Decompilation
Excerpt from U.S. Cong./Off. Technol. Assessment, Finding a Balance: Computer Software, Intellectual Property, and the Challenge of Technological Change, ch. 4 (1992)
“Decompilation” is a procedure for translating a machine language program into a more understandable form. It is thought by some to be a copyright infringement, and by others to be a necessary tool for software engineering. The decompilation issue intersects many of the policy issues outlined earlier in this chapter. For example, decompilation may be used in the development of functionally compatible products; whether or not the development of such products should be permitted is itself a policy issue.
Introduction
To understand a program, there are three things you can do: read about it (e.g., documentation), read it (e.g., source code), or run it (e.g. watch execution, get trace data, examine dynamic storage, etc.). Understanding a program is made easier when the high-level language or assembly language representations are available. In most cases, however, only the machine language version is distributed. Decompilation is a procedure by which a high-level language representation of a program is derived from a machine language program, and “disassembly” is a procedure for translating the machine language program into an assembly language program.
The legal status of efforts to discover assembly language or high-level language representations of a program has become the subject of an intense debate. Both decompilation and disassembly involve the making of at least a partial reproduction or derivative of the machine language program, and some people believe that reverse engineering using these techniques is a copyright infringement. The policy question is the extent to which limiting access to information about someone else's program through the workings of the copyright law is socially desirable.
Important factors cited in the policy debate are the uses of decompilation, the ease with which it can be done, and the degree to which the information is available from sources other than decompilation. It has been argued that limits placed on decompilation are required to provide sufficient incentives for the development of original programs. Those who take this position claim that decompilation is a straightforward and routine process that allows clone programs to be implemented at much lower cost.141
| 141. “Decompilation of a computer program does not provide an imitator with just a good start in producing a competing product; it gives him virtually everything necessary to produce a functionally identical product,” William T. Lake, John H. Harwood, and Thomas P. Olson, “Tampering With Fundamentals: A Critique of Proposed Changes in EC Software Protection,” The Computer Lawyer, vol. 6, No. 12, December 1989, pp. 1-10. |
Programs are decompiled and then:
. . . without the necessity for the significant R&D expenditures made by the innovator, the pirate goes on to alter the program to disguise the copying, and create a second, similar program which it markets as an allegedly different product for a much lower price.142
| 142. Testimony of James M. Burger, Chief Counsel, Apple Computer, Inc., on behalf of the Computer and Bus. Equip. Mfrs. Assn, at hearings before the House Subcom. on Intellectual Property and Judicial Administration, May 30, 1991. |
Others argue that decompilation is technically difficult, and is therefore unlikely to be used for piracy. They emphasize the disassembly and decompilation can be used for a variety of other purposes, many of which would have a less direct economic impact on the developer of the program being reverse engineered. For example, some of the information gained by decompilation may be used in developing an “attaching” product that is to exchange data with the program being reverse engineered. Decompilation also could be used for maintenance, debugging, detecting viruses, investigating safety or reliability concerns, or systems integration. Indeed, some of these uses of decompilation represent situations in which an organization might reverse engineer its own programs, not just those developed by someone else.
Decompilation and Disassembly
The product of decompilation or disassembly would never be identical to the original source program. At the very least, comments and the names of labels, variables, and procedures would be lost in the assembly or compilation process and could not be recovered. In addition, the structure of the decompiled program would not necessarily be the same as that of the original program, although this would depend on the compiler that had been used. Because of the loss of mnemonics and much of the structure of the program, considerable work is required to understand the decompiled or disassembled program.
Disassembly is easier than decompilation. There is essentially a one-to-one conversion between the machine language statements and assembly language statements, simplifying the process of translating the machine language program into a more readable form. However, it takes a great deal of effort to understand the disassembled code from a large program.145 Because disassemblers are widely available, some developers assume that their programs will be disassembled, and try to write sensitive parts of their code in ways which make disassembly more difficult or make the disassembled code more difficult to understand. Figure 4-2 shows a high-level language program, the corresponding machine language (compiled) program, and the results of disassembling the machine language program.
| 145. An 80,000-byte machine language program for an IBM PC-class computer would result in 32,000 lines of assembly code. Clark Calkins, “Tailoring the MD86 Disassembler for Turbo Pascal,” Tech Specialist, vol. 2, No. 6, June 1991, pp. 41-46. |
Decompilation is much more difficult; at this time there appear to be no commercially available decompilers. For this reason, it is unclear whether decompilation is widely used by “pirates” to decompile entire programs and then rearrange the code in an attempt to hide the copying.148 It is possible that the term “decompilation” is being used in the policy debate to include disassembly, as “decompilation” is often characterized as any technique that is used to transform “machine readable” code into “human readable” code.
| 148. Following the testimony of the Computer and Business Equipment Manufacturers Association at the May 30, 1991 hearings (see footnote 142), OTA asked for specific examples of piracy using decompilation and descriptions of the state of the art in automated decompilation (Joan Winston, OTA, letters to James Burger, Apple Computer, July 5, 1991 and Sept. 23, 1991). To date, OTA has not been provided with this information. |
Today, any effort to decompile a program would start with disassembly. Then, if one knew something about the compiler that had been used, it might be possible to match certain patterns of assembly language statements to higher level constructs. However, decompilation would be much more difficult in cases where a sophisticated compiler had been used; optimizing compilers delete and rearrange some of the instructions in order to make the machine language program more efficient, and the correspondence between sequences of machine language instructions and high-level instructions becomes less direct. While a pseudo-source code program could still be derived, it would be less likely that the decompiled program would immediately reveal the original program's structure.
Uses of Decompilation
The information gained by reverse engineering techniques such as decompilation can be put to a variety of uses, each with a different economic effect. The effect on the developer of the program being decompiled is most direct when the information is being used to develop a competitive product. In some cases the reverse engineer may be interested in learning about a small part of a program, such as an algorithm, that gives the program's owner a competitive advantage. In other cases, decompilation could be used to develop the specifications for a program that is functionally compatible — a cloned program. Sometimes the specifications are used in a clean-room process that is intended to ensure that the new program does not share expression with the original, for it is the protected expression that is protected by copyright.
Decompilation may also be used to develop program or hardware device that is not competitive but “complementary” or “attaching.” This would not affect the market for the original product directly, but would create more competition in the second market. For example, knowing interface information might allow the development of competition in the market for peripheral devices such as printers. Decompilation can also be used to confirm published interface specifications; for example, in the course of debugging a program an unexpected problem may arise with another program in the system, such as an operating system.
Finally, there are a variety of uses for decompilation for which no product is developed at all. First of all, someone may wish to “maintain” a program for which no source code is available. Decompilation or disassembly would help in understanding the program so that it could be adapted to new requirements, or to fix bugs if no other support was available. Disassembly is also used to find viruses, to examine the output of a compiler to see what it had done, and finally to examine a competitor's program to see if they had taken any protected expression.
Other Methods of Reverse Engineering
There are other methods of reverse engineering programs whose legal status is less controversial because they do not involve the making of unauthorized reproductions of the machine language program. In practice, a reverse engineer would probably employ a combination of methods, depending on the application, the information being sought, the effort involved, and legal considerations. Some information can be obtained by simply executing the program: it can be run with many different data sets and its behavior observed. There are a number of different software and hardware tools that could be used to follow the course of execution of the program. However, the program code is the best specification of the behavior of the program — it may be impossible to develop tests to explore all the cases that a program may have to handle.
Other information is available from published specifications, manuals, and standards documents. In some cases companies will publish interface specifications because it is in their commercial interest to do so. Even if the information is not published, they may be willing to make it available through contractual arrangements. However, in other cases, such as when a company is active in the market for both the primary product and a complementary product, it may want to limit competition in the secondary market by not making the interface information available. Published documents may not be at the appropriate level of detail. For example, there may be scope for differences between implementations of a standard, and manuals may be inaccurate or out of date, or leave some elements undocumented.
Sega Enterprises Ltd. v. Accolade, Inc.
United States Court of Appeals
977 F.2d 1510 (9th Cir. 1992), cert. denied, 113 S. Ct. 1582 (1993)
Reinhardt, Circuit Judge.
This case presents several difficult questions of first impression involving our copyright laws. We are asked to determine whether the Copyright Act permits persons who are neither copyright holders nor licensees to disassemble a copyrighted computer program in order to gain an understanding of the unprotected functional elements of the program.
In light of the public policies underlying the Act, we conclude that, when the person seeking the understanding has a legitimate reason for doing so and when no other means of access to the unprotected elements exists, such disassembly is as a matter of law a fair use of the copyrighted work.
Accordingly, we reverse the district court's grant of a preliminary injunction in favor of plaintiff-appellee Sega Enterprises, Ltd. on its claim of copyright infringement.
Background
Plaintiff-appellee Sega Enterprises, Ltd. (“Sega”) develops and markets video entertainment systems, including the “Genesis” console and video game cartridges. Defendant-appellant, Accolade, Inc., is an independent developer, manufacturer, and marketer of computer entertainment software, including game cartridges that are compatible with the Genesis console, as well as game cartridges that are compatible with other computer systems.
Sega licenses its copyrighted computer code to a number of independent developers of computer game software. Those licensees develop and sell Genesis-compatible video games in competition with Sega. Accolade is not and never has been a licensee of Sega. Prior to rendering its own games compatible with the Genesis console, Accolade explored the possibility of entering into a licensing agreement with Sega, but abandoned the effort because the agreement would have required that Sega be the exclusive manufacturer of all games produced by Accolade.
Accolade used a two-step process to render its video games compatible with the Genesis console. First, it “reverse engineered” Sega's video game programs in order to discover the requirements for compatibility with the Genesis console. As part of the reverse engineering process, Accolade transformed the machine-readable object code contained in commercially available copies of Sega's game cartridges into human-readable source code using a process called “disassembly” or “decompilation”.2 Accolade purchased a Genesis console and three Sega game cartridges, wired a decompiler into the console circuitry, and generated printouts of the resulting source code. Accolade engineers studied and annotated the printouts in order to identify areas of commonality among the three game programs. They then loaded the disassembled code back into a computer, and experimented to discover the interface specifications for the Genesis console by modifying the programs and studying the results. At the end of the reverse engineering process, Accolade created a development manual that incorporated the information it had discovered about the requirements for a Genesis–compatible game. According to the Accolade employees who created the manual, the manual contained only functional descriptions of the interface requirements and did not include any of Sega's code.
| 2. Computer programs are written in specialized alphanumeric languages, or “source code”. In order to operate a computer, source code must be translated into computer readable form, or “object code”. Object code uses only two symbols, 0 and 1, in combinations which represent the alphanumeric characters of the source code. A program written in source code is translated into object code using a computer program called an “assembler” or “compiler”, and then imprinted onto a silicon chip for commercial distribution. Devices called “disassemblers” or “decompilers” can reverse this process by “reading” the electronic signals for “0” and “1” that are produced while the program is being run, storing the resulting object code in computer memory, and translating the object code into source code. Both assembly and disassembly devices are commercially available, and both types of devices are widely used within the software industry. |
In the second stage, Accolade created its own games for the Genesis. According to Accolade, at this stage it did not copy Sega's programs, but relied only on the information concerning interface specifications for the Genesis that was contained in its development manual. Accolade maintains that with the exception of the interface specifications, none of the code in its own games is derived in any way from its examination of Sega's code. In 1990, Accolade released “Ishido”, a game which it had originally developed and released for use with the Macintosh and IBM personal computer systems, for use with the Genesis console.
