William Malson, Associate Member, University of Cincinnati Law Review
The advent of computer programming presents a metaphysical quandary for intellectual property law. Software seems to be a “Writing” and therefore the kind of thing protectible by copyright. But despite being “written,” software has the kind of utility typically associated with the “Discoveries” protected by patent. Because the Supreme Court has yet to render a definitive ruling on which kind of protection software should receive, the owners of software have treated it in both ways. The chimeric nature of software not only creates doubt about what kind of protection should be applied to software; it also raises a secondary set of issues when courts are called on to apply the doctrines of either area of law to software.
One of the above-mentioned problems arises in the context of infringement cases under copyright law. The doctrines of merger and scenes a faire impose limits on the protection afforded to an expressive work by copyright. Each doctrine, in its own way, denies the originality of some expression. Courts are divided on how these doctrines should be applied in an infringement action. Depending on the Circuit in which the action is brought, merger and scenes a faire either go to the question of whether there is a valid copyright or exist as affirmative defenses to a claim of infringement. This article argues that software, as a unique form of expression with inherent functionality, should be treated differently from other copyrightable subject matter. Specifically, the doctrine of merger should be applied as an affirmative defense, and scenes a faire should apply to the existence of a valid copyright. Part II argues for the appropriateness of a unique application of copyright law to software by demonstrating that proposals to apply copyright law differently in different contexts are not new, discusses the uniqueness of computer programs, and argues that software should continue to be copyrightable. This framework would modify the test announced by the Second Circuit in Computer Associates International. v. Altai which is discussed in Part III of this article. Part IV discusses the merit of modifying the Altai test by removing the doctrine of merger from the Filtration step.
II. Copyright Law Should be Applied Differently to Software
A. Applying Copyright and Copyright Doctrines Differently is Not a New Proposal
In 1970, George Harrison recorded the song “My Sweet Lord,” using a progression of musical motifs identical to Ronald Mack’s 1962 song “He’s So Fine.” According to Harrison, he arrived at the musical progression independently, and the court accepted his assertion that he did not deliberately or even consciously utilize the theme from “He’s So Fine.” Rather than accept his defense of independent creation, the court determined that Harrison “subconscious[ly] knew [the musical progression] already had worked in a song his conscious mind did not remember.” This inquiry is not found in the U.S. Code but is nevertheless good law.
Applying copyright law differently to different areas is clearly not new, and neither is proposing different applications of copyright doctrines. For example, for visual works, one author has argued that the doctrines of merger and scenes a faire should only be applied if the defendant can show that the original artist copied from something else. Some courts apply these doctrines to visual works in a “reductionist dissection” approach that strips away copyright protection for obviously creative works. For example, when considering Leonardo da Vinci’s Mona Lisa,
The long dark hair is common to many women; we must disregard it.
The . . . expression on the woman’s face must also be disregarded . . . .
Posing a sitter in a window overlooking a landscape of twisting roads and bending rivers surely is a stock image in portraiture . . . . The simple frock she wears is not of intricate design and ornamentation – features that da Vinci might have claimed were products of his creative fancy and thus substantially original and protected. . . . Lastly, the colors chosen by da Vinci for the work must be ignored . . . . At the end of this reductionist dissection process, there is little or no protected element to compare to the second work.
Therefore, even if a second work is substantially similar to the Mona Lisa, the original work has so few protected elements that da Vinci has no real claim of copyright infringement. The author’s proposed solution is to require an alleged plagiarist to produce the image the original work copied from in order to trigger the application of these doctrines, showing that the plaintiff’s work was not original, and allow the finder of fact to evaluate the “total look and feel” of the two competing works, but imposing a check on the process with special verdict forms.
Computer programs should likewise be considered separately: their unique functionality and the likelihood of identical independent creation perfectly position them for different treatment.
B. Software is Unique
Software is different from other copyrightable expression because, although it’s regarded as literature, it has a function that hinges on its precise expression. Unlike other literary works, the transposition of even a single character can have catastrophic consequences for a computer program. If even one character out of hundreds or thousands is incorrect, the entire program may not run. A misplaced comma in a book, however, does not affect the functioning of the entire work.
