Developments in Copyright and Trademark Law

I. Introduction

During the survey year, the U.S. Supreme Court issued one blockbuster decision involving copyrights on software and another significant decision involving trademark law applied to domain names. Part II discusses the Court’s decision ending a dispute between Oracle and Google over software code used in the Android mobile platform, as well as some cases on the volitional-conduct requirement and the making-available right. Part III addresses the Court’s decision about the eligibility for trademark registration of domain names having the form “generic.com” and a Sixth Circuit case about the circumstances under which an online marketplace may be directly liable for trademark infringement occurring in transactions it facilitates.

II. Copyright Law Developments

A. Copying of Application Programming Interface Procedure Names Is Fair Use

In Google LLC v. Oracle America, Inc., the U.S. Supreme Court resolved a long-running dispute between two behemoths of the computing industry, and placed its imprimatur on a type of copying that is aimed at making it easier for programmers to create new software.

The dispute arose when Google set about developing a software platform, called Android, that would run on mobile telephones. One element of the platform was a set of ready-made routines that software developers could incorporate, free of charge, into their own programs designed to run on Android phones. Using these routines would save them the effort they would otherwise need to expend to code standard and relatively low-level functions. These routines constitute what is called an application programming interface (“API”).

When Google began working on Android, there was an existing API with which millions of programmers were already familiar, associated with the Java SE programming language, which was used to write programs that ran on desktop and laptop computers. Google decided to design its Android API so that Java programmers could use it without learning a new programming language. To accomplish this, the Android API’s routines had to be organized and named in the same way as those in the Java API. The Java API organized the routines hierarchically. The individual routines, called “methods,” were grouped together into “classes,” and the classes were grouped into “packages.” Each method was also given a name. To access a particular method, a programmer would call it up using a statement reflecting both the organizational hierarchy and the method’s name. For example, to call a method that determines which of two integers is larger, the programmer would write a statement that includes the string “java.lang.Math.max,” where “java.lang” designates the package where the routine is found, “Math” designates the applicable class within that package, and “max” is the name of the method that does the job.

To mirror in the Android API the organizational and naming conventions of the Java API, Google copied about 11,500 lines of what is called “declaring code.” This is not the code that constitutes the routine and performs the desired task (such as determining which of two integers is larger), but only the code that names the routines and allows them to be called up, corresponding to thirty-seven of the Java API’s packages. Google did not copy the code that runs when a Java routine is called up, but rather wrote its own implementing code from scratch. The implementing code amounted to 2.86 million lines. The code that Google copied therefore represented only a small proportion of the code constituting the Java API.

Oracle was the successor to Sun Microsystems, the creator of Java, and owned the copyright to the code that constituted it. Oracle sued Google, claiming that Google infringed its copyright both by copying the literal declaring code and by replicating the hierarchical structure of the Java API. After a jury trial, the district court held that the API was not protected by copyright, as it constituted a “system or method of operation” beyond the scope of the Copyright Act. On appeal, the Federal Circuit reversed, holding that both the declaring code and its organizational structure were protected. After a second trial, the jury found that Google’s copying of these elements of the Java API was fair use. On a second appeal, the Federal Circuit once again reversed, holding that, as a matter of law, the use was not fair.

The U.S. Supreme Court granted certiorari and held that the Federal Circuit had erred in its determination that Google’s use was not fair. The Court declined to decide the threshold issue of whether the Java API code or organizational structure was protected by copyright. Instead, it assumed that the API was protected, and held that Google’s use of it was fair.

In its analysis of fair use, the Court applied the familiar four-factor test from section 107 of the Copyright Act. But in a departure from the usual judicial exposition of the factors, the Court began with, and accorded substantial significance to, the second factor, which considers “the nature of the copyrighted work.” The Court observed that the nature of the work in question, namely the declaring code, differs from ordinary computer code (such as the implementing code that actually performs the function for which a method is called) in that it is “inextricably bound up” with elements of the Java API that all agree are not protected by copyright: the “general system” of computing tasks, the “idea” of organizing software routines hierarchically, and the use of particular method calls. Likewise, it is “inextricably bound up” with the implementing code, which is protected by copyright, but which Google did not copy. On account of these features, the Court found, “the declaring code is, if copyrightable at all, further than are most computer programs (such as the implementing code) from the core of copyright.” This factor accordingly tends to support a finding of fair use.