Even before Accolade began to reverse engineer Sega's games, Sega had grown concerned about the rise of software and hardware piracy in Taiwan and other Southeast Asian countries to which it exported its products. Taiwan is not a signatory to the Berne Convention and does not recognize foreign copyrights. Taiwan does allow prosecution of trademark counterfeiters. However, the counterfeiters had discovered how to modify Sega's game programs to blank out the screen display of Sega's trademark before repackaging and reselling the games as their own. Accordingly, Sega began to explore methods of protecting its trademark rights in the Genesis and Genesis–compatible games. While the development of its own trademark security system (TMSS) was pending, Sega licensed a patented TMSS for use with the Genesis home entertainment system.
The most recent version of the Genesis console, the “Genesis III”, incorporates the licensed TMSS. When a game cartridge is inserted, the microprocessor contained in the Genesis III searches the game program for four bytes of data consisting of the letters “S-E-G-A” (the “TMSS initialization code”). If the Genesis III finds the TMSS initialization code in the right location, the game is rendered compatible and will operate on the console. In such case, the TMSS initialization code then prompts a visual display for approximately three seconds which reads “produced by or under license from sega enterprises ltd” (the ” Sega Message”). All of Sega's game cartridges, including those disassembled by Accolade, contain the TMSS initialization code.
Accolade learned of the impending release of the Genesis III in the United States in January, 1991, when the Genesis III was displayed at a consumer electronics show. When a demonstration at the consumer electronics show revealed that Accolade's “Ishido” game cartridges would not operate on the Genesis III, Accolade returned to the drawing board. During the reverse engineering process, Accolade engineers had discovered a small segment of code — the TMSS initialization code — that was included in the “power-up” sequence of every Sega game, but that had no identifiable function. The games would operate on the original Genesis console even if the code segment was removed. Mike Lorenzen, the Accolade engineer with primary responsibility for reverse engineering the interface procedures for the Genesis console, sent a memo regarding the code segment to Alan Miller, his supervisor and the current president of Accolade, in which he noted that “it is possible that some future Sega peripheral device might require it for proper initialization.”
In the second round of reverse engineering, Accolade engineers focused on the code segment identified by Lorenzen. After further study, Accolade added the code to its development manual in the form of a standard header file to be used in all games. The file contains approximately twenty to twenty-five bytes of data. Each of Accolade's games contains a total of 500,000 to 1,500,000 bytes. According to Accolade employees, the header file is the only portion of Sega's code that Accolade copied into its own game programs. In this appeal, Sega does not raise a separate claim of copyright infringement with respect to the header file.
In 1991, Accolade released five more games for use with the Genesis III, “Star Control”, “Hardball!”, “Onslaught”, “Turrican”, and “Mike Ditka Power Football.” With the exception of “Mike Ditka Power Football”, all of those games, like “Ishido”, had originally been developed and marketed for use with other hardware systems. All contained the standard header file that included the TMSS initialization code. According to Accolade, it did not learn until after the Genesis III was released on the market in September, 1991, that in addition to enabling its software to operate on the Genesis III, the header file caused the display of the Sega Message. All of the games except “Onslaught” operate on the Genesis III console; apparently, the programmer who translated “Onslaught” for use with the Genesis system did not place the TMSS initialization code at the correct location in the program.
All of Accolade's Genesis-compatible games are packaged in a similar fashion. The front of the box displays Accolade's “Ballistic” trademark and states “for use with Sega Genesis and Mega Drive Systems.” The back of the box contains the following statement: ” Sega and Genesis are registered trademarks of Sega Enterprises, Ltd. Game 1991 Accolade, Inc. All rights reserved. Ballistic is a trademark of Accolade, Inc. Accolade, Inc. is not associated with Sega Enterprises, Ltd. All product and corporate names are trademarks and registered trademarks of their respective owners.”
After expedited discovery and a hearing, the district court granted Sega's motion for a preliminary injunction. With respect to Sega's copyright claim, the district court rejected Accolade's contention that intermediate copying of computer object code does not constitute infringement under the Copyright Act. It found that Accolade had disassembled Sega's code for a commercial purpose, and that Sega had likely lost sales of its games as a result of Accolade's copying. The court further found that there were alternatives to disassembly that Accolade could have used in order to study the functional requirements for Genesis compatibility. Accordingly, it also rejected Accolade's fair use defense to Sega's copyright infringement claim.
Based on its conclusion that Sega is likely to succeed on the merits of its claim for copyright infringement, the district court enjoined Accolade from: (1) disassembling Sega's copyrighted code; (2) using or modifying Sega's copyrighted code; (3) developing, manufacturing, distributing, or selling Genesis-compatible games that were created in whole or in part by means that included disassembly; and (4) manufacturing, distributing, or selling any Genesis-compatible game that prompts the Sega Message. In response to a request from Sega, the district court modified the preliminary injunction order to require the recall of Accolade's infringing games within ten business days.
Accolade raises four arguments in support of its position that disassembly of the object code in a copyrighted computer program does not constitute copyright infringement. First, it maintains that intermediate copying does not infringe the exclusive rights granted to copyright owners in section 106 of the Copyright Act unless the end product of the copying is substantially similar to the copyrighted work. Second, it argues that disassembly of object code in order to gain an understanding of the ideas and functional concepts embodied in the code is lawful under section 102(b) of the Act, which exempts ideas and functional concepts from copyright protection. Third, it suggests that disassembly is authorized by section 117 of the Act, which entitles the lawful owner of a copy of a computer program to load the program into a computer. Finally, Accolade contends that disassembly of object code in order to gain an understanding of the ideas and functional concepts embodied in the code is a fair use that is privileged by section 107 of the Act.
Neither the language of the Act nor the law of this circuit supports Accolade's first three arguments. Accolade's fourth argument, however, has merit. Although the question is fairly debatable, we conclude based on the policies underlying the Copyright Act that disassembly of copyrighted object code is, as a matter of law, a fair use of the copyrighted work if such disassembly provides the only means of access to those elements of the code that are not protected by copyright and the copier has a legitimate reason for seeking such access. Accordingly, we hold that Sega has failed to demonstrate a likelihood of success on the merits of its copyright claim. Because on the record before us the hardships do not tip sharply (or at all) in Sega's favor, the preliminary injunction issued in its favor must be dissolved, at least with respect to that claim.
Intermediate Copying
We have previously held that the Copyright Act does not distinguish between unauthorized copies of a copyrighted work on the basis of what stage of the alleged infringer's work the unauthorized copies represent. Walker v. University Books, 602 F.2d 859, 864 (9th Cir. 1979) (“The fact that an allegedly infringing copy of a protected work may itself be only an inchoate representation of some final product to be marketed commercially does not in itself negate the possibility of infringement.”). Our holding in Walker was based on the plain language of the Act. Section 106 grants to the copyright owner the exclusive rights “to reproduce the work in copies”, “to prepare derivative works based upon the copyrighted work”, and to authorize the preparation of copies and derivative works. 17 U.S.C. § 106 (1)-(2). Section 501 provides that “anyone who violates any of the exclusive rights of the copyright owner as provided by sections 106 through 118...is an infringer of the copyright.” Id. § 501(a). On its face, that language unambiguously encompasses and proscribes “intermediate copying”.
In order to constitute a “copy” for purposes of the Act, the allegedly infringing work must be fixed in some tangible form, “from which the work can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device.” 17 U.S.C. § 101. The computer file generated by the disassembly program, the printouts of the disassembled code, and the computer files containing Accolade's modifications of the code that were generated during the reverse engineering process all satisfy that requirement. The intermediate copying done by Accolade therefore falls squarely within the category of acts that are prohibited by the statute.
Accolade points to a number of cases that it argues establish the lawfulness of intermediate copying. Most of the cases involved the alleged copying of books, scripts, or literary characters. In each case, however, the eventual lawsuit alleged infringement only as to the final work of the defendants. We conclude that this group of cases does not alter or limit the holding of Walker.
The remaining cases cited by Accolade, like the case before us, involved intermediate copying of computer code as an initial step in the development of a competing product. Computer Assoc. Int'l v. Altai; NEC Corp. v. Intel Corp.; E.F. Johnson Co. v. Uniden Corp. In each case, the court based its determination regarding infringement solely on the degree of similarity between the allegedly infringed work and the defendant's final product. A close reading of those cases, however, reveals that in none of them was the legality of the intermediate copying at issue. Sega cites an equal number of cases involving intermediate copying of copyrighted computer code to support its assertion that such copying is prohibited. SAS Institute, Inc. v. S&H Computer Systems, Inc.; Hubco Data Products v. Management Assistance, Inc. Again, however, it appears that the question of the lawfulness of intermediate copying was not raised in any of those cases.
In summary, the question whether intermediate copying of computer object code infringes the exclusive rights granted to the copyright owner in section 106 of the Copyright Act is a question of first impression. In light of the unambiguous language of the Act, we decline to depart from the rule set forth in Walker for copyrighted works generally.
Accordingly, we hold that intermediate copying of computer object code may infringe the exclusive rights granted to the copyright owner in section 106 of the Copyright Act regardless of whether the end product of the copying also infringes those rights. If intermediate copying is permissible under the Act, authority for such copying must be found in one of the statutory provisions to which the rights granted in section 106 are subject.
The Idea/Expression Distinction
Accolade next contends that disassembly of computer object code does not violate the Copyright Act because it is necessary in order to gain access to the ideas and functional concepts embodied in the code, which are not protected by copyright. 17 U.S.C. § 102(b). Because humans cannot comprehend object code, it reasons, disassembly of a commercially available computer program into human-readable form should not be considered an infringement of the owner's copyright. Insofar as Accolade suggests that disassembly of object code is lawful per se, it seeks to overturn settled law.
Accolade's argument regarding access to ideas is, in essence, an argument that object code is not eligible for the full range of copyright protection. Although some scholarly authority supports that view, we have previously rejected it based on the language and legislative history of the Copyright Act.
As recommended by the National Commission on New Technological Uses of Copyrighted Works (CONTU), the 1980 amendments to the Copyright Act unambiguously extended copyright protection to computer programs. Nor does the Act require that a work be directly accessible to humans in order to be eligible for copyright protection. Rather, it extends protection to all original works “which...can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device.” 17 U.S.C. § 102(a). The statutory language, read together with the CONTU report, leads inexorably to the conclusion that the copyright in a computer program extends to the object code version of the program.
Nor does a refusal to recognize a per se right to disassemble object code lead to an absurd result. The ideas and functional concepts underlying many types of computer programs, including word processing programs, spreadsheets, and video game displays, are readily discernible without the need for disassembly, because the operation of such programs is visible on the computer screen. The need to disassemble object code arises, if at all, only in connection with operations systems, system interface procedures, and other programs that are not visible to the user when operating—and then only when no alternative means of gaining an understanding of those ideas and functional concepts exists.
In our view, consideration of the unique nature of computer object code thus is more appropriate as part of the case-by-case, equitable “fair use” analysis authorized by section 107 of the Act. Accordingly, we reject Accolade's second argument.
Section 117
Section 117 of the Copyright Act allows the lawful owner of a copy of a computer program to copy or adapt the program if the new copy or adaptation “is created as an essential step in the utilization of the computer program in conjunction with a machine and...is used in no other manner.” 17 U.S.C. § 117(1). Accolade contends that section 117 authorizes disassembly of the object code in a copyrighted computer program.