Additionally, software is different from other copyrightable works because independent creation is much more common. Given the same idea, two people may arrive at identical expressions. In the computer context, “when specific instructions, even though previously copyrighted, are the only and essential means of accomplishing a given task, their later use by another will not amount to infringement.” Given that many programmers strive to write software in the most efficient manner, the possibility of identical expression is much higher than other contexts. And while a single line of code or a simple program may not possess the de minimis of creativity required for copyright, its inclusion in a larger work, the order in the program, and other nonliteral elements may be copyrightable. In other areas of copyright, identical expression outside of common lexical phrases simply does not occur. While certainly one can see similarities in expression everywhere—song melodies, movie scenes, photography styles—it is highly unlikely that separate authors could independently create the exact same work. Because of these unique aspects of software, the doctrines of merger and scenes a faire should be discussed separately for computer programs.
C. Software Should Continue to be Copyrightable
Given software’s functionality and the likelihood of independent identical creation, it would seem reasonable to suggest that software should be protected by patent and not copyright. I argue that copyright is an appropriate method of protection for computer programs as not every program meriting protection would be protected by patent. While patents protect any novel, nonobvious, and useful process, copyright can protect “a multitude of expressions that implement that process.” Patents are not issued merely for lines of code, but for process and apparatus. Firstly, processes must be novel in order to be patented. If the functionality of a computer program were to be described in a publication such that it could not be patented under 35 U.S.C. § 102(a), copyright could still protect the expression of that program. The originality requirement for copyright means that the work was created independently by the author, even if it’s not novel. A programmer could thus express a previously hypothesized process with computer code and still have that expression be protected. Secondly, copyright protects expression that possesses more than a “de minimis quantum of creativity.” The bar for creativity is quite low. For a patent, software must be nonobvious to a person having ordinary skill in the art under 35 U.S.C. § 103, a higher bar than the creativity requirement for copyright. This would preclude protection for obvious yet creative programs. Any app idea with an obvious software implementation that could be written by another programmer—but wasn’t—would thus receive no protection.
III. The AltaiTest
Some courts apply the three-step Abstraction, Filtration, and Comparison test for copyright protection of computer programs announced in Computer Associates International v. Altai, Inc. This test requires courts to first break down the program’s structure and “isolate each level of abstraction contained within it,” beginning with the code itself and ending with “an articulation of the program’s ultimate function.” The abstractions include the individual instructions organized into modules, the functions of those modules, their implementation, and the ultimate purpose of the program. Courts must then filter out protectable expression from non-protectable material by determining whether each abstraction was an idea or was included for reasons of efficiency “so as to be necessarily incidental to that idea;” required by external factors—scenes a faire—or taken from the public domain and is thus nonprotectable. Filtering out these noncopyrightable components allows courts to compare the copyrightable portions of the first program to the second for substantial similarity. The application of copyright law doctrines suggested in this article would modify the Altai test by removing the filtration of abstractions that are “ideas” before comparison to the allegedly infringing program.
A. Merger Doctrine: An Affirmative Defense
When the doctrine of merger is applied as a bar to copyrightability, plaintiffs must show that their software is written in a way that is not simply the most efficient or logical given its intended function. In order to meet this burden, plaintiffs would be all but required to present a second program that executes the first program’s intended function in a way not substantially similar to the first. For complex software that cost potentially hundreds of thousands of dollars to develop—perhaps millions—plaintiffs would have to bear that burden again. Defendants, however, would not bear this burden. Defendants arriving at nearly identical programs would be able to show independent creation or demonstrate that the first program’s expression was the most logical or efficient without having to rewrite the original program.
This burden is precisely what Nintendo bore in Atari Games Corp. v. Nintendo of Am., Inc. Nintendo designed its video game system to prevent it from accepting unauthorized game cartridges by programming a lock and key into both the system and cartridges. When an unauthorized cartridge was inserted into the system, the console would refuse to run the game because it detected no unlocking program. Atari attempted to reverse engineer the software and created their own program to mimic Nintendo’s cartridge key, allowing unauthorized cartridges to be run on Nintendo’s system. The reason Nintendo prevailed in its copyright infringement suit against Atari was because it was able to show that there were other ways of unlocking the gaming system—in effect, designing a second program that executes the first program’s function in a different way.
This cost is unique to copyright infringement cases involving software. The copyrightability of other types of copyrightable works does not depend on whether there were other ways of designing a writing intended to perform a certain function—nor is it used as an affirmative defense.
