brief Amicus curiae
Applying Fundamental Copyright Principles to Lotus Development Corp. v. Borland International, Inc. *
Dennis S. Karjala †
Peter S. Menell ††
I. INTRODUCTION 177
II. ISSUE IN THE CASE 179
III. SUMMARY OF ARGUMENT 179
IV. ARGUMENT 180
B. The Critical Relationship between Copyright and Patent Law 182
C. Application of these Principles to the Lotus Case 187
Lotus v. Borland raises the fundamental question of whether, and to what extent, the copyright in a computer program protects the user interface generated by that program. Lotus owns the copyright in and markets the very popular spreadsheet program known as Lotus 1-2-3. Borland produced and marketed a competing set of spreadsheet programs known as Quattro. Borland designed and implemented in program code its own user interface, as to which no claim of infringement was made. In addition, however, the Borland programs included a Lotus "emulation mode," which allowed users familiar with the Lotus program more easily to make the switch to Quattro. That is, when operating Quattro in the Lotus emulation mode, the Borland program presented to the user the same hierarchy of menu commands as Lotus 1-2-3 and responded identically to those commands. Lotus brought action for copyright infringement based on the similarity of the Quattro user interface when operated in the emulation mode. No copying of program code was alleged.
The federal district court concluded that the Lotus 1-2-3 user interface as a whole, as well as specific elements of the interface, were copyright protected as nonliteral aspects of the underlying Lotus 1-2-3 computer program.1 The court based its finding of infringement on the copying of "menu commands, menu command structure, macro language, and keystroke sequences."2 The district court's reasoning was based largely on the merger doctrine: Because a wide variety of formats and methods of operation are possible for spreadsheet programs, any given set of choices could be copyright protected without hindering the development of other spreadsheet programs. The court essentially ignored functionality as a limit on copyright protection.
The authors of this brief are law professors at Arizona State University and the University of California (Berkeley), respectively, who teach and write on copyright issues, and specifically on the proper application of copyright law to the protection of computer software.3 We submitted our views on what we perceive to be the crucial issue in this case to the court of appeals because we believe that the district court fundamentally misconstrued both the traditional copyright principles applicable to the facts and the role played by copyright in the overall system of intellectual property protection in the United States. As a result, the district court decision effectively overrules the Supreme Court's decision in Baker v. Selden,4 a decision that for over 100 years has served to demarcate the boundary between the patent and copyright regimes. The result of the district court's decision is to extend copyright protection far beyond its traditional bounds. We believe that Congress is the only body authorized to make such an expansion of copyright protection, and therefore urged the court of appeals to reverse the district court's decision.
II. ISSUE IN THE CASE
III. SUMMARY OF ARGUMENT
Second, even if aspects of the Lotus 1-2-3 user interface qualify as works of authorship independently of the underlying computer program-for example, as graphic or audiovisual works-the traditional limits on the scope of any copyright in that interface still apply. These limits include not only the copyright doctrines of merger and scènes á faire but also the foundational doctrines of Baker v. Selden and section 102(b) of the Copyright Act. These doctrines exclude function from the scope of copyright protection. The policies underlying that tradition are and always have been firmly grounded in the fundamental allocation of protection established by the intellectual property system as a whole. Each of the intellectual property regimes-copyright, patent, trademark, and trade secret-must be understood in the context of this larger system and cannot be interpreted in isolation. Moreover, only Congress, and not the courts, has authority to expand the scope of the intellectual property laws. Because we are addressing an asserted independent copyright in the outputs generated by the Lotus 1-2-3 program and not the program code itself, protection of functional features of the user interface are not governed by Congress' decision to protect some aspects of program code under copyright. This case centers on the interpretation of traditional copyright principles applied to visual images and systems for processing data, and absent instructions from Congress to the contrary, courts must adhere to the doctrines reflected in the Copyright Act and the case law interpreting the intellectual property statutes.
A. Copyright in a Program Does Not Extend to those "Certain Results" Brought about by Program Execution5
This straightforward conclusion-that what a program does, and therefore its "appearance" to the outside world when viewed solely through input and output devices, is not protected by the copyright in the program-comports with well over 100 years of copyright tradition and was explicitly recognized by the Second Circuit in Computer Associates International, Inc. v. Altai:10
B. The Critical Relationship between Copyright and Patent Law
Despite the similar basic approach of patent and copyright law-both draw a balance between providing an incentive for the creation of works and ensuring that the public and later creators can enjoy and build upon an expanding "public domain"-they are distinct components of a coherent scheme of intellectual property protection. Patent law seeks to promote the advancement of technology while copyright law seeks to encourage culture and the arts. Given their different foci, it is not surprising that Congress has crafted very different balances in these two statutes. In exchange for protection of claimed functional attributes of an invention, the patent law requires an inventor to demonstrate to a technically trained examiner that the asserted invention is novel, nonobvious, and useful. After meeting these exacting requirements, the holder of a United States patent is entitled to 17 years of exclusive rights to practice the invention. By contrast, the copyright law presents a much lower threshold for receiving protection-a minimal level of "creativity"-and forgoes any requirement of examination. In return, the author is entitled to protection limited to the expressive content of the work. The term of this protection, however, extends for the life of the author plus 50 years (or 75 years for entity authors).