Section 117 was enacted on the recommendation of CONTU, which noted that “because the placement of any copyrighted work into a computer is the preparation of a copy [since the program is loaded into the computer's memory], the law should provide that persons in rightful possession of copies of programs be able to use them freely without fear of exposure to copyright liability.” CONTU Report at 13. We think it is clear that Accolade's use went far beyond that contemplated by CONTU and authorized by section 117. Section 117 does not purport to protect a user who disassembles object code, converts it from assembly into source code, and makes printouts and photocopies of the refined source code version.
D. Fair Use
Accolade contends, finally, that its disassembly of copyrighted object code as a necessary step in its examination of the unprotected ideas and functional concepts embodied in the code is a fair use that is privileged by section 107 of the Act. Because, in the case before us, disassembly is the only means of gaining access to those unprotected aspects of the program, and because Accolade has a legitimate interest in gaining such access (in order to determine how to make its cartridges compatible with the Genesis console), we agree with Accolade. Where there is good reason for studying or examining the unprotected aspects of a copyrighted computer program, disassembly for purposes of such study or examination constitutes a fair use.
1. As a preliminary matter, we reject Sega's contention that the assertion of a fair use defense in connection with the disassembly of object code is precluded by statute. First, Sega argues that not only does section 117 of the Act not authorize disassembly of object code, but it also constitutes a legislative determination that any copying of a computer program other than that authorized by section 117 cannot be considered a fair use of that program under section 107. That argument verges on the frivolous. Each of the exclusive rights created by section 106 of the Copyright Act is expressly made subject to all of the limitations contained in sections 107 through 120. 17 U.S.C. § 106. Nothing in the language or the legislative history of section 117, or in the CONTU Report, suggests that section 117 was intended to preclude the assertion of a fair use defense with respect to uses of computer programs that are not covered by section 117, nor has section 107 been amended to exclude computer programs from its ambit.
Moreover, sections 107 and 117 serve entirely different functions. Section 117 defines a narrow category of copying that is lawful per se. 17 U.S.C. § 117. Section 107, by contrast, establishes a defense to an otherwise valid claim of copyright infringement. It provides that particular instances of copying that otherwise would be actionable are lawful, and sets forth the factors to be considered in determining whether the defense applies. Id. § 107. The fact that Congress has not chosen to provide a per se exemption to section 106 for disassembly does not mean that particular instances of disassembly may not constitute fair use.
Second, Sega maintains that the language and legislative history of section 906 of the Semiconductor Chip Protection Act of 1984 (SCPA) establish that Congress did not intend that disassembly of object code be considered a fair use. Section 906 of the SCPA authorizes the copying of the “mask work” on a silicon chip in the course of reverse engineering the chip. 17 U.S.C. § 906. The mask work in a standard ROM chip, such as those used in the Genesis console and in Genesis-compatible cartridges, is a physical representation of the computer program that is embedded in the chip. The zeros and ones of binary object code are represented in the circuitry of the mask work by open and closed switches. Sega contends that Congress's express authorization of copying in the particular circumstances set forth in section 906 constitutes a determination that other forms of copying of computer programs are prohibited.
The legislative history of the SCPA reveals, however, that Congress passed a separate statute to protect semiconductor chip products because it believed that semiconductor chips were intrinsically utilitarian articles that were not protected under the Copyright Act. H.R. Rep. No. 781, 98th Cong., 2d Sess. 8-10, reprinted in 1984 U.S.C.C.A.N. 5750, 5757-59. Accordingly, rather than amend the Copyright Act to extend traditional copyright protection to chips, it enacted “a sui generis form of protection, apart from and independent of the copyright laws.” Because Congress did not believe that semiconductor chips were eligible for copyright protection in the first instance, the fact that it included an exception for reverse engineering of mask work in the SCPA says nothing about its intent with respect to the lawfulness of disassembly of computer programs under the Copyright Act. Nor is the fact that Congress did not contemporaneously amend the Copyright Act to permit disassembly significant, since it was focusing on the protection to be afforded to semiconductor chips. Here we are dealing not with an alleged violation of the SCPA, but with the copying of a computer program, which is governed by the Copyright Act. Moreover, Congress expressly stated that it did not intend to “limit, enlarge or otherwise affect the scope, duration, ownership or subsistence of copyright protection ... in computer programs, data bases, or any other copyrightable works embodied in semiconductor chip products.” Accordingly, Sega's second statutory argument also fails. We proceed to consider Accolade's fair use defense.
2. Section 107 lists the factors to be considered in determining whether a particular use is a fair one. Those factors include: (1) the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes; (2) the nature of the copyrighted work; (3) the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and (4) the effect of the use upon the potential market for or value of the copyrighted work. 17 U.S.C. § 107. The statutory factors are not exclusive. Rather, the doctrine of fair use is in essence “an equitable rule of reason.” Fair use is a mixed question of law and fact. “Where the district court has found facts sufficient to evaluate each of the statutory factors,” an appellate court may resolve the fair use question as a matter of law.
In determining that Accolade's disassembly of Sega's object code did not constitute a fair use, the district court treated the first and fourth statutory factors as dispositive, and ignored the second factor entirely. Given the nature and characteristics of Accolade's direct use of the copied works, the ultimate use to which Accolade put the functional information it obtained, and the nature of the market for home video entertainment systems, we conclude that neither the first nor the fourth factor weighs in Sega's favor. In fact, we conclude that both factors support Accolade's fair use defense, as does the second factor, a factor which is important to the resolution of cases such as the one before us.
(a) Purpose and character of use
With respect to the first statutory factor, we observe initially that the fact that copying is for a commercial purpose weighs against a finding of fair use. Harper & Row, 471 U.S. at 562. However, the presumption of unfairness that arises in such cases can be rebutted by the characteristics of a particular commercial use. Further “the commercial nature of a use is a matter of degree, not an absolute.”
Sega argues that because Accolade copied its object code in order to produce a competing product, the Harper & Row presumption applies and precludes a finding of fair use. That analysis is far too simple and ignores a number of important considerations. We must consider other aspects of “the purpose and character of the use” as well. As we have noted, the use at issue was an intermediate one only and thus any commercial “exploitation” was indirect or derivative.
The declarations of Accolade's employees indicate, and the district court found, that Accolade copied Sega's software solely in order to discover the functional requirements for compatibility with the Genesis console — aspects of Sega's programs that are not protected by copyright. 17 U.S.C. § 102(b). With respect to the video game programs contained in Accolade's game cartridges, there is no evidence in the record that Accolade sought to avoid performing its own creative work. Indeed, most of the games that Accolade released for use with the Genesis console were originally developed for other hardware systems. Moreover, with respect to the interface procedures for the Genesis console, Accolade did not seek to avoid paying a customarily charged fee for use of those procedures, nor did it simply copy Sega's code; rather, it wrote its own procedures based on what it had learned through disassembly. Taken together, these facts indicate that although Accolade's ultimate purpose was the release of Genesis-compatible games for sale, its direct purpose in copying Sega's code, and thus its direct use of the copyrighted material, was simply to study the functional requirements for Genesis compatibility so that it could modify existing games and make them usable with the Genesis console. Moreover, as we discuss below, no other method of studying those requirements was available to Accolade. On these facts, we conclude that Accolade copied Sega's code for a legitimate, essentially non-exploitative purpose, and that the commercial aspect of its use can best be described as of minimal significance.
We further note that we are free to consider the public benefit resulting from a particular use notwithstanding the fact that the alleged infringer may gain commercially. Public benefit need not be direct or tangible, but may arise because the challenged use serves a public interest. In the case before us, Accolade's identification of the functional requirements for Genesis compatibility has led to an increase in the number of independently designed video game programs offered for use with the Genesis console. It is precisely this growth in creative expression, based on the dissemination of other creative works and the unprotected ideas contained in those works, that the Copyright Act was intended to promote. See Feist. The fact that Genesis-compatible video games are not scholarly works, but works offered for sale on the market, does not alter our judgment in this regard. We conclude that given the purpose and character of Accolade's use of Sega's video game programs, the presumption of unfairness has been overcome and the first statutory factor weighs in favor of Accolade.
(b) Effect on market for copyrighted work
As applied, the fourth statutory factor, effect on the potential market for the copyrighted work, bears a close relationship to the “purpose and character” inquiry in that it, too, accommodates the distinction between the copying of works in order to make independent creative expression possible and the simple exploitation of another's creative efforts. We must, of course, inquire whether, “if [the challenged use] should become widespread, it would adversely affect the potential market for the copyrighted work,” Sony Corp. v. Universal City Studios, 464 U.S. 417, 451 (1984), by diminishing potential sales, interfering with marketability, or usurping the market. If the copying resulted in the latter effect, all other considerations might be irrelevant. The Harper & Row Court found a use that effectively usurped the market for the copyrighted work by supplanting that work to be dispositive. However, the same consequences do not and could not attach to a use which simply enables the copier to enter the market for works of the same type as the copied work.
Unlike the defendant in Harper & Row, which printed excerpts from President Ford's memoirs verbatim with the stated purpose of “scooping” a Time magazine review of the book, Accolade did not attempt to “scoop” Sega's release of any particular game or games, but sought only to become a legitimate competitor in the field of Genesis-compatible video games. Within that market, it is the characteristics of the game program as experienced by the user that determine the program's commercial success. As we have noted, there is nothing in the record that suggests that Accolade copied any of those elements.
By facilitating the entry of a new competitor, the first lawful one that is not a Sega licensee, Accolade's disassembly of Sega's software undoubtedly “affected” the market for Genesis-compatible games in an indirect fashion. We note, however, that while no consumer except the most avid devotee of President Ford's regime might be expected to buy more than one version of the President's memoirs, video game users typically purchase more than one game. There is no basis for assuming that Accolade's “Ishido” has significantly affected the market for Sega's “Altered Beast”, since a consumer might easily purchase both; nor does it seem unlikely that a consumer particularly interested in sports might purchase both Accolade's “Mike Ditka Power Football” and Sega's “Joe Montana Football”, particularly if the games are, as Accolade contends, not substantially similar. In any event, an attempt to monopolize the market by making it impossible for others to compete runs counter to the statutory purpose of promoting creative expression and cannot constitute a strong equitable basis for resisting the invocation of the fair use doctrine. Thus, we conclude that the fourth statutory factor weighs in Accolade's, not Sega's, favor, notwithstanding the minor economic loss Sega may suffer.
(c) Nature of copyrighted work
The second statutory factor, the nature of the copyrighted work, reflects the fact that not all copyrighted works are entitled to the same level of protection. The protection established by the Copyright Act for original works of authorship does not extend to the ideas underlying a work or to the functional or factual aspects of the work. 17 U.S.C. § 102(b). To the extent that a work is functional or factual, it may be copied, Baker v. Selden, as may those expressive elements of the work that “must necessarily be used as incident to” expression of the underlying ideas, functional concepts, or facts, Works of fiction receive greater protection than works that have strong factual elements, such as historical or biographical works or works that have strong functional elements, such as accounting textbooks, Baker. Works that are merely compilations of fact are copyrightable, but the copyright in such a work is “thin.” Feist.
Computer programs pose unique problems for the application of the “idea/expression distinction” that determines the extent of copyright protection. To the extent that there are many possible ways of accomplishing a given task or fulfilling a particular market demand, the programmer's choice of program structure and design may be highly creative and idiosyncratic. However, computer programs are, in essence, utilitarian articles—articles that accomplish tasks. As such, they contain many logical, structural, and visual display elements that are dictated by the function to be performed, by considerations of efficiency, or by external factors such as compatibility requirements and industry demands. Computer Assoc. Int'l, Inc. v. Altai, Inc. In some circumstances, even the exact set of commands used by the programmer is deemed functional rather than creative for purposes of copyright. “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.” CONTU Report at 20.
