This article was originally published in Massachusetts Lawyers Weekly in April 1995
On March 9, 1995, the U.S. Court of Appeals for the First Circuit issued a radically conservative copyright decision in the case of Lotus Development Corporation v. Borland International, Inc. To the surprise of many legal observers, and to the almost certain shock of Lotus Corporation, the Court struck out in a new direction for software copyright jurisprudence by becoming the first Circuit Court to hold that a system of menus through which users interface with a software program are a “method of operation,” and therefore are entitled to no protection under U.S. copyright law.
The microcomputer software industry is the jewel in the crown of late 20th century U.S. capitalism. The United States has been the unquestioned leader in this industry since its inception in the 1970s. The proliferation and growth of software companies of all possible sizes has been nothing short of astounding. Vast fortunes have been created, hundreds of thousands of jobs have been created and eliminated, and society may have been changed forever by an industry that didn’t exist 20 years ago. Even more amazing is the prediction by many computer scientists that we are only at the beginning of a technological growth curve that in 10 or 20 years will make software technology of the 1990’s seem as antiquated as turn of the century automobile technology seems today.
It is probably no secret to readers of this article to hear that the phenomenal growth of the computer industry growth has been accompanied by turmoil over the nature and extent of legal protection that should be granted to computer software. Broadly speaking, there are two schools of thought: one group argues forcefully that in the absence of strong legal protection for computer programs the industry will stagnate and decline. The other group argues with equal conviction that minimal legal protection will encourage rapid technical advancement and discourage the growth of monopolists. Which of these two positions (if either) is correct has never been (and may never be) empirically determined. However, the deep philosophical and economic differences between these two groups is being determined not in the schools of economics or industrial science, but rather, on a case-by-case basis as the federal courts struggle to apply federal copyright law to infringement disputes involving computer software.
Development of Software Copyright Law
In 1980 Congress enacted an amendment to the copyright law which explicitly brought computer programs within the protection of the copyright statute. A number of cases in the early and mid-1980s quickly established the non-controversial legal proposition that copyright law protects computer source code (what programmers write) and object code (the “0s” and “1s” that the computer actually processes) from “literal” copying. However, after this “first wave” of decisions a far more problematic set of issues emerged: whether, and to what extent, copyright law protects the nonliteral elements of computer programs. The second wave of cases has falls into two sometimes overlapping categories: cases involving nonliteral copying of source code, and cases involving copying of the user interface of software programs. The popular and computer industry press has often used the phrase “look and feel” as a shorthand label to describe cases involving copyright protection of user interfaces.
Although cases involving copying of the appearance of video games were decided in the late 1970s and early 1980s (often in favor of the copyright owner), it was not until the late 1980s that business software cases involving the protection of user interfaces began in earnest. The early cases, decided almost exclusively by the District Courts, granted a high degree of protection to software authors. For example, Digital Communications Assocs. v. SoftKlone Distributing Corp. , 659 F. Supp. 449 (N.D. Ga. 1987), held that a single screen containing command terms was protectible, and that the defendant’s verbatim copying was infringement. In Manufacturers Technologies, Inc. v. CAMS Inc., 706 F. Supp. 984 (D. Conn. 1989), the court held that menus consisting of textual material were protectible, and went beyond the SoftKlone case to find the defendant liable, even though the defendant had not copied the plaintiff’s screens verbatim. Other courts have been reluctant to protect user interfaces and have developed tests that result in a relatively low level of protection. In particular, the Second, Ninth and Tenth Circuits have developed tests that will filter out elements that are limited by considerations of efficiency or dictated by various “external factors” recognized by those courts. The clear trend appears to be to label more and more elements as unprotectible based on an increasing number of reasons. However, until the First Circuit’s decision in Lotus v. Borland no court has held that screen commands are outside the scope of copyrightable material and therefore entitled to no protection at all.
Background of Lotus v. Borland
The impact and importance of the First Circuit’s decision in Lotus/Borland can only be fully appreciated with an understanding of the background of the case. The nature of Lotus’s claim is easy to understand. Lotus claimed that Borland had infringed Lotus’s copyright when Borland copied Lotus’s menu command hierarchy — the system of menus that are used to operate the Lotus 1-2-3 “electronic spreadsheet” program. An electronic spreadsheet is a computer program that displays related numbers in rows and columns, similar to a traditional paper spreadsheet. What makes it so valuable, and what made electronic spreadsheet programs one of the driving forces of the growth of the microcomputer industry in the 1980s, is the program’s ability to recalculate the spreadsheet if any number or any formula is changed. For example, in a highly complex pro forma income statement the user can change one basic assumption, and the program will automatically (and to the perception of the user instantaneously) recalculate all parts of the spreadsheet that are dependent on that assumption. The software also makes it possible to perform a wide variety of formatting operations, and to copy, move and revise data in a variety of ways.