On the other fair use factors too, the Supreme Court disagreed with the Federal Circuit. The Federal Circuit found that Google’s use of the copied code was not “transformative” under the first factor, because, among other things, Google’s Android API served the same function as the Java API. But the Court held that the Android API “provided a new collection of tasks operating in a distinct and different computing environment,” which “were carried out through the use of new implementing code (that Google wrote) designed to operate within that new environment.” The Court also disagreed with the Federal Circuit’s application of the third factor, “the amount and substantiality of the portion used in relation to the copyrighted work as a whole.” The Federal Circuit found that Google had copied too much, because “the parties stipulated that only 170 lines of code were necessary to write in the Java language,” but Google had copied 11,500 lines, and Google’s desire “to meet the expectations of intended customers” did not justify such a large appropriation. But the Court said that the relevant criterion was the ratio between the number of lines of the Java API that Google copied (11,500) and the number it did not copy (2.86 million), namely 0.4 percent.

The Court’s analysis suggests that courts will be more open to finding fair use in the context of computer programs when the purpose of the copying was to promote the creation of new computer programs.

B. The Volitional-Conduct Requirement

Since the early days of the commercial Internet, courts have held that a provider of online services can be held directly liable for copyright infringement only if the conduct alleged to be infringing is “volitional.” Two district courts arrived at different conclusions on the issue of whether a website operator acted volitionally when it allowed users to upload copyrighted materials.

In Sid Avery & Associates, Inc. v. Pixels.com, LLC, the court found that the defendant did not engage in volitional conduct when it operated a website that allowed users to upload photographs that others could purchase as prints or have printed on items, such as coffee mugs and tote bags. Relying on a Ninth Circuit decision, the court explained that, to demonstrate volitional conduct, the plaintiff had to show that the defendant “exercised control (other than by general operation of [its website]); selected any material for upload, download, transmission, or storage; or instigated any copying, storage, or distribution” of the photographs. Because defendant did not select the images to be uploaded (its uploading users did), or determine which items were purchased (its purchasers did), and because it in fact prohibited the uploading of infringing content and took down such content as soon as possible, the alleged infringement did not result from defendant’s volitional conduct and defendant could not be held directly liable.

In Atlantic Recording Corp. v. Spinrilla, LLC, a court reached the opposite result on somewhat similar facts by applying a notably different analysis. The website in question allowed approved users to upload mp3 files of hip-hop songs and “mixtapes” consisting of multiple songs. Other users could select from the uploaded music files and listen to them streamed via the Internet. The defendant had a policy of removing music files once notified they were infringing. The court relied on prior decisions to conclude that the website operator engaged in volitional conduct when it allowed music to be streamed from the site. Because the streaming constituted an unauthorized public performance of copyrighted sound recordings, the website operator was directly liable for the infringement.

C. Latest Case on the Making-Available Right

A district court in Washington State became the latest court to hold that merely making a copyrighted work available to the public, without actually distributing copies of it, does not infringe the public distribution right.

In SA Music, LLC v. Amazon.com, Inc., the copyright owners of numerous musical works composed between the 1920s and the 1960s (including jazz standards such as Stormy Weather and the music of the 1939 motion picture The Wizard of Oz) charged that defendants made unauthorized copies of the songs, compiled them into albums accompanied by copies of the original album artwork, and made the recordings available to purchase via digital download on Amazon’s online music store. Amazon, which was named as a defendant, moved to dismiss the claim that it infringed the plaintiffs’ public distribution right merely by making the unauthorized copies available for purchase. The court agreed with Amazon, holding “that distribution of a copyrighted work under § 106(3) requires ‘actual dissemination’ of the copyrighted work and, in the context of a digital music store, actual dissemination means the transfer (or download) of a file containing the copyrighted work from one computer to another.”

