“The value of a mystery novel is in the last page where you find out about the murder…. Google only copied 37 of these interfaces but they didn’t care about the others—only 37 of them were relevant to Android smartphones.” – Llewellyn Gibbons
Yesterday at IPWatchdog LIVE, a morning panel discussed “The Impact of Google v. Oracle on Fair Use.” That decision of the U.S. Supreme Court, issued this April, has many ramifications for the future of copyrightability of computer code as well as the analysis of transformative uses of computer code. Speaking on the panel was Chad Rutkowski, Partner at Baker & Hostetler LLP; Van Lindberg, Partner at Taylor English and Counsel and Director of the Python Software Foundation; and Llewellyn Gibbons, Distinguished University Professor, University of Toledo College of Law.
SCOTUS Punts on Copyrightability of Computer Code, Providing Very Little Clarity
Along with discussing the Java programming language at issue in the case and tracing the procedural history of Google v. Oracle, including both Federal Circuit decisions in the case, the panel focused on the problematic nature of copyright protection, which is limited to artistic expression, and computer software and source code, which is highly functional in nature. Recognition of computer code as potentially protectable by U.S. copyright law extends back to the early 1980s and the recommendations of the National Commission on New Technological Uses of Copyrighted Works (CONTU). As Rutkowski noted, however, there’s a natural tension between the goals of copyright law and the nature of computer programming languages. “We think of copyright as protecting tangible expression, but it protects what leaps off the page like plot, character and setting,” he said. “What is the plot, character and setting of source code?”
The copyrightability of software has been addressed by several circuit courts in the intervening years since the CONTU report, with varying results. In 1986, the Third Circuit decided Whelan Associates v. Jaslow Dental Laboratory, in which the court affirmed a lower ruling that federal copyright protection extended to the non-literal structure, sequence and organization (SSO) of computer code and not simply its literal elements. Whereas Rutkowski felt that the doctrine established by the Third Circuit in that case had weight to it, Lindberg contended that the appellate court’s analysis went no further than establishing that every expression of the idea of dental office software that was manifested in computer code, and that Whelan Associates has been disfavored in subsequent cases on the copyrightability of computer code. Over in the Second Circuit, the SSO test from Whelan Associates was rejected and replaced with the abstraction-filtration-comparison test for substantial similarity of computer program structure in its 1992 decision in Computer Associates International v. Altai.
One of the problematic aspects of copyrightability for the computer code at the center of Google v. Oracle, an issue which the panel largely agreed that the Supreme Court punted on in the case, is the functional nature of the declaring code copied by Google for its Android application programming interface (API). “Here, the implementing code is like a set of street signs, it only works one way and if you change one thing, it all breaks down,” Rutkowski said. “SSO in this context doesn’t make a lot of sense to me.” Gibbons raised issues with the order of the Supreme Court’s analysis in Google in which the court assumed copyrightability for computer code and then assessed fair use factors, including the transformative nature of Google’s use. Gibbons indicated that he felt Google’s use was less than transformative. “If you have a motion picture in Spanish and then you translate it to French, you may have a different product but I’m not sure how transformative it is,” he said.
Lindberg: ‘There Is Such a Thing as a Beautiful API’
Along with the copyrightability and transformative use analyses, Gibbons took issue with how the Supreme Court analyzed the fair use factor regarding the amount and substantiality of the portion of the copyrighted work that Google used. Whereas the Supreme Court focused on the fact that Google only used 0.4% of Oracle’s code, Gibbons noted that this factor is supposed to focus on the value of the portion used, not simply the percentage of the whole. “The value of a mystery novel is in the last page where you find out about the murder,” Gibbons said. “Many songs, only a short part of it is memorable. Google only copied 37 of these interfaces but they didn’t care about the others—only 37 of them were relevant to Android smartphones.”
Lindberg, who noted that he enjoys writing computer code as a hobby, said that copyright protection is appropriate for computer code that has originality of personal expression. “There is such a thing as a beautiful API,” he said. “There is expression in code, and it does actually matter how you express things.” Lindberg pushed back on Gibbons’ assertion that beauty in API and programming languages only encompasses the efficiency of language that can achieve the most in a very few lines of computer language. Lindberg pointed out that source code is primarily for human consumption, whereas a computer has to translate the language into binary code for processing. “Computers are not ideal things, and sometimes you need to make ugly hacks to make it run faster,” he said. “Computer code can be beautiful when it accurately and succinctly expresses the intent of the programmer.”