Because of the hybrid nature of computer programs, there is no settled standard for identifying what is protected expression and what is unprotected idea in a case involving the alleged infringement of a copyright in computer software. We are in wholehearted agreement with the Second Circuit's recent observation that “thus far, many of the decisions in this area reflect the court's attempt to fit the proverbial square peg in a round hole.” In 1986, the Third Circuit attempted to resolve the dilemma by suggesting that the idea or function of a computer program is the idea of the program as a whole, and “everything that is not necessary to that purpose or function [is] part of the expression of that idea.” Whelan Assoc., Inc. v. Jaslow Dental Laboratory, Inc. The Whelan rule, however, has been widely — and soundly — criticized as simplistic and overbroad. See Altai (citing cases, treatises, and articles). In reality, “a computer program's ultimate function or purpose is the composite result of interacting subroutines. Since each subroutine is itself a program, and thus, may be said to have its own `idea,' Whelan's general formulation...is descriptively inadequate.” Id. For example, the computer program at issue in the case before us, a video game program, contains at least two such subroutines — the subroutine that allows the user to interact with the video game and the subroutine that allows the game cartridge to interact with the console. Under a test that breaks down a computer program into its component subroutines and sub-subroutines and then identifies the idea or core functional element of each, such as the test recently adopted by the Second Circuit in Altai, many aspects of the program are not protected by copyright. In our view, in light of the essentially utilitarian nature of computer programs, the Second Circuit's approach is an appropriate one.
Sega argues that even if many elements of its video game programs are properly characterized as functional and therefore not protected by copyright, Accolade copied protected expression. Sega is correct. The record makes clear that disassembly is wholesale copying. Because computer programs are also unique among copyrighted works in the form in which they are distributed for public use, however, Sega's observation does not bring us much closer to a resolution of the dispute.
The unprotected aspects of most functional works are readily accessible to the human eye. The systems described in accounting textbooks or the basic structural concepts embodied in architectural plans, to give two examples, can be easily copied without also copying any of the protected, expressive aspects of the original works. Computer programs, however, are typically distributed for public use in object code form, embedded in a silicon chip or on a floppy disk. For that reason, humans often cannot gain access to the unprotected ideas and functional concepts contained in object code without disassembling that code — i.e., making copies. Atari Games Corp. v. Nintendo of America, 975 F.2d 832 (Fed. Cir. 1992).
Sega argues that the record does not establish that disassembly of its object code is the only available method for gaining access to the interface specifications for the Genesis console, and the district court agreed. An independent examination of the record reveals that Sega misstates its contents, and demonstrates that the district court committed clear error in this respect.
First, the record clearly establishes that humans cannot read object code. Sega makes much of Mike Lorenzen's statement that a reverse engineer can work directly from the zeros and ones of object code but “it's not as fun.” In full, Lorenzen's statements establish only that the use of an electronic decompiler is not absolutely necessary. Trained programmers can disassemble object code by hand. Because even a trained programmer cannot possibly remember the millions of zeros and ones that make up a program, however, he must make a written or computerized copy of the disassembled code in order to keep track of his work. See generally Johnson-Laird, Technical Demonstration of “Decompilation”, reprinted in Reverse Engineering: Legal and Business Strategies for Competitive Design in the 1990's 102 (Prentice Hall Law & Bus. ed. 1992). The relevant fact for purposes of Sega's copyright infringement claim and Accolade's fair use defense is that translation of a program from object code into source code cannot be accomplished without making copies of the code.
Second, the record provides no support for a conclusion that a viable alternative to disassembly exists. The district court found that Accolade could have avoided a copyright infringement claim by “peeling” the chips contained in Sega's games or in the Genesis console, as authorized by section 906 of the SCPA, 17 U.S.C. § 906. Even Sega's amici agree that this finding was clear error. The declaration of Dr. Harry Tredennick, an expert witness for Accolade, establishes that chip peeling yields only a physical diagram of the object code embedded in a ROM chip. It does not obviate the need to translate object code into source code.
The district court also suggested that Accolade could have avoided a copyright infringement suit by programming in a “clean room”. That finding too is clearly erroneous. A “clean room” is a procedure used in the computer industry in order to prevent direct copying of a competitor's code during the development of a competing product. Programmers in clean rooms are provided only with the functional specifications for the desired program. As Dr. Tredennick explained, the use of a clean room would not have avoided the need for disassembly because disassembly was necessary in order to discover the functional specifications for a Genesis-compatible game.
In summary, the record clearly establishes that disassembly of the object code in Sega's video game cartridges was necessary in order to understand the functional requirements for Genesis compatibility. The interface procedures for the Genesis console are distributed for public use only in object code form, and are not visible to the user during operation of the video game program. Because object code cannot be read by humans, it must be disassembled, either by hand or by machine. Disassembly of object code necessarily entails copying. Those facts dictate our analysis of the second statutory fair use factor. If disassembly of copyrighted object code is per se an unfair use, the owner of the copyright gains a de facto monopoly over the functional aspects of his work — aspects that were expressly denied copyright protection by Congress. 17 U.S.C. § 102(b). In order to enjoy a lawful monopoly over the idea or functional principle underlying a work, the creator of the work must satisfy the more stringent standards imposed by the patent laws. Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141 (1989). Sega does not hold a patent on the Genesis console.
Because Sega's video game programs contain unprotected aspects that cannot be examined without copying, we afford them a lower degree of protection than more traditional literary works. In light of all the considerations discussed above, we conclude that the second statutory factor also weighs in favor of Accolade.
(d) Extent of taking
As to the third statutory factor, Accolade disassembled entire programs written by Sega. Accordingly, the third factor weighs against Accolade. The fact that an entire work was copied does not, however, preclude a finding a fair use. Sony. In fact, where the ultimate (as opposed to direct) use is as limited as it was here, the factor is of very little weight.
(e) Summary of fair–use factors
In summary, careful analysis of the purpose and characteristics of Accolade's use of Sega's video game programs, the nature of the computer programs involved, and the nature of the market for video game cartridges yields the conclusion that the first, second, and fourth statutory fair use factors weigh in favor of Accolade, while only the third weighs in favor of Sega, and even then only slightly. Accordingly, Accolade clearly has by far the better case on the fair use issue.
We are not unaware of the fact that to those used to considering copyright issues in more traditional contexts, our result may seem incongruous at first blush. To oversimplify, the record establishes that Accolade, a commercial competitor of Sega, engaged in wholesale copying of Sega's copyrighted code as a preliminary step in the development of a competing product. However, the key to this case is that we are dealing with computer software, a relatively unexplored area in the world of copyright law. We must avoid the temptation of trying to force “the proverbial square peg into a round hole.” Altai.
In determining whether a challenged use of copyrighted material is fair, a court must keep in mind the public policy underlying the Copyright Act:
The immediate effect of our copyright law is to secure a fair return for an author's creative labor. But the ultimate aim is, by this incentive, to stimulate artistic creativity for the general public good. [Sony.]
When technological change has rendered an aspect or application of the Copyright Act ambiguous, “the Copyright Act must be construed in light of this basic purpose.” Id. As discussed above, the fact that computer programs are distributed for public use in object code form often precludes public access to the ideas and functional concepts contained in those programs, and thus confers on the copyright owner a de facto monopoly over those ideas and functional concepts. That result defeats the fundamental purpose of the Copyright Act — to encourage the production of original works by protecting the expressive elements of those works while leaving the ideas, facts, and functional concepts in the public domain for others to build on. Feist.
Sega argues that the considerable time, effort, and money that went into development of the Genesis and Genesis-compatible video games militate against a finding of fair use. Borrowing from antitrust principles, Sega attempts to label Accolade a “free rider” on its product development efforts. In Feist Publications, however, the Court unequivocally rejected the “sweat of the brow” rationale for copyright protection. Under the Copyright Act, if a work is largely functional, it receives only weak protection. “This result is neither unfair nor unfortunate. It is the means by which copyright advances the progress of science and art.” Id. Here, while the work may not be largely functional, it incorporates functional elements which do not merit protection. The equitable considerations involved weigh on the side of public access. Accordingly, we reject Sega's argument.
We conclude that where disassembly is the only way to gain access to the ideas and functional elements embodied in a copyrighted computer program and where there is a legitimate reason for seeking such access, disassembly is a fair use of the copyrighted work, as a matter of law. Our conclusion does not, of course, insulate Accolade from a claim of copyright infringement with respect to its finished products. Sega has reserved the right to raise such a claim, and it may do so on remand.
Notes
1. Sega produces a video game console generally similar in operation to the Nintendo console described above in the Galoob case. The Sega hardware “platform” differs from Nintendo's, as a Macintosh platform differs from that of a “PC clone.” Hence, a video game cartridge that works on one system does not work on the other. Sega and Nintendo sought to “license” access to their hardware platforms. A company that writes video game software would obtain such a license by contracting with Sega (or Nintendo) to modify the computer program for the game (the video game software) sufficiently to make it compatible with the hardware, encode the computer program into a cartridge designed to work with the hardware, and sell the finished cartridge to the software company for it to resell through video game software distribution channels. The license fee is incorporated into the price that the software company pays for the cartridge.
In the Accolade case, Accolade had negotiated with Sega for such a license, but negotiations broke down over price. According to Accolade, Sega demanded a price per cartridge that greatly exceeded the price for a comparable unlicensed cartridge. Accolade claims that the difference between the two prices meant the difference to it between operating at a profit or at a loss.
The district court rejected Accolade's copyright misuse defense and antitrust tying allegations, which referred to the circumstances surrounding the license negotiations. What are the copyright misuse implications of the opinion of the court of appeals?
2. What intellectual property or proprietary right is the subject matter of Sega's licenses to video game software companies?
What is Sega's legally protected interest in controlling access to the Sega Genesis console? What is the legal theory? What should be the protection?
3. The operation of the TMSS (security system) is explained in general terms in the court's opinion; an elaboration follows. A video game software cartridge contains approximately one million bytes of code. Near the first address of the ROM in which the software is encoded, Sega placed the ASCII code for the letters S-E-G-A and a few bytes of initialization code. (A total of approximately 20 bytes out of the million bytes in the ROM comprise the “key” to “unlock” the TMSS “lock” on the Sega console.) If a microprocessor in the console is able to read the code for S-E-G-A in the appropriate locations of the cartridge's ROM, the microprocessor causes the Sega Message to flash on the screen and starts the game playing. If the microprocessor does not read S-E-G-A at those locations, nothing happens; the game simply does not play.
Why did Sega rely on an “intermediate copying” theory of copyright infringement in this case, instead of suing Accolade for its blatant reproduction and distribution of the entire S-E-G-A key in Accolade's regular commercial video game cartridges?
4. Accolade reverse-engineered Sega's previous model of Genesis in 1989, announced “Ishido” for Sega consoles in June 1990, and began shipping Ishido in December 1990. Sega demonstrated the revised Genesis III console (the one with lock and key) at a January 1991 trade show, showing that Accolade's Ishido would not run on Genesis III, and shipped Genesis III to U.S. customers in September 1991. When end users started returning Accolade cartridges to retailers (such as Toys–R–Us), because they would not work on Genesis III consoles, the retailers stopped buying Accolade cartridges and returned their inventory to Accolade, demanding refunds. Even though, initially, only a small fraction of the installed Sega console base had the security system, causing customer returns to occur, the retailers did not want to be bothered with any such returns. The adoption of the security system thus had an immediate effect far out of proportion to the share of the market for which consoles with security systems accounted.
5. The court considered the defense of fair use. Are any further defenses suggested by viewing the case in terms of Accolade's vicariously asserting rights of the end user?