B. Scenes a Faire: A Bar to Copyrightability
It is “virtually impossible” to write software to perform a particular function in a specific computing environment without employing standard techniques. Since the creation of computer programs—especially complex ones—involves industry standards already, plaintiffs are in a convenient position to demonstrate that their expression does not follow standard procedure. The Altai test already tasks courts with determining whether code or modules are standard in the filtration step, and this article does not suggest that it change the point at which scenes a faire is applied. The alternative would allow copyright protection to industry-standard code or design and require defendants to show how the plaintiff’s code complies with these norms as a defense to a claim of infringement.
The current circuit split regarding the application of the doctrines of merger and scenes a faire should be resolved on a subject-matter basis. This type of proposal is not entirely new, as courts apply a modified test for copyright to song recordings, and others have proposed additional conditions on the application of these doctrines for visual works. Computers are unique and should be discussed separately from all other subjects eligible for copyright. For computer programs, the doctrine of merger should be allowed as an affirmative defense, and the doctrine of scenes a faire should be a bar to copyrightability. This modifies the Altai test’s Filtration step, eliminating potentially enormous costs to plaintiffs and putting courts in a more convenient position to filter out unprotected expression before comparing allegedly infringing programs to plaintiffs’ programs.
U.S. Const., art. I, § 8, cl. 8.
See Oracle Am., Inc. v. Google Inc., 750 F.3d 1339 (Fed. Cir. 2014).
The doctrine of merger applies when an idea, by its nature, dictates the form of the idea’s expression. Because the idea so limits the creativity of its expression, the idea and expression are said to “merge,” and be ineligible for protection under copyright.
The doctrine of scenes a faire states that certain expressive elements of works are so linked to the context of the expression as to be excluded from the ownership of any one author.
See Kregos v. AP, 937 F.2d 700 (2d Cir.1991) (Second and Ninth Circuits apply merger as a defense to infringement); Mason v. Montgomery Data, Inc., 967 F.2d 135 (5th Cir.1992) (Fifth Circuit applies merger to determine copyrightability); Reed-Union Corp. v. Turtle Wax, Inc., 77 F.3d 909 (7th Cir.1996) (Seventh Circuit treats scenes a faire doctrine separately from the validity of a copyright); Hoehling v. Universal City Studios, Inc., 618 F.2d 972 (2d Cir.1980) (Second Circuit treats scenes a faire as not copyrightable).
982 F.2d 693 (2d Cir.1992).
See Bright Tunes Music Corp. v. Harrisongs Music, Ltd., 420 F.Supp. 177 (S.D.N.Y.1976).
Id. at 180.
Michael D. Murray, Copyright, Originality, and the End of the Scenes a Faire and Merger Doctrines for Visual Works, 58 Baylor L. Rev. 779, 782 (2006).
Id. at 782.
Id. at 782-783.
Id. at 858-859.
Computer Assocs. Int’l. v. Altai, 982 F.2d 693, 708 (2d Cir.1992).
See Christian Owens, Move Over Actors And Artists, Software Developers Are True Creative Heroes, Forbes(Feb. 14, 2019, 4:12 AM), https://www.forbes.com/sites/christianowens/2019/02/14/move-over-actors-and-artists-software-developers-are-true-creative-heroes/#3726de1d7127 (discussing the “creativity” in software engineering as one’s approach to problem solving); Computer Assocs. Int’l. v. Altai, 982 F.2d 693 (2d Cir.1992).
Atari Games Corp. v. Nintendo of Am., Inc., 975 F.2d 832, 839 (Fed.Cir.1992).
David Kappos, An Examination of Software Patents, USPTO (Nov. 20, 2012), https://www.uspto.gov/about-us/news-updates/examination-software-patents.
Feist Publications, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 345, 111 S.Ct. 1282, 113 L.Ed.2d 358 (1991).
Id. at 363.
Id. at 707.
Atari Games Corp. v. Nintendo of Am., Inc., 975 F.2d 832 (Fed. Cir. 1992).
Id. at 836.
Oracle Am., Inc. v. Google Inc., 750 F.3d 1339, 1360 (Fed.Cir.2014).
4 Nimmer on Copyright § 13.03 (2019)