Like every computer program, Lotus 1-2-3 required a user interface. While it may sound technical, the term “user interface” describes a concept that we all encounter continuously. The user interface between you and your alarm clock is the digital or analogue display and the knobs and switches you use to set the time and stop the alarm. The user interface between you and your car is the steering wheel, the knobs and gauges on the dashboard and the pedals you operate with your feet. If you use an automatic teller machine you interact with the machine through its keyboard and a series of menus that comprise the machine’s user interface.
In a paper-based spreadsheet the “user interface” is paper, pencil and eraser. However, to use an electronic spreadsheet on a computer requires that you use a non-human user interface at two levels. At the most general level the user interface between you and your microcomputer is the monitor screen, the keyboard and, perhaps, the “mouse” that you use to direct actions on the monitor screen. However, every program requires an additional user interface specific to its uses. The user interface of Lotus 1-2-3 (and indeed, of virtually every computer spreadsheet), is a two dimensional grid of cells, each of which can be uniquely identified based on its column and row location. Typically, across the top of the screen each column is labeled with a letter, and down the left side of the screen, each row is labeled with a number. Thus, for example, the upper left-most “cell” in the spreadsheet has the address of “1A.” In the 1980s, before graphical user interfaces became popular, the 1-2-3 user interface also had a “two-line moving cursor menu” at the top of the screen. The top line contained a series of command terms, which could be chosen, or “highlighted,” by the user. The second line of the menu usually consisted of a description of the command that happened to be highlighted. If one “selected” a command (pressed the Enter key while a command was highlighted or pressed the first letter of the command), either a spreadsheet function occurred (for example, a cell was copied) or a new menu of commands was presented. Given the vast number of operations that Lotus 1-2-3 made available to its users, this system of menus consisted of a total of roughly 450 commands.
Borland — a competitor of Lotus — had at least two very good reasons to want to copy the Lotus 1-2-3 menu commands. First, users are understandably slow to gain proficiency in a complex menu system of this sort. Once proficiency (or even familiarity) is achieved, a user is not anxious to have to learn a new menu system, even if the new program may be superior to the old program. This principle of user inertia is illustrated by the ubiquitous QWERTY keyboard (named after the first six letters on the top row of letter keys on the standard keyboard), which was actually designed to slow down typists in the long-ago era when humans could type faster than typewriters would allow, and has survived many challenges from far more efficient keyboard arrangements. Inertia can sometimes be an immovable force, and Borland was understandably reluctant to present Lotus customers — who comprised most of the market — with a new menu system that would have to be learned from scratch.
Borland’s second reason for wanting to use the Lotus menu commands was derived from the fact that moving through a series of five or ten menus time and again to get to the desired command can become tedious. Lotus 1-2-3 allowed users to create “macros”. A macro is simply a mini-program created by the user that creates a single shortcut command to replace what would otherwise be a lengthy series of commands. The concept of macros is present in almost all computer programs, since it permits experienced users, who know what functions they use frequently, to automate those functions. By copying the Lotus 1-2-3 command system Borland made it possible for users of the Lotus program to move to the Borland product and take their valuable macros with them, confident that the macro would work in identical fashion on the Borland program.
Lotus’s infringement case against Borland was simplified by the fact that Borland admitted copying the Lotus menu system. Given this admission, there was little doubt concerning the outcome of Lotus’s challenge in the District Court. Lotus’s suit was assigned to Judge Robert Keeton, who had already spent several years considering the very same legal issues that would arise in the Borland case in Lotus’s copyright infringement suit against two smaller companies, Paperback Software Int’l and Mosaic Software. Both of these companies had also copied the 1-2-3 command structure, and argued in defense to an infringement suit by Lotus that the Lotus menus were not protected by copyright law. Judge Keeton rejected these arguments in his decision in Lotus Development Corp. v. Paperback Software Int’l , 740 F. Supp. 37 (D. Mass. 1990). Moreover, he posited a complex and carefully thought out methodology for evaluating claims of this nature.
Paperback and Mosaic may have been correct (although there is no guarantee that the First Circuit would have decided this issue the same way in 1989 or 1990), but lacking resources to appeal their cases both companies were forced to withdraw their products. Mosaic was forced out of business. Given this background, it was clear from the start that Borland’s only hope lay beyond Judge Keeton, in an eventual appeal. However, the road to the Court of Appeals was long and bitter. Lotus filed suit against Borland in the Summer of 1990. Over the next three years Judge Keeton published four lengthy opinions — all adverse to Borland — dealing with whether Lotus’s menu structure was protected by copyright law and addressing various affirmative defenses raised by Borland. Finally, in the Fall of 1993 the Judge permitted Borland to appeal his rulings on liability while the parties prepared for a damages trial in which Lotus reportedly sought $100 million in damages from Borland. The damages trial was scheduled to begin on March 14, 1995. The Court of Appeals reversal, which was without a remand for further proceedings, was issued on March 9, 1995, only 5 days before the damages trial was to begin.