Would it make a difference if the hardware platform is a Macintosh instead of a Genesis? If the software is an application program (such as Lotus 1-2-3 or WordPerfect) instead of a video game?
6. The court of appeals said that Sega had no patent on the security system. In fact, such a Sega patent was then pending and subsequently issued.
How could that fact affect the outcome? (There is no statutory counterpart of 17 U.S.C. § 107 in title 35. See also Roche Prods., Inc. v. Bolar Pharmaceutical Co., 733 F.2d 858 (Fed. Cir.), cert. denied, 469 U.S. 856 (1984), limiting the so-called experimental use defense to noncommercial use.)
7. A detailed examination of this case is available on the Internet. See Julie E. Cohen, Reverse Engineering and the Rise of Electronic Vigilantism: Intellectual Property Implications of "Lock-Out" Programs, 68 S. Cal. L. Rev. 1091 (1995), on line at this
link.
Sony Computer Entertainment v. Connectix Corp.
United States Court of Appeals
203 F.3d 596 (9th Cir. 2000)
Before: CHOY, CANBY, SILVERMAN.
CANBY, Circuit Judge.
In this case we are called upon once again to apply the principles of copyright law to computers and their software, to determine what must be protected as expression and what must be made accessible to the public as function. Sony Computer Entertainment, Inc., which brought this copyright infringement action, produces and markets the Sony PlayStation console, a small computer with hand controls that connects to a television console and plays games that are inserted into the PlayStation on compact discs (CDs). Sony owns the copyright on the basic input-output system or BIOS, which is the software program that operates its PlayStation. Sony has asserted no patent rights in this proceeding.
The defendant, Connectix Corporation, makes and sells a software program called “Virtual Game Station.” The purpose of the Virtual Game Station is to emulate on a regular computer the functioning of the Sony PlayStation console, so that computer owners who buy the Virtual Game Station software can play Sony PlayStation games on their computers. The Virtual Game Station does not contain any of Sony's copyrighted material.
In the process of producing the Virtual Game Station, however, Connectix repeatedly copied Sony's copyrighted BIOS during a process of “reverse engineering” that Connectix conducted in order to find out how the Sony PlayStation worked. Sony claimed infringement and sought a preliminary injunction. The district court concluded that Sony was likely to succeed on its infringement claim because Connectix's “intermediate copying” was not a protected “fair use” under 17 U.S.C. § 107. The district court enjoined Connectix from selling the Virtual Game Station or from copying or using the Sony BIOS code in the development of other Virtual Game Station products.
Connectix now appeals. We reverse and remand with instructions to dissolve the injunction. The intermediate copies made and used by Connectix during the course of its reverse engineering of the Sony BIOS were protected fair use, necessary to permit Connectix to make its non-infringing Virtual Game Station function with PlayStation games. Any other intermediate copies made by Connectix do not support injunctive relief, even if those copies were infringing.
I. Background
A. The products
Sony is the developer, manufacturer and distributor of both the Sony PlayStation and Sony PlayStation games. Sony also licenses other companies to make games that can play on the PlayStation. The PlayStation system consists of a console (essentially a mini-computer), controllers, and software that produce a three-dimensional game for play on a television set. The PlayStation games are CDs that load into the top of the console.
The PlayStation console contains both (1) hardware components and (2) software known as firmware that is written onto a read-only memory (ROM) chip. The firmware is the Sony BIOS. Sony has a copyright on the BIOS. It has claimed no patent relevant to this proceeding on any component of the PlayStation.
Connectix's Virtual Game Station is software that “emulates” the functioning of the PlayStation console. That is, a consumer can load the Virtual Game Station software onto a computer, load a PlayStation game into the computer's CDROM drive, and play the PlayStation game. The Virtual Game Station software thus emulates both the hardware and firmware components of the Sony console. The Virtual Game Station does not play PlayStation games as well as Sony's PlayStation does. At the time of the injunction, Connectix had marketed its Virtual Game Station for Macintosh computer systems but had not yet completed Virtual Game Station software for Windows.
B. Reverse engineering
Copyrighted software ordinarily contains both copyrighted and unprotected or functional elements. Sega Enters. Ltd. v. Accolade, Inc., 977 F.2d 1510, 1520 (9th Cir. 1993) (amended opinion); see 17 U.S.C. § 102(b) (copyright protection does not extend to any “idea, procedure, process, system, method of operation, concept, principle, or discovery” embodied in the copyrighted work). Software engineers designing a product that must be compatible with a copyrighted product frequently must “reverse engineer” the copyrighted product to gain access to the functional elements of the copyrighted product. See Andrew Johnson-Laird, Software Reverse Engineering in the Real World, 19 U. Dayton L. Rev. 843, 845-46 (1994).
Reverse engineering encompasses several methods of gaining access to the functional elements of a software program. They include: (1) reading about the program; (2) observing “the program in operation by using it on a computer;” (3) performing a “static examination of the individual computer instructions contained within the program;” and (4) performing a “dynamic examination of the individual computer instructions as the program is being run on a computer.” Id. at 846. Method (1) is the least effective, because individual software manuals often misdescribe the real product. See id. It would be particularly ineffective in this case because Sony does not make such information available about its PlayStation. Methods (2), (3), and (4) require that the person seeking access load the target program on to a computer, an operation that necessarily involves copying the copyrighted program into the computer's random access memory or RAM.
Method (2), observation of a program, can take several forms. The functional elements of some software programs, for example word processing programs, spreadsheets, and video game displays may be discernible by observation of the computer screen. See Sega, 977 F.2d at 1520. Of course, the reverse engineer in such a situation is not observing the object code itself, only the external visual expression of this code's operation on the computer. Here, the software program is copied each time the engineer boots up the computer, and the computer copies the program into RAM.
Other forms of observation are more intrusive. Operations systems, system interface procedures, and other programs like the Sony BIOS are not visible to the user when they are operating. See id. One method of “observing” the operation of these programs is to run the program in an emulated environment. In the case of the Sony BIOS, this meant operating the BIOS on a computer with software that simulated the operation of the PlayStation hardware; operation of the program, in conjunction with another program known as a “debugger,” permitted the engineers to observe the signals sent between the BIOS and other programs on the computer. This latter method required copying the Sony BIOS from a chip in the PlayStation onto the computer. The Sony BIOS was copied again each time the engineers booted up their computer and the computer copied the program into RAM. All of this copying was intermediate; that is, none of the Sony copyrighted material was copied into, or appeared in, Connectix's final product, the Virtual Game Station.
Methods (3) and (4) constitute “disassembly” of object code into source code. In each case, engineers use a program known as a “disassembler” to translate the ones and zeros of binary machine-readable object code into the words and mathematical symbols of source code. This translated source code is similar to the source code used originally to create the object code but lacks the annotations drafted by the authors of the program that help explain the functioning of the source code. In a static examination of the computer instructions, method (3), the engineer disassembles the object code of all or part of the program. The program must generally be copied one or more times to perform disassembly.
In a dynamic examination of the computer instructions, method (4), the engineer uses the disassembler program to disassemble parts of the program, one instruction at a time, while the program is running. This method also requires copying the program and, depending on the number of times this operation is performed, may require additional copying of the program into RAM every time the computer is booted up.
C. Connectix's reverse engineering of the Sony BIOS
Connectix began developing the Virtual Game Station for Macintosh on about July 1, 1998. In order to develop a PlayStation emulator, Connectix needed to emulate both the PlayStation hardware and the firmware (the Sony BIOS).
Connectix first decided to emulate the PlayStation's hardware. In order to do so, Connectix engineers purchased a Sony PlayStation console and extracted the Sony BIOS from a chip inside the console. Connectix engineers then copied the Sony BIOS into the RAM of their computers and observed the functioning of the Sony BIOS in conjunction with the Virtual Game Station hardware emulation software as that hardware emulation software was being developed by Connectix. The engineers observed the operation of the Sony BIOS through use of a debugging program that permitted the engineers to observe the signals sent between the BIOS and the hardware emulation software. During this process, Connectix engineers made additional copies of the Sony BIOS every time they booted up their computer and the Sony BIOS was loaded into RAM.
Once they had developed the hardware emulation software, Connectix engineers also used the Sony BIOS to “debug” the emulation software. In doing so, they repeatedly copied and disassembled discrete portions of the Sony BIOS.
Connectix also used the Sony BIOS to begin development of the Virtual Game Station for Windows. Specifically, they made daily copies to RAM of the Sony BIOS and used the Sony BIOS to develop certain Windows-specific systems for the Virtual Game Station for Windows. Although Connectix had its own BIOS at the time, Connectix engineers used the Sony BIOS because it contained CD-ROM code that the Connectix BIOS did not contain.
Early in the development process, Connectix engineer Aaron Giles disassembled a copy of the entire Sony BIOS that he had downloaded from the Internet. He did so for the purpose of testing a “disassembler” program he had written. The print-out of the source code was not used to develop the Virtual Game Station emulator. Connectix engineers initially used this copy of the Sony BIOS to begin the reverse engineering process, but abandoned it after realizing that it was a Japanese-language version.
During development of the Virtual Game Station, Connectix contacted Sony and requested “technical assistance” from Sony to complete the development of the Virtual Game Station. Connectix and Sony representatives met during September 1998. Sony declined Connectix's request for assistance.
Connectix completed Virtual Game Station for Macintosh computers in late December 1998 or early January 1999. Connectix announced its new product at the MacWorld Expo on January 5, 1999. At MacWorld, Connectix marketed the Virtual Game Station as a “PlayStation emulator.” The materials stated that the Virtual Game Station permits users to play “their favorite Playstation games” on a computer “even if you don't yet have a Sony PlayStation console.”
II. Discussion
Connectix admits that it copied Sony's copyrighted BIOS software in developing the Virtual Game Station but contends that doing so was protected as a fair use under 17 U.S.C. § 107.
A. Fair use
The fair use issue arises in the present context because of certain characteristics of computer software. The object code of a program may be copyrighted as expression, 17 U.S.C. § 102(a), but it also contains ideas and performs functions that are not entitled to copyright protection. See 17 U.S.C. § 102(b). Object code cannot, however, be read by humans. The unprotected ideas and functions of the code therefore are frequently undiscoverable in the absence of investigation and translation that may require copying the copyrighted material. We conclude that, under the facts of this case and our precedent, Connectix's intermediate copying and use of Sony's copyrighted BIOS was a fair use for the purpose of gaining access to the unprotected elements of Sony's software.
The general framework for analysis of fair use is established by statute, 17 U.S.C. § 107. We have applied this statute and the fair use doctrine to the disassembly of computer software in the case of Sega Enterprises Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1993) (amended opinion). Central to our decision today is the rule set forth in Sega:
[W]here disassembly is the only way to gain access to the ideas and functional elements embodied in a copyrighted computer program and where there is a legitimate reason for seeking such access, disassembly is a fair use of the copyrighted work, as a matter of law.
In Sega, we recognized that intermediate copying could constitute copyright infringement even when the end product did not itself contain copyrighted material. But this copying nonetheless could be protected as a fair use if it was “necessary” to gain access to the functional elements of the software itself. We drew this distinction because the Copyright Act protects expression only, not ideas or the functional aspects of a software program. We also recognized that, in the case of computer programs, this idea/expression distinction poses “unique problems” because computer programs are “in essence, utilitarian articles - articles that accomplish tasks. As such, they contain many logical, structural, and visual display elements that are dictated by the function to be performed, by considerations of efficiency, or by external factors such as compatibility requirements and industry demands.” Thus, the fair use doctrine preserves public access to the ideas and functional elements embedded in copyrighted computer software programs. This approach is consistent with the “`ultimate aim [of the Copyright Act], to stimulate artistic creativity for the general public good.'” Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 432 (1984) (quoting Twentieth Century Music Corp. v. Aiken, 422 U.S. 151, 156 (1975)).
