It’s Becoming Harder and Harder to Protect Software Function with Copyright

In SAS Institute, Inc. v. World Programming Limited, Case No. 2021-1542 (Fed. Cir. April 6, 2023) (opinion here), the general question raise is as to the scope of protection afforded to computer software and computer software function specifically. SAS offers a software product called the Sas System that enables computer users to analyze and visualize data. WPL offers a competing product – agreed to be based upon an earlier, unprotected version f the SAS system – that performs much the same way. Importantly for consideration of this opinion, the WPL system is generally compatible in the formats of data and expected language of the software used. This benefits WPL in offering a lower-cost and generally compatible alternative to the SAS System to its customers and to SAS System customers.

SAS sued, claiming among other things, copyright infringement. At the district court, Judge Gilstrap in the Eastern District of Texas found that SAS held valid copyright in some software elements, that those elements were not included in WPLs version of its software, and then applied the so-called abstraction-filtration-comparison test found that the elements claimed to be protected by SAS were not copyrightable. The suit was dismissed with prejudice. A patent claim was present in the case early, but later dropped from the case. So, the appeal here was to the Federal Circuit which challenged the district court’s determination of copyright ability generally, the use of a “special hearing” to determine copyright ability, and the court’s rejection of SAS’s expert report.

Following Google v. Oracle America, copyright protection in software has generally been viewed to extend not only to the literal elements of the software as a work of literature, but also to the overall structure of the software. In Google, this structure came down to the function calls and variable names that were part of the overall structure of the Java programming language and that were coopted for Android. These are generally called the “nonliteral elements” of the software. The same issue arose here. SAS and WPL appeared to be in agreement that the literal elements that were copied were not protectable as from an open source version of SAS. But, the structuring of the output and operation methods for the WPL software, SAS argued, was an infringement of the non-literal elements of its software.

To address the non-literal elements, several circuits – and here the Federal Circuit – have adopted the so-called abstraction-filtration-comparison test to determine the scope of copyright protection for non-literal elements of software. First the court breaks down the software into its constituent parts – e.g. an abstraction. This helps the court to separate ideas from expression – a fundamental element of copyright protection. Ideas cannot be protected, but expressions of those ideas can. Next, the court filters out all non-protectable material. These are merely ideas, as opposed to specific expression. Finally, the court is asked to compare the remaining “for protectable expression” with the allegedly infringing program to see if there is substantial similarity. The disctrict court’s decision was on the first two portions of this test. Judge Gilstrap found no copyright protection extended to non-literal elements of the SAS software. The Federal Circuit affirms.

In particular, SAS argued that WPL copied its “input formats” which are essentially a format or syntax for input of SAS-style data for the SAS system as well as its “output designs” which are similar, but for output from the SAS System. The EDTX asked SAS to come forward with proof its code’s copyright ability, which it did, then shifted the burden asking WPL to demonstrate uncopyrightability, which it did. In essence, WPL demonstrated that many of the input and output formats were identical with open source SAS or very nearly identical. The district court then required SAS to demonstrate which specific elements, copied, were projectable and found that SAS failed to show any.

On appeal, the Federal circuit indicates that the district court largely comported with the abstraction-filtration-comparison process and that SAS failed to perform any filtering. Therefore, its claims to protectable non-literal elements failed. The procedure of the copyright ability hearing was well within the discretion of the district court and was one option among many that the district court could have used to establish copyright ability of a claimed work. In addition, the district court’s decision to exclude testimony was considered under an abuse of discretion standard. At that point, the issue was basically decided. The SAS expert pointed to copying, instead of to copyrightable subject matter, as evidence that it was copyrightable. The district court was correct to exclude this expert testimony in addition because SAS and its expert failed to articulate its infringement theory in any meaningful way.

This case can be viewed as narrowing the opportunity for protection of non-literal elements by the Federal Circuit adopting the abstraction-filtration-comparison test. This forces copyright plaintiffs, even ones with one or more valid registrations, into a process of demonstrating a right to protection twice over (abstraction and filtration) before the court will even reach infringement. It also appears to make it more likely that cases will be dismissed on summary judgment as lacking in protectable subject matter, rather than ever reaching a jury.

Judge Newman offered a strong dissent. She argues for a strong, but narrow copyright protection. Her essential argument is that the court’s adoption of this test and its application to “filter out” elements as unprotect able is in contravention of the small, narrow display of creativity required to qualify for copyright protection. She points to cases related to the “selection and arrangement” of well-known data or to conform with the explicit requirements of statute. Even those types of rote materials can and do qualify for protection. She argues that SAS’s program is entitled to protection. The abstraction-filtration-comparison process is better suited (and only properly applied) in the context of an infringement analysis. Once copyright ability is established – and it should be an easy bar – then the court may consider what elements are protected while deciding what is infringed. That, she argues, is the correct approach and one that is thrown out here.

And Judge Newman Under Pressure

Over the last week or so, several reports regarding the Federal Circuit have indicated that Chief Judge Moore has threatened Judge Pauline Newman with a petition to remove her from the bench if she doe snot agree to take senior status. The threat apparently alleges that Judge Newman is incompetent to continue working in her post. There appear to be claims that she is exceedingly slow to issue opinions and has to be cajoled into issuing opinions. A brief scan of her history on Pacer indicates that she is by-and-large a dissenting voice on the Federal Circuit, at least in the last several years. She tends to favor protections for patents and copyrights, and to speak out against opinions that undermine rights holders. She also has a long view of cases. In SAS, her dissent cites the copyright act and cases from the 1950s and 1970s for propositions related to software. She was appointed in 1984 by President Reagan, and has served at the Federal Circuit almost forty years.

The Judicial Counsel of the Federal Circuit unsealed several orders relating to this issue which appear to confirm the earlier reports. Among them appear to be issues having orders issued and an outright refusal to participate in the hearings and proceedings related to the complaint. Neither are nice to have been made publicly-available, but they do appear to demonstrate somewhat of a pattern. For example, SAS v. World Programming was argued January 24, 2022 and the opinion issued April 6 of 2023. That is a lapse of nearly 18 months from argument to decision. There may be some merit to the position of Chief Judge Moore that Judge Newman’s decisions take longer than necessary.

In contravention, several attorneys and peers have spoken out saying that they have recently spoken to Judge Newman or have heard her speak and debate at professional settings in the last several weeks. All appear to suggest she is quite cogent, thoughtful, and completely in control of her faculties. One suggested that she discussed 101 jurisprudence with some clarity. Anyone who can speak to 101 has to have some scruples about them.

Judge Newman is almost 96 years old. It is unclear why she is uninterested in senior status. Though, perhaps she enjoys the mental engagement and the overall process of her work and has no wish to slow down. Some other prominent jurists appear to have held similar views, working as long as they were able, and by all appearances working thoughtfully and well in that time.


Jonathan Pearce, the author of this post, will lead a discussion of this case at our weekly SoCal IP Institute Meeting on April 17, 2023.