employ a procedural device known as a Markman hearing to force crisp definition of patent claims and filter out unsupportable lawsuits. SAS drew Gilstrap in its most recent case, and Gilstrap deployed the procedural device that he knows best, the Markman hearing, on SAS’ software copyright claim, creating a new filtering device known as a “Copyrightability Hearing.” Gilstrap also deployed something that is not new at all, the “abstraction-filtering-comparison” (AFC) test for assessing infringements of software copyright, first developed by the Second Circuit in Computer Associates v. Altai in 1992 and largely adopted everywhere.

The AFC test was crafted as a “let’s make the best of a bad situation” fix for the uneasy fit of computer software code under the rubric of copyright. Copyright does not protect “any idea, procedure, process, system, method of operation, concept, principle, or discovery. …” 17 USC 102. However, copyright protects “computer programs,” defined as a “set of statements or instructions to be used directly or indirectly in a computer in order to bring about a certain result.” How can copyright simultaneously protect a set of instructions used to directly or indirectly bring about a result but not also protect a procedure, process, system or method of operation? Heck if I know – and this is pretty much all I do. To try and manage this bit of cognitive dissonance, courts rely on the AFC test, whereby computer programs are abstracted out into telescoping levels of abstraction (like the actual code itself as most protectable, the ultimate function of the code as likely not protectable as an idea, and steps in between like hierarchy of modules having possible protection), filtering out of unprotectable ideas and standard coding techniques at each level of abstraction, and comparing what’s left to determine whether there has been infringement.

Gilstrap apparently thought this sounded a lot like a Markman hearing, so he ordered the parties to participate in a “Copyrightability Hearing,” and he imposed alternating burdens of proof – plaintiff SAS had to meet a burden of production on copyrightability, easily met by its copyright registrations. WPL had to rebut copyrightability with identification of unprotectable elements in the code (pure ideas/functions, math functions, conventional displays, public domain material), and SAS was to respond to the rebuttal with an explanation of why what WPL identified as unprotectable was wrong or what expressive parts they missed. Except … SAS refused to engage in that last part. According to the Federal Circuit’s opinion in this case, “SAS does not appear to dispute that its expert did not conduct the filtration analysis that the district court adopted.” SAS had several principled reasons for doing so, many of which were ably amplified by the numerous amici that participated in the appeal, including that the AFC test is an infringement analysis tool and should not be imported into the copyrightability analysis and that the burden shifting was unfairly applied as improperly requiring the plaintiff to identify every copyrightable element of its generally copyrightable work. But the district court ordered it to, and it didn’t – “[b]ecause SAS did not show what identifiable protectable elements remained in the SAS System, even after the court instructed SAS to do so, the court dismissed the case.”

So you probably know where this is going. On appeal, the Federal Circuit found that “[a] court may reasonably adopt an analysis to determine what the ‘core of protectable expression’ is to provide the jury with accurate elements to compare in its role of determining whether infringement has occurred,” found that Gilstrap appropriately exercised his discretion, found that SAS failed to “articulate a legally viable theory” and affirmed dismissal. Judge Pauline Newman dissented, asserting that the “court today holds that these software programs are not copyrightable.”

Maybe. Maybe the Input Formats and Output Designs are copyrightable. Maybe they’re not. But we’ll never know because SAS stood on principle (defensible principle for sure). What we do know is that the majority in the decision reiterated many of the copyrightable principles that animated the expansive view of protection in Oracle v. Google:

  • “As a general matter, and to varying degrees, copyright protection can extend beyond literal elements to nonliteral elements.”
  • “The nonliteral elements of a computer program are those aspects that are not reduced to written code.”
  • “These elements include the program architecture, structure, sequence and organization, operational modules, and user interface.”
  • “Using a literary novel as an analogy, the novel’s written words would be the literal elements (e.g., code) and the organization of the chapters, characters, and plot would be the nonliteral elements.”

So no, the Federal Circuit did not just kill off software copyrights. And if you’re going around saying that, you should probably stop.