We turn then to the statutory fair use factors, as informed by our precedent in Sega.
1. Nature of the copyrighted work
Under our analysis of the second statutory factor, nature of the copyrighted work, we recognize that “some works are closer to the core of intended copyright protection than others.” Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 586 (1994). Sony's BIOS lies at a distance from the core because it contains unprotected aspects that cannot be examined without copying. We consequently accord it a “lower degree of protection than more traditional literary works.” As we have applied this standard, Connectix's copying of the Sony BIOS must have been “necessary” to have been fair use. We conclude that it was.
There is no question that the Sony BIOS contains unprotected functional elements. Nor is it disputed that Connectix could not gain access to these unprotected functional elements without copying the Sony BIOS. Sony admits that little technical information about the functionality of the Sony BIOS is publicly available. The Sony BIOS is an internal operating system that does not produce a screen display to reflect its functioning. Consequently, if Connectix was to gain access to the functional elements of the Sony BIOS it had to be through a form of reverse engineering that required copying the Sony BIOS onto a computer. Sony does not dispute this proposition.
The question then becomes whether the methods by which Connectix reverse-engineered the Sony BIOS were necessary to gain access to the unprotected functional elements within the program. We conclude that they were. Connectix employed several methods of reverse engineering (observation and observation with partial disassembly) each of which required Connectix to make intermediate copies of copyrighted material. Neither of these methods renders fair use protection inapplicable. Sega expressly sanctioned disassembly. We see no reason to distinguish observation of copyrighted software in an emulated computer environment. Both methods require the reverse engineer to copy protected as well as unprotected elements of the computer program. Because this intermediate copying is the gravamen of the intermediate infringement claim, see 17 U.S.C. § 106(1); Sega, 977 F.2d at 1518-19, and both methods of reverse engineering require it, we find no reason inherent in these methods to prefer one to another as a matter of copyright law. Connectix presented evidence that it observed the Sony BIOS in an emulated environment to observe the functional aspects of the Sony BIOS. When this method of reverse engineering was unsuccessful, Connectix engineers disassembled discrete portions of the Sony BIOS to view directly the ideas contained therein. We conclude that intermediate copying in this manner was “necessary” within the meaning of Sega.
We decline to follow the approach taken by the district court. The district court did not focus on whether Connectix's copying of the Sony BIOS was necessary for access to functional elements. Instead, it found that Connectix's copying and use of the Sony BIOS to develop its own software exceeded the scope of Sega: “[T]hey disassembled Sony's code not just to study the concepts. They actually used that code in the development of [their] product.” This rationale is unpersuasive. It is true that Sega referred to “studying or examining the unprotected aspects of a copyrighted computer program.” But in Sega, Accolade's copying, observation and disassembly of Sega's game cartridges was held to be fair use, even though Accolade “loaded the disassembled code back into a computer, and experimented to discover the interface specifications for the Genesis console by modifying the programs and studying the results.” Thus, the distinction between “studying” and “use” is unsupported in Sega. Moreover, reverse engineering is a technically complex, frequently iterative process. Within the limited context of a claim of intermediate infringement, we find the semantic distinction between “studying” and “use” to be artificial, and decline to adopt it for purposes of determining fair use.
We also reject the argument, urged by Sony, that Connectix infringed the Sony copyright by repeatedly observing the Sony BIOS in an emulated environment, thereby making repeated copies of the Sony BIOS. These intermediate copies could not have been “necessary” under Sega, contends Sony, because Connectix engineers could have disassembled the entire Sony BIOS first, then written their own Connectix BIOS, and used the Connectix BIOS to develop the Virtual Game Station hardware emulation software. We accept Sony's factual predicate for the limited purpose of this appeal. Our doing so, however, does not aid Sony.
Sony contends that Connectix's reverse engineering of the Sony BIOS should be considered unnecessary on the rationale that Connectix's decision to observe the Sony BIOS in an emulated environment required Connectix to make more intermediate copies of the Sony BIOS than if Connectix had performed a complete disassembly of the program. Under this logic, at least some of the intermediate copies were not necessary within the meaning of Sega. This construction stretches Sega too far. The “necessity” we addressed in Sega was the necessity of the method, i.e., disassembly, not the necessity of the number of times that method was applied. In any event, the interpretation advanced by Sony would be a poor criterion for fair use. Most of the intermediate copies of the Sony BIOS were made by Connectix engineers when they booted up their computers and the Sony BIOS was copied into RAM. But if Connectix engineers had left their computers turned on throughout the period during which they were observing the Sony BIOS in an emulated environment, they would have made far fewer intermediate copies of the Sony BIOS (perhaps as few as one per computer).
Even if we were inclined to supervise the engineering solutions of software companies in minute detail, and we are not, our application of the copyright law would not turn on such a distinction. [fn9] Such a rule could be easily manipulated. More important, the rule urged by Sony would require that a software engineer, faced with two engineering solutions that each require intermediate copying of protected and unprotected material, often follow the least efficient solution. (In cases in which the solution that required the fewest number of intermediate copies was also the most efficient, an engineer would pursue it, presumably, without our urging.) This is precisely the kind of “wasted effort that the proscription against the copyright of ideas and facts . . . [is] designed to prevent.” Feist. Such an approach would erect an artificial hurdle in the way of the public's access to the ideas contained within copyrighted software programs. These are “aspects that were expressly denied copyright protection by Congress.” Sega (citing 17 U.S.C. § 102(b)). We decline to erect such a barrier in this case. If Sony wishes to obtain a lawful monopoly on the functional concepts in its software, it must satisfy the more stringent standards of the patent laws. See Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 160-61 (1989); Sega, 977 F.2d at 1526. This Sony has not done. The second statutory factor strongly favors Connectix.
2. Amount and substantiality of the portion used
With respect to the third statutory factor, amount and substantiality of the portion used in relation to the copyrighted work as a whole, Connectix disassembled parts of the Sony BIOS and copied the entire Sony BIOS multiple times. This factor therefore weighs against Connectix. But as we concluded in Sega, in a case of intermediate infringement when the final product does not itself contain infringing material, this factor is of “very little weight.” Sega, 977 F.2d at 1526-27; see also Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 449-50 (1984) (copying of entire work does not preclude fair use).
3. Purpose and character of the use
Under the first factor, purpose and character of the use, we inquire into whether Connectix's Virtual Game Station
merely supersedes the objects of the original creation, or instead adds something new, with a further purpose or different character, altering the first with new expression, meaning, or message; it asks, in other words, whether and to what extent the new work is “transformative.”
Campbell v. Acuff-Rose Music, Inc.
As an initial matter, we conclude that the district court applied an erroneous legal standard; the district court held that Connectix's commercial purpose in copying the Sony BIOS gave rise to a “presumption of unfairness that . . . can be rebutted by the characteristics of a particular commercial use.” Since Sega, however, the Supreme Court has rejected this presumption as applied to the first and fourth factor of the fair use analysis. ., 510 U.S. at 584, 594 (clarifying Sony). Instead, the fact that Connectix's copying of the Sony BIOS was for a commercial purpose is only a “separate factor that tends to weigh against a finding of fair use.”
We find that Connectix's Virtual Game Station is modestly transformative. The product creates a new platform, the personal computer, on which consumers can play games designed for the Sony PlayStation. This innovation affords opportunities for game play in new environments, specifically anywhere a Sony PlayStation console and television are not available, but a computer with a CD-ROM drive is. More important, the Virtual Game Station itself is a wholly new product, notwithstanding the similarity of uses and functions between the Sony PlayStation and the Virtual Game Station. The expressive element of software lies as much in the organization and structure of the object code that runs the computer as it does in the visual expression of that code that appears on a computer screen. See 17 U.S.C. § 102(a) (extending copyright protection to original works of authorship that “can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device”).
Sony does not claim that the Virtual Game Station itself contains object code that infringes Sony's copyright. We are therefore at a loss to see how Connectix's drafting of entirely new object code for its VGS program could not be transformative, despite the similarities in function and screen output.
Finally, we must weigh the extent of any transformation in Connectix's Virtual Game Station against the significance of other factors, including commercialism, that militate against fair use. Connectix's commercial use of the copyrighted material was an intermediate one, and thus was only “indirect or derivative.” Sega. Moreover, Connectix reverse-engineered the Sony BIOS to produce a product that would be compatible with games designed for the Sony PlayStation. We have recognized this purpose as a legitimate one under the first factor of the fair use analysis. Upon weighing these factors, we find that the first factor favors Connectix.
The district court ruled, however, that the Virtual Game Station was not transformative on the rationale that a computer screen and a television screen are interchangeable, and the Connectix product therefore merely “supplants” the Sony PlayStation console. The district court clearly erred. For the reasons stated above, the Virtual Game Station is transformative and does not merely supplant the PlayStation console. In reaching its decision, the district court apparently failed to consider the expressive nature of the Virtual Game Station software itself. Sony's reliance on Infinity Broadcast Corp. v. Kirkwood, 150 F.3d 104 (2d Cir. 1998), suffers from the same defect. The Infinity court reasoned that a “change of format, though useful, is not technically a transformation.” But the infringing party in that case was merely taking copyrighted radio transmissions and retransmitting them over telephone lines; there was no new expression. Infinity does not change our conclusion; the purpose and character of Connectix's copying points toward fair use.
4. Effect of the use upon the potential market
We also find that the fourth factor, effect of the use upon the potential market, favors Connectix. Under this factor, we consider
not only the extent of market harm caused by the particular actions of the alleged infringer, but also “whether unrestricted and widespread conduct of the sort engaged in by the defendant . . . would result in a substantially adverse impact on the potential market” for the original.
Acuff-Rose.
Whereas a work that merely supplants or supersedes another is likely to cause a substantially adverse impact on the potential market of the original, a transformative work is less likely to do so.
The district court found that “[t]o the extent that such a substitution [of Connectix's Virtual Game Station for Sony PlayStation console] occurs, Sony will lose console sales and profits.” We recognize that this may be so. But because the Virtual Game Station is transformative, and does not merely supplant the PlayStation console, the Virtual Game Station is a legitimate competitor in the market for platforms on which Sony and Sony-licensed games can be played. For this reason, some economic loss by Sony as a result of this competition does not compel a finding of no fair use. Sony understandably seeks control over the market for devices that play games Sony produces or licenses. The copyright law, however, does not confer such a monopoly. See Sega (“[A]n attempt to monopolize the market by making it impossible for others to compete runs counter to the statutory purpose of promoting creative expression and cannot constitute a strong equitable basis for resisting the invocation of the fair use doctrine.”). This factor favors Connectix.
The four statutory fair use factors must be “weighed together, in light of the purposes of copyright.” Acuff-Rose. Here, three of the factors favor Connectix; one favors Sony, and it is of little weight. Of course, the statutory factors are not exclusive, Harper & Row, 471 U.S. at 560, but we are unaware of other factors not already considered that would affect our analysis. Accordingly, we conclude that Connectix's intermediate copying of the Sony BIOS during the course of its reverse engineering of that product was a fair use under 17 U.S.C. § 107, as a matter of law. With respect to its claim of copyright infringement, Sony has not established either a likelihood of success on the merits or that the balance of hardships tips in its favor. Accordingly, we need not address defenses asserted by Connectix under 17 U.S.C. § 117(a)(1) and our doctrine of copyright misuse.