III. Trademark Law Developments

A. Registering “generic.com” as a Trademark

Booking.com is a website that enables its users to make vacation rentals and other travel arrangements. The operator of the website sought to register “Booking.com” as a trademark for the services it offers. The U.S. Patent and Trademark Office (“USPTO”) refused to issue the registration on the ground that the proposed mark is generic. A generic mark is one that refers to a type of good or service rather than the source of a particular good or service. For example, “beer” is a generic term for a type of beverage, while “Bell’s Two Hearted Ale” is a trademark that distinguishes a particular beer that comes from a particular source. Trademark law provides that a generic mark cannot be registered.

The district court and the Fourth Circuit held that the USPTO’s analysis was erroneous, and, in U.S. Patent & Trademark Office v. Booking.com B.V., the U.S. Supreme Court agreed. The USPTO based its analysis on a nineteenth-century case in which the U.S. Supreme Court held that “a generic corporate designation added to a generic term does not confer trademark eligibility.” Thus, a brewery called “Beer Company” could not successfully argue that, while “Beer” is unregistrable as generic, the addition of the word “Company” made the name distinctive and therefore registrable. The Court held that it is different with domain names: The addition of a generic top-level domain (like “.com”) to a generic name (like “beer”) may result in a distinctive designation that is eligible for registration. Because “only one entity can occupy a particular Internet domain name at a time,” consumers may associate a domain name with a particular source of goods or services.

Whether a domain name of the form “generic.com” actually is distinctive depends on consumer perception. What counts is “[t]he primary significance of the registered mark to the relevant public.” The lower courts determined that consumers do in fact perceive “Booking.com” as indicating a particular online source of travel-related services, rather than as a type of travel-service providers: Nobody would ever say “Travelocity is my favorite Booking.com.”

The Court thus rejected a per se rule that a domain name of the form “generic.com” is ineligible for registration as a trademark. Whether a particular domain name is registrable will depend on how consumers perceive it.

B. Liability of an Online Platform for Facilitating Trademark-Infringing Transactions

Redbubble operates an online marketplace that allows artists to upload images they create and allows consumers to order items (such as clothing, wall art, and jigsaw puzzles) that display those images. Some artists uploaded designs that included trademarks owned by Ohio State University (“OSU”), and consumers purchased items bearing those designs. OSU sued Redbubble for trademark infringement. The district court granted summary judgment to Redbubble, finding that its role in the transactions was too limited to render it directly liable for the resulting trademark infringement.

In Ohio State University v. Redbubble, Inc., the Sixth Circuit held that the district court should not have granted summary judgment. The appellate court began its analysis by noting that, in previous cases, courts have determined that some online marketplaces, including Amazon (to the extent that it hosts third-party sellers) and eBay, are not directly liable for trademark infringement occurring by virtue of sales that they facilitate. Such platforms employ a hands-off business model. Their role is to connect willing buyers with willing sellers, earning commissions from the transactions. On the other hand, the manufacturers of trademark-infringing items, and the retail stores (online or brick-and-mortar) that sell them, are generally held liable. These entities may therefore be placed on a “spectrum,” with Amazon and eBay at one end and retailers and manufacturers at the other end. The question that the court addressed was where Redbubble falls on this spectrum. More generally, “what level of involvement and control must a defendant exercise over the creation, manufacture, or sale of offending goods to be considered akin to a ‘seller’ or ‘manufacturer’” and therefore potentially directly liable for the resulting infringement?

Redbubble has a greater involvement in the transactions it facilitates than do Amazon and eBay. When a consumer places an order, “Redbubble automatically contacts the artist and arranges the manufacturing and shipping of the product with independent third parties.” Is this degree of involvement enough to render Redbubble liable? The appellate court determined that “one key distinction” relevant to this question “is the degree to which the party represents itself, rather than a third-party vendor, as the seller, or somehow identifies the goods as its own.” The record included some findings of relevance to this criterion. For example, the goods that consumers order “often arrive in Redbubble packaging and contain Redbubble tags.” These facts created a triable issue on the degree of Redbubble’s involvement, making the district court’s grant of summary judgment erroneous.