Lotus v. Borland
The First Circuit’s opinion in Lotus v. Borland was the first software copyright case to be decided by this Court, and it revealed a heretofore unanticipated radically conservative, economically-driven view of software copyright law. Although Borland had asserted many affirmative defenses, Borland’s core defense — and indeed, the central defenses of long-forgotten Paperback and Mosaic — was that the menus were a “system” or “method of operation.” Federal copyright law expressly states that systems and methods of operation (as well as ideas, processes, concepts, principles and discoveries) are not protected by copyright law. 17 U.S.C. 102(b). Judge Keeton had recognized that the 1-2-3 commands enabled users to operate a computer spreadsheet, but he held that there were also “expressive” elements in the choice and arrangement of the terms, and that therefore the command structure was protected under the copyright laws.
The opinion from the First Circuit Court of Appeals was extraordinarily straightforward, even simplistic, in its approach. The Court noted that section 102(b) of the Copyright Act excludes from protection a “method of operation”. This, the Court observed, is the “means by which a person operates something, whether it be a car, a food processor or a computer.” Without expressly disagreeing with Judge Keeton’s finding that Lotus had made “some expressive choices” in selecting and arranging the menu terms, the Court held that the expression in the menus was not copyrightable because these commands provided the “means” by which users controlled and operated the program.
In essence, the First Circuit agreed with the argument that had been made by Paperback and Mosaic as far back as 1987, and had been pressed unsuccessfully by Borland since 1990. Likening the 1-2-3 menu commands to the buttons on a video cassette recorder, the Court held that the menu commands were essentially a “method” for operating the computer, and therefore were uncopyrightable.
The Court also gave great weight to economic efficiency arguments that had been expressly rejected by Judge Keeton, such as compatibility and the transferability of macros from one program to another. The Court noted that under Lotus’s theory of copyright law a computer user would have to learn how to perform the same operation a different way for each program used. This consequence of Lotus’s argument, the Court observed, was “absurd.” Moreover, the Court correctly noted that under Lotus’s approach a user would be required to rewrite macros to operate on different computer programs.
Judge Boudin, in a concurring opinion, went even further, and wondered why customers of Lotus who had devised macros using 1-2-3 should remain “captives of Lotus because of an investment in learning made by the users and not by Lotus.” He noted that while a new menu may involve creative work, over time the value and importance of the menu resides in the investment of users using the menu and building macros to utilize it. Apparently swayed by the policy issues of the case Judge Boudin noted that the question was “not whether Borland should prevail, but on what basis.” After discussing the possible application of the copyright fair use doctrine and what he described as a possible “privileged use” doctrine, Judge Boudin concluded that the majority’s formulation was “as good, if not better, than any other that occurs to me now as within the reach of the courts.”
The First Circuit’s decision in the Lotus/Borland case appears driven by the Court’s perception of economics in the microcomputer industry. Depending on which of the two schools of thought described earlier in this article one endorses, the First Circuit’s decision is either a catastrophe for copyright law and the policies that encourage the risks associated with investments in the software industry, or it is a long-awaited Circuit Court acknowledgment that copyright law is an altogether inappropriate means to protect a utilitarian object like software.
However, the economic effects of these competing policy views may be impossible to measure. At the most basic level both sides must concede that the U.S. software industry has shown strong and consistent growth, seemingly without discernable regard for judicial trends in copyright protection of user interfaces. It seems unlikely that advances in technology and the proliferation of new products, or in society’s adoption of that technology or those products, has been retarded or encouraged depending on whether the courts are inclined to grant weak protection or strong protection. To date, at least, society’s demand for new technology and the creative ferment of the software industry seem to have generated as much growth as society can absorb independent of copyright incentives. We are left to speculate on whether the First Circuit’s decision will encourage or obstruct investment, risk-taking and creative growth.
However, from a more pragmatic legal (as opposed to economic) perspective the First Circuit’s decision may do neither. The Borland decision goes considerably farther than any earlier software copyright decision in any other Federal Circuit. Cases in other circuits, while sometimes sharply circumscribing what may be protected, have invariably operated on the underlying assumption that user interfaces are subject to copyright protection. The arguments made by the Borland Court, if taken to their logical extremes, would lead to the conclusion that copyright protection does not subsist in any computer user interface, regardless of its level of originality. For example, the Court did not challenge Judge Keeton’s conclusion that there was original expression in the Lotus menus; instead, it held that because these menus were the “means” of operating the program, they were uncopyrightable regardless of this expression. Presumably, this analysis would be equally applicable to the graphically-oriented user interfaces that increased computer power and new software technologies have made ubiquitous in the 1990s. While graphical user interfaces may have “more” expressive content than the Lotus 1-2-3 menus, this is only a matter of degree; the First Circuit’s analysis suggests that “more” expressive content would change its analysis and conclusions. No other court has even hinted at a result this extreme.
Therefore, the real question may be whether the First Circuit has now become the leader in the area of software user interfaces, with other courts around the country soon to follow, or whether the First Circuit has wandered so far afield from the majority trend that it will find itself isolated and the First Circuit will find itself a backwater of software copyright law and a strategic factor in forum shopping decisions in software copyright cases.