Conclusion
Connectix's reverse engineering of the Sony BIOS extracted from a Sony PlayStation console purchased by Connectix engineers is protected as a fair use. Other intermediate copies of the Sony BIOS made by Connectix, if they infringed Sony's copyright, do not justify injunctive relief. We reverse the district court's grant of a preliminary injunction on the ground of copyright infringement.11
| 11. We do not accept Sony's argument that the downloading of Sony's BIOS from the Internet was itself an infringement justifying the injunction. The evidence of record suggests that the downloaded BIOS played a minimal role, if any, in development of the Virtual Game Station. We conclude that, on this record, the downloading infringement, if such it was, would not justify our upholding the injunction on the development and sale of the Virtual Game Station. The Virtual Game Station itself infringes no copyright. Bearing in mind the goals of the copyright law, "to stimulate artistic creativity for the general public good," Sony, we conclude that there is a legitimate public interest in the publication of Connectix's software, and that this interest is not overborne by the record evidence related to the downloaded BIOS. The imposition of an injunction is discretionary. See 17 U.S.C. § 502(a). On this record, we conclude that it would be inappropriate to uphold the injunction because of Connectix's copying and use of the downloaded Sony BIOS; damages would adequately protect Sony's interest with respect to that alleged infringement. See Acuff-Rose, 510 U.S. at 578 n. 10 (discussing factors to be evaluated in deciding whether to enjoin product found to have exceeded bounds of fair use). |
We also reverse the district court's finding that Connectix's Virtual Game Station has tarnished the Sony PlayStation mark. For these reasons, the district court's injunction is dissolved and the case is remanded to the district court.
Notes
1. In Connectix the infringing use was transformative and also caused Sony to lose sales of consoles. The court excused this, saying: “because the Virtual Game Station is transformative, and does not merely supplant the PlayStation console, the Virtual Game Station is a legitimate competitor in the market for platforms on which Sony and Sony-licensed games can be played.” Why didn't the court just say that loss of console sales is immaterial since the statute looks only to the effect of the infringement on the market for the copyrighted work, not the effect on the market for other products of the copyright owner? What is the copyrighted work, here? What was the copyrighted work in the Sega case?
2. How does Connectix go beyond Sega?
3. Is there a derivative work here? Is the Virtual Game Station of Connectix a derivative work based on Sony's work? Connectix did make intermediate copies. The court held that the reproduction of intermediate copies was fair use because the use was transformative. What did Connectix transform?
4. Connectix is a rare case in that the use is transformative and also diverts sales from the copyright owner. The oranges of transformation are held to outweigh the apples of sales diversion. A
hypothetical case of both competition and transformation is presented in the LHOOQ pages: A Professor Annotator prepares a transformative derivative work that competes with Professor Chaucer's licensed derivative work, but the later work provides much greater functionality. Arguments are presented for each side, and the conclusion is drawn that finding fair use more greatly benefits the public because it "will in the long run more encourage the progress of knowledge and the arts." Further arguments on this issue are presented in a
related LHOOQ page.
DSC Communications Corp. v. DGI Technologies, Inc.
United States Court of Appeals
81 F.3d 597 (5th Cir. 1996)
Garza, Circuit Judge:
DSC Communications Corporation obtained a preliminary injunction enjoining DGI Technologies, Inc. from making removable copies of DSC Communication Corporation's copyrighted software. DSC Communications Corporation appeals from the granting of this preliminary injunction, arguing that it is too narrowly drawn. Because we hold that the district court did not abuse its discretion in crafting the injunction, we affirm.
Facts
DSC Communications Corporation (DSC) manufactures telephone switching systems (phone switches). DGI Technologies, Inc. (DGI) manufactures various devices, including microprocessor cards, that are used in DSC phone switches. DSC sued DGI for unfair competition on various grounds, including copyright infringement. DSC obtained a preliminary injunction in that suit, which prohibits DGI from making copies of DSC's copyrighted operating system software that can be removed from DSC's customer's premises. The injunction, however, does not prohibit DGI from making copies of DSC's copyrighted software that cannot be removed from DSC's customer's premises. DSC's appeals from the preliminary injunction, arguing that DGI should also be enjoined from making copies that cannot be removed from DSC's customer's premises.
A phone switch routes long distance telephone calls to their destinations. It consists of three principal components: (1) the switch matrix, which actually routes the telephone calls; (2) the trunk/line interface system, which converts long distance telephone signals into a form and sequence that can be handled by the switch matrix; and (3) a mass storage frame, which contains the software that operates the entire switching system.
Long distance signals must be in digital form and properly sequenced before they can be routed through the switch matrix. The trunk/line interface system converts the data arriving from the long distance telephone line into proper digital form and sequence. The trunk/line interface system is controlled by DSC's copyrighted software when it converts these signals. Once these signals have been “switched” — i.e. routed to their destination — the trunk/line system converts them back into a form in which they can be transmitted through the phone lines to their destination.
The trunk/line interface system is housed in metal cabinets called frames. These frames contain a number of shelves. The front of the shelves is open, and at the back of the shelves is a “backplane.” Cables carrying incoming telephone signals from the telephone line and outgoing signals to the switch matrix and the telephone lines are attached to the backplane. Groups of printed circuit boards called cards are inserted into the shelves of the frames from the front and connect to the backplane. These cards contain the components that translate the data from the telephone line into a format that can be used by the switch matrix and vice versa.
The principal cards in the frames are microprocessor cards. The microprocessor cards contain firmware, which is software embedded in a memory chip on the card. When a microprocessor card is inserted into the frame, it must boot up. That is, it must download DSC's copyrighted operating system software into its random access memory (“RAM”). The booting up process is similar to that used in personal computers, which also boot up by downloading operating system software from a floppy disk or hard disk when the computer is turned on or reset. A microprocessor card must download DSC's copyrighted operating system software when it is used in the phone switch.
DSC manufactures the entire phone switch system, and has a copyright on the software used in the phone switch. DSC sells phone switches, but does not sell the software necessary to operate them. Instead, it licenses the software to its customers. One of the customers to whom DSC sold a phone switch and licensed its software is NTS Communications Corporation (NTS). The licensing agreement between DSC and NTS prohibits NTS from copying the software, and only allows NTS to use the software in conjunction with the phone switch purchased from DSC.
DGI is attempting to develop a microprocessor card that can be used in DSC phone switches. Customers would use this card instead of using a DSC-manufactured card. DSC contends that DGI engaged in several acts of copyright infringement in its attempt to develop a microprocessor card. The alleged infringement at issue in this appeal is DGI's copying of DSC's copyrighted operating system software.
Because DSC did not sell its operating system software on the open market, the only way to gain access to the software was to license it from DSC. DGI needed to gain access to DSC's operating system software in order to develop a microprocessor card, because the microprocessor card had to be able to download the software into RAM, and had to be compatible with the software. To obtain access to the operating system software, DGI obtained access to a DSC phone switch owned by NTS. NTS gave DGI permission to use its phone switch to test microprocessor cards. In return for this permission, DGI gave NTS a ten percent discount on purchases of DGI cards, shelves and frames. DGI did more than merely test its cards, however. Without NTS's knowledge, it made copies of DSC's copyrighted software, and removed these copies from NTS's premises.
DGI used two methods to copy DSC's copyrighted software. First, it downloaded DSC's operating system into the memory of a DSC microprocessor card, out through a port on that card, and into a lap top computer. Second, DGI modified a DSC microprocessor card by adding chips designed to capture the communications between the card and another microprocessor card from which the operating system software would be obtained and a chip designed to hold and retain information on the downloading function when the microprocessor card was removed. DGI copied DSC's copyrighted operating system software using this modified microprocessor card.
DSC and DGI were already involved in litigation at the time that DGI was copying DSC's operating system software. DSC sued DGI for allegedly misappropriating its trade secrets to develop microprocessor cards for use is DSC's phone switch and for violating the Lanham Act in selling its cards. DGI counter-sued, alleging that DSC violated antitrust laws, misappropriated DGI's trade secrets, engaged in unfair competition and committed tortious interference with DGI's business relationships. When DSC learned that DGI was copying its operating system software, it amended its complaint to allege copyright infringement, and moved for a preliminary injunction to prevent DGI from continuing to copy, and benefitting from copying, the operating system software. The district court granted a preliminary injunction, prohibiting DGI from making any copies of DSC's operating system software that could be removed from NTS's premises. However, the injunction did not prohibit DGI from “downloading into dynamic RAM on a microprocessor or test microprocessor card which is incidental to the testing or operating of a compatible [microprocessor] card so long as the copy is not capable of being removed from the customer location and transported to any other location.” In other words, DGI could not continue to make copies of the operating system to take back to its lab and study, but it could test its microprocessor card on NTS's phone switch, even though DSC's operating system software would be downloaded into the microprocessor card's RAM.
Discussion
The decision to grant or deny a preliminary injunction lies within the discretion of the district court and will be reversed on appeal only upon a showing of abuse of discretion.
In order to obtain a preliminary injunction, DSC was required to demonstrate, among other things, a substantial likelihood of success on the merits. Because we hold that the district court did not abuse its discretion in finding that DSC did not have a substantial likelihood of success on the merits, we affirm the injunction.
DSC claims that it has a substantial likelihood of prevailing on its claim that DGI infringes upon its copyright every time DGI boots up one of its microprocessor cards on a DSC phone switch. When a DGI microprocessor card boots up, DSC's copyrighted operating system software is downloaded into the card's RAM. DSC contends that this downloading constitutes making a copy under copyright law. Therefore, DSC argues, booting up a DGI microprocessor card infringes upon DSC's copyright by making an unauthorized copy of DSC's copyrighted operating system software.
To prevail on its claim of copyright infringement, DSC will have to prove that DGI impermissibly copied or otherwise infringed upon DSC's copyright. Further, it will have to overcome DGI's affirmative defenses, including the defense of copyright misuse.
The parties agree that DSC owns a copyright on the operating system software. However, they disagree on whether booting up a DGI microprocessor card constitutes impermissible copying. DSC argues that an impermissible copy is made every time software is loaded onto a computer's RAM. See MAI Sys. Corp. v. Peak Computer, Inc., 991 F.2d 511, 518 (9th Cir. 1993), cert. dism'd, 114 S.Ct. 671 (1994). Because the licensing agreement between DSC and NTS only allows the software to be booted up on DSC equipment, this copying is not authorized. Therefore, DSC argues, booting up a microprocessor card violates its copyright. DGI does not dispute that a copy is made when the microprocessor cards are booted up. Instead, DGI argues, inter alia, that it is entitled to the defense of copyright misuse.
The district court did not err in implicitly finding that DSC did not have a substantial likelihood of prevailing on the merits because DGI may well prevail on its affirmative defense of copyright misuse. The defense of copyright misuse “bars a culpable plaintiff from prevailing on an action for the infringement of the misused copyright.” Lasercomb Am., Inc. v. Reynolds, 911 F.2d 970, 972 (4th Cir. 1990). See Mitchell Bros. Film Group v. Cinema Adult Theater, 604 F.2d 852, 865 n. 27 (5th Cir. 1979), cert. denied, 445 U.S. 917 (1980)(recognizing the copyright misuse defense). The defense is a recognition that while “copyright law [seeks] to increase the store of human knowledge and arts by awarding...authors with the exclusive rights to their works for a limited time...the granted monopoly power does not extend to property not covered by the...copyright.” Lasercomb.