In view of the very different thresholds for as well as scope and terms of protection, to extend copyright by protecting functional works14 without explicit congressional authorization presents grave risks to the integrity of the entire intellectual property law system. Thus, in the seminal case of Baker v. Selden,15 decided more than a century ago, the United States Supreme Court explicitly recognized the fundamentally different roles played by copyright and patent law within that system:
The policy basis for the distinct approaches of patent and copyright law is the social desirability of allowing later technological creators-creators of functional works-to build on and improve, often in small ways, the earlier works of others.19 Technological improvements are often substantially similar to the products they improve and would infringe if the copyright standard were applied. Such improvements do not infringe a patent in the technology unless they adopt all of the elements, or their substantial equivalents, of a patent claim. Authors of copyright-protected works, as well, build upon public domain works and unprotected elements of protected works, but in taking from the latter, copyright law limits them to the more general or abstract features, together with those aspects denied copyright protection under Baker v. Selden, section 102(b) of the United States Copyright Act, and other limiting doctrines.
Later inventors thus can apply a "ground up" approach to reliance on protected works that precede. Once they are outside all patent claims, they are safe. Later authors, however, must adopt a "top down" approach and take a serious risk of being held liable for infringement as their reliance becomes more detailed. The different foci of patent and copyright law explain their different approaches to infringement. The social utility of allowing subsequent authors to make minor variations on a copyright-protected novel is minimal. For works of fiction, art, and music, variety is the spice of both legal and real life. We prefer to have one hundred different war novels than one hundred versions of "War and Peace" that differ only in their final chapter. Consequently, the broad scope of copyright protection for novels and the long period of its duration fulfills the goal of recognizing the author's creativity without unduly hindering later authors or depriving society of desirable works.20
Technology, however, improves incrementally, as later inventors add a bell or a whistle to an earlier invention to make it more desirable or useful to consumers. Many improvements on existing products, being rather straightforward or "obvious" in the sense of patent law, are given no intellectual property protection once they are released to the public. Such products often show at least as much intellectual creativity as many copyright-protected works, but their creators have a monopoly only for the period that is required for competitors to recognize the value or popularity of the improved product, figure out its "secret," if any, and gear up for production and marketing. In the case of technological products we have drawn the social policy balance at a different point than for traditional works of authorship because we believe that to grant intellectual property rights in ordinary engineering advances would hinder the development of more and better products than it would encourage.21 Hence the "nonobviousness" requirement of patent law, as well as its shorter term and disclosure requirements, recognize and promote the incremental nature of technological innovation.
Computer programs, of course, at least in object-code form, are technological in nature: They sequentially set the switches inside a computing machine in such a way that the results can be interpreted as "information processing" by human beings. To the extent that code is now copyright protected, Congress has indicated that a functional work should come under the copyright umbrella. However, the National Commission on New Technological Uses of Copyright Works (CONTU)-established by Congress "to assist the President and Congress in developing a national policy for protecting both the rights of copyright owners and insuring public access to copyrighted works when they are used in computer . . . systems, bearing in mind the public and consumer interest,"22-concluded that literal copying of computer programs would dramatically reduce the development costs of second comers and thereby create a disincentive to invest in computer program development.23 Even if we assume that the application of Baker v. Selden and section 102(b) to program code is now more limited than to other technological works-in that the literal code of every program can be said to implement a process or method of operation and yet remains copyright protected-there is no basis for concluding that Congress intended to remove any barriers to the copyright protection of function in other classes of works. The Supreme Court in Feist,24 the Second and Ninth Circuits in Computer Associates25 and Sega,26 and the Northern District of California in Apple Computer, Inc. v. Microsoft Corp.27 have recently reaffirmed the fundamental role of Baker v. Selden in copyright analysis.
C. Application of these Principles to the Lotus Case
In stark contrast to the Supreme Court's analysis in Baker v. Selden, the district court in Lotus v. Borland begins its infringement analysis with an explicit refusal to treat as unprotected idea the very systemic aspects of the Lotus 1-2-3 interface that are denied protection under Baker. The court sets up an abstractions scale going from an electronic spreadsheet at the high end down to the precise Lotus 1-2-3 command set as hierarchically arranged by Lotus at the low end.33 It rejects Borland's argument that the latter is also unprotected for the following reasons:
The second step of the district court's infringement standard compounds this error by simply assessing whether there were many ways of representing spreadsheet interfaces and designing menu trees for inputting and processing information:
Had it been dealing with a fanciful work, the district court's analysis would have been appropriate, but Baker v. Selden and section 102(b) of the Copyright Act mean more than the merger doctrine for functional features. The whole point of Baker's channeling principle between patent and copyright law is that systems and methodologies are not copyright protected. In this case, users of the Lotus 1-2-3 have learned the Lotus system and many want to use it, not some other system. Other users have developed "macros" that must combine Lotus commands precisely as Lotus defines them in order to run on the Lotus system. In order to use the Lotus system, there must be identity of commands and command structures. Of course, it is possible for Borland and other spreadsheet developers to create other systems. Borland, in fact, did that with its native mode interface. But this possibility is irrelevant to the application of Baker v. Selden to the facts of Lotus: Under Baker, the Lotus system is simply unprotected by copyright.38
Applying the district court's analysis to the facts of Baker v. Selden would therefore require a different outcome in that case. There are countless ways of bookkeeping and methods for inputting and processing accounting information. Hence, Selden would be entitled to protection for his system as well as his forms for recording the information in the manner demanded by his system. Had Selden implemented his accounting system by way of a computer program, as Lotus did, rather than simply describe it in a book, he would under this analysis be entitled to protect the "ruled lines and headings of accounts" and structured methodology of his system. Thus, the district court in this case has effectively overruled Baker v. Selden, a central pillar in the United States intellectual property protection system.
This is not to say that there can never be copyright-protected "expression" in a computer program user interface. A straightforward example would be the fanciful characters and the environments in which they chase or avoid enemies in video games.39 It is even possible that some aspects of the Lotus 1-2-3 user interface, at least its screen displays, are included solely for aesthetic purposes and are not part of any of the processes, operations, or methods by which the user actually employs the program. Textual information under the "Help" operation will in many cases be copyright protected like any other literary work. But that should be the test: Is an aspect of the interface chosen solely to portray an appearance or to convey information or does it have another utilitarian function-is it something that the user must do in order to make this program operate in the intended manner?40 If it is, section 102(b) and Baker v. Selden require that it be excluded from copyright protection.
Lotus may feel that it has made a fundamental advance in making personal computers useful to nonspecialists. If, in fact, Lotus has made a nonobvious advance in the art over the products of predecessors on whom Lotus relied, it should have sought patent protection for those advances.41 Lotus has a copyright in its program, which is undoubtedly a complex piece of work, that affords Lotus important protection against piracy of program code. Lotus does not allege, however, that Borland copied its program code or protectable nonliteral elements (if any) of that program code. Although it is not strictly relevant to the copyright analysis, Borland did not even gain the advantage over Lotus of avoiding the design costs of a graphical user interface for a spreadsheet program because Borland designed both what it felt would be a better interface and wrote code that would implement the Lotus interface to accommodate users who, like the beleaguered users of the QWERTY typewriter keyboard, had learned the Lotus system and were reluctant to change. It is precisely that kind of productive use of unpatented functional works that the exclusion from copyright protection of Baker v. Selden is designed to promote.
We also do not mean to suggest that aspects of user interfaces or other outputs of computer programs might not merit greater protection than currently exists under intellectual property law. Rather, we believe, as did CONTU,42 that the legislature is the appropriate body to expand the scope of intellectual property protection. In the 1980 amendments to the Copyright Act and the Semiconductor Chip Protection Act of 1984, Congress has, after careful study and broad involvement of affected interests and the public at large, extended intellectual property protection to new classes of computer-related works. Over the past four years, the Office of Technology Assessment has undertaken, at the behest of Congress, studies of the efficacy of existing legal protection for computer software.43 We fully expect that Congress will seek to address perceived gaps in this system should existing protection, properly applied by the courts, prove inadequate. It is not, however, the province of courts to take over this lawmaking function by extending protection beyond the limits established by traditional doctrines undergirding the intellectual property system. The district court in Lotus v. Borland clearly exceeded these bounds, expanding protection afforded outputs of computer programs-user interfaces-and effectively overruling Baker v. Selden without the authorization of Congress. Even if additional protection were warranted for computer-user interfaces, it is not at all clear that traditional copyright protection-with its long term of protection and exclusivity of rights-would be appropriate.44 As other courts grappling with software cases have recognized, the judiciary is ill-suited to weighing the complex issues involved in expanding or reforming intellectual property law.45
Computer software, by its very nature as written work intended to serve utilitarian purposes, defies easy categorization within our intellectual property system. Cases involving computer software have therefore presented difficult challenges for courts, which have not always correctly applied traditional copyright doctrines in conjunction with Congress's adoption of CONTU's recommendations to a complex and evolving technology. As a prime example, Whelan Associates, Inc. v. Jaslow Dental Laboratory, Inc.,46 stood for six years as an influential decision on the scope of protection for nonliteral aspects of program code despite cogent academic criticism. The Second Circuit's sweeping repudiation of the Whelan approach in the Computer Associates case in 1992, building upon this academic literature and reasserting essential limiting doctrines of copyright law, has supplanted the Whelan approach in substantially all subsequent decisions.47 Similarly with regard to copyright protection for interfaces, most early district courts addressing the issue-resting upon an erroneous assumption that the copyright in the computer program extends to the outputs of the program and/or building upon the faulty foundation of Whelan-have gone astray of fundamental copyright doctrines.48 As the first appellate court to review this issue, the First Circuit in this case should seize the opportunity to bring copyright protection for outputs of computer programs back into line with the fundamental principles of copyright law.