The copyright misuse defense is analogous to the patent misuse defense. The patent misuse defense was recognized by the Supreme Court in Morton Salt Co. v. G.S. Suppiger, 314 U.S. 488 (1942). In that case, the plaintiff Morton Salt brought suit on the basis that the defendant had infringed upon Morton's patent in a salt-depositing machine. The salt tablets that the machine deposited were not themselves a patented item, but Morton's patent license required that licensees use only salt tablets produced by Morton. Morton was thereby using its patent to restrain competition in the sale of an item that was not within the scope of the patent's privilege. The Supreme Court held that, as a court of equity, it would not aid Morton in protecting its patent when Morton was using that patent in a manner contrary to public policy.
In Lasercomb, the Fourth Circuit extended the rationale behind Morton Salt to copyright misuse. Paraphrasing Morton Salt, the Fourth Circuit stated:
The grant to the author of the special privilege of a copyright carries out a public policy adopted by the Constitution and laws of the United States, “to promote the Progress of Science and useful arts, by securing for limited Times to [Authors]...the exclusive Right...” to their “original” works. U.S. Const., Art. I, § 8, cl. 8, 17 U.S.C. § 102. But the public policy which includes original works within the granted monopoly excludes from it all that is not embraced in the original expression. It equally forbids the use of the copyright to secure an exclusive right or limited monopoly not granted by the Copyright Office and which is contrary to public policy to grant.
We concur with the Fourth Circuit's characterization of the copyright misuse defense.
DGI may well prevail on the defense of copyright misuse, because DSC seems to be attempting to use its copyright to obtain a patent-like monopoly over unpatented microprocessor cards. Any competing microprocessor card developed for use on DSC phone switches must be compatible with DSC's copyrighted operating system software. In order to ensure that its card is compatible, a competitor such as DGI must test the card on a DSC phone switch. Such a test necessarily involves making a copy of DSC's copyrighted operating system, which copy is downloaded into the card's memory when the card is booted up. If DSC is allowed to prevent such copying, then it can prevent anyone from developing a competing microprocessor card, even though it has not patented the card. The defense of copyright misuse “forbids the use of the copyright to secure an exclusive right or limited monopoly not granted by the Copyright Office,” including a limited monopoly over microprocessor cards. See Lasercomb. Therefore, DGI's asserting the misuse defense could cast substantial doubt on the predictability of success by DSC.
Of course, we do not hold that DGI will successfully avail itself of the copyright misuse defense. After a trial on the merits, the district court may well decide that DSC did not commit copyright misuse, or that DGI cannot avail itself of the defense because it has “unclean hands.” We simply hold that the district court did not abuse its discretion in implicitly holding that DSC did not have a substantial likelihood of success on the merits because—based on the evidence before the district court — DGI may prevail on its misuse to the defense.
Because we hold that the district court did not abuse its discretion in crafting the preliminary injunction, we affirm.
Notes
1. Is it consistent with the rationale of the misuse doctrine to deny the infringer here the right to assert a misuse defense because it has “unclean hands”? Here, DSC wanted to protect the secrecy of its system, and therefore “did not sell its operating system software on the open market.” Hence, “the only way to gain access to the software was to license it from DSC.” Therefore, DGI gave NTS, a DSC licensee, a 10% discount to get NTS to let DGI “test” (nudge, nudge, wink, wink) microprocessor cards that NTS had obtained from DSC.
We may assume that DSC sold the cards to NTS subject to a restriction like that in shrinkwrap license cases, or like that in the MAI case. That is, NTS did not buy the software embedded in the microprocessor cards; DSC merely licensed NTS to use the software subject to restrictions, such as no copying. Thus, the district court said:
DSC...has a copyright on the software used in the phone switch. DSC sells phone switches, but does not sell the software necessary to operate them. Instead, it licenses the software to its customers. ...The licensing agreement between DSC and NTS prohibits NTS from copying the software, and only allows NTS to use the software in conjunction with the phone switch purchased from DSC.
It may therefore be assumed for purposes of this question that DGI downloaded DSC's operating system software in intentional contravention of the contractual arrangement between DSC and its licensee NTS.
a. Does public policy require that the court ignore or give no legal effect to the DSC–NTS contract, at least insofar as DGI's downloading is concerned?
b. What effect on the public interest does it have to treat the DSC–NTS contract as binding DGI not to download the software? To what extent is the sanctity of contracts (Pacta sunt servanda!) a part of public policy or a public interest factor?
c. Suppose that the Fifth Circuit thinks that both DSC and DGI are behaving badly. What does it do? Should there be a principle of comparative naughtiness?
If the Fifth Circuit says, “Unclean! Unclean! We refuse to have anything to do with either of you and your respective arguments!” — does it just reject application of the misuse defense, or does it refuse to take any legal action at all, including refusing to adjudicate the copyright infringement controversy? Does it just refuse to consider the misuse defense? See Atari Games Corp. v. Nintendo of America, Inc., 975 F.2d 832 (Fed. Cir. 1992).
d. What happens in an antitrust case when both the plaintiff and defendant are naughty persons? See, e.g., Perma Life Mufflers, Inc. v. International Parts Corp., 392 U.S. 134 (1968). In General Leaseways, Inc. v. National Truck Leasing Ass'n, 744 F.2d 588, 597 (7th Cir. 1984), the court stated that “whenever some maxim of equity (such as that to get equitable relief you must have clean hands) collides with the objectives of the antitrust laws, the equity maxim must give way.” Is that principle sound? Does it apply here?
e. Is the application of this concept appropriately calibrated in terms of whether legal (damages) or equitable (injunctive) relief is sought? Is an accounting for profits equitable rather than legal in this sense?
2. What is the rationale for the district court in letting DGI test its microprocessor cards on site at NTS but not letting DGI cart the code off to DGI's laboratory for closer study of the code? Why not in for a penny, in for a dollar?
3. Consider the language at the end of the Fifth Circuit's opinion. Is this just a holding of “no abuse of discretion” on the terms of a preliminary injunction? Is there any ruling on the merits of whether the Fifth Circuit takes a different view of the merits of these restrictions than the Ninth Circuit did in Triad Sys. Corp. v. Southeastern Express Co., 64 F.3d 1330 (9th Cir. 1995)?
4. Should the wideness of distribution of the product or the magnitude of the price make any copyright infringement law difference in these cases? Is something you buy off the shelf in a store different from a product on which the seller and buyer sign actual contracts?
5. The court considers patent–like protection and says:
If DSC is allowed to prevent such copying [downloading operating system software to RAM of microprocessor of card to test card's compatibility], then it can prevent anyone from developing a competing microprocessor card, even though it has not patented the card.
Under current patent law, however, it might be possible to patent the software, irrespective of whether the microprocessor card is patented. While the matter remains unsettled, the PTO's current software guidelines leave that door open. Under patent law, reverse engineering is generally permitted when it does not itself involve any act of infringement. But Roche Prods., Inc. v. Bolar Pharmaceutical Co., 733 F.2d 858 (Fed. Cir.), cert. denied, 469 U.S. 856 (1984), holds it patent infringement to practice the patented invention, for example, by making a patented product (as distinguished from merely using a patented product that was purchased from the patentee) in the course of developing a competing product. Do you see any possibility that the Federal Circuit would hold it misuse to bring such a patent infringement suit against a software reverse engineer who infringed a software patent in the course of her reverse engineering? Suppose article of manufacture claims for computer programs (i.e., a storage medium encoded with a computer program) are sustained as patentable. It would then appear that saving to disk a copy of a modified version of a computer program subject to an article of manufacture claim would be an act of patent infringement — i.e., a making. Executing the derivative work version would appear to be infringement by using.
What are the implications for reverse engineering? Consider DSC redux as a patent case.
Reverse Engineering Under the Semiconductor Chip
Protection Act of 1984 (SCPA) – 17 U.S.C. § 906(a)
§ 906. Limitation on exclusive rights: reverse engineering
(a) Notwithstanding the provisions of section 905 [which describe mask work owners' exclusive rights], it is not infringement of the exclusive rights of the owner of a mask work for—
(1) a person to reproduce the mask work solely for the purpose of teaching, analyzing, or evaluating the concepts or techniques embodied in the mask work or the circuitry, logic flow, or organization of components used in the mask work; or
(2) a person who performs the analysis or evaluation described in paragraph (1) to incorporate the results of such conduct in an original mask work which is made to be distributed.
Notes
1. Does this provision do anything more than make copyright law's fair-use doctrine applicable to chip layout mask works?
2. Comment on how the following statement from the House Report on the SCPA (H.R. Rep. No. 98–781, 98th Cong., 2d Sess. (1984)) captures (or not) the essence of copyright's fair use doctrine (id. at 22):
It is the intent of the Committee to permit, under the reverse engineering limitation, the “unauthorized” creation of a second mask work whose layout, in substantial part, is similar to the layout of a protected mask work — if the second mask work was the product of substantial study and analysis, and not the mere result of plagiarism accomplished without such study or analysis.
See also 130 Cong. Rec. S12917 (daily ed. Oct. 3, 1984):
If the resulting semiconductor chip product is not substantially identical to the original, and its design involved significant toil and investment so that it is not a mere plagiarism, it does not infringe the original chip, even if the layout of the two chips is, in substantial part, similar.
How do these passages square with the subsequent Feist decision?
3. What do you think is the rationale for the requirement in § 906(a) that the second mask work be “original”? What does “original” mean? Does it mean not derivative of the first mask work, so that there is independent creation? Does the term mean here what it means in copyright law (per Feist)?
What happens if a computer program does the chip layout? For example, an automatically routed gate array. Is the customization (“personalization”) of a gate array protectable under the SCPA? See 37 C.F.R. § 214.4(c)(2). What if it is customized by means a computer program? What if that customized gate array is subjected to reverse engineering and a new customization is prepared using the same computer program? A different computer program?
Does copyright law require originality for the defense of fair use to apply? Explain the reason for that. Compare the SCPA and its purposes.
Does the INS case have anything to do with this? Is the SCPA a legislative enactment of INS for the semiconductor industry?
4. How would you adapt § 906(a) to software? Would that be a good or bad idea?
5. Comment on the following excerpt from a legislative proposal, found in a recent issue (March 1996) of AIPLA Q.J. The proposal is directed to computer algorithms and other computer–related subject matter on 17 U.S.C. § 102(b)'s index prohibitorum.
(e) Research, Reverse Engineering.—It is not infringement for a person to use, or traffic in an embodiment of, protected subject matter solely for the purpose of teaching, analyzing, or evaluating the concepts or techniques embodied in the protected subject matter. It is not infringement to utilize the results of such analysis or evaluation to prepare a further computer program or other product unless trafficking in the program or product would be infringement even without occurrence of the preceding analysis or evaluation of the subject matter. It is not infringement to traffic in such an otherwise noninfringing further computer program or other product.
How is it the same and how is it different from the SCPA's reverse engineering provision? Are there any reasons why differences should be made for legislation governing software reverse engineering, such as different interests or policies at stake?
6. There is not much case law under the SCPA. Brooktree Corp. v. Advanced Micro Devices, Inc., 977 F.2d 1555 (Fed. Cir. 1992), is the only appellate decision. In that case, defendant AMD relied on a reverse engineering defense but the jury ruled against it. The Federal Circuit affirmed, on the ground that the jury had substantial evidence before it on which it could reasonably have found unprivileged copying rather than reverse engineering, for example, because AMD's design was not original.
AMD presented evidence showing that its chip had improved functionality over plaintiff's. (It was smaller and faster.) Plaintiff put in evidence that this was the result of a later technology permitting closer spacing of elements, not a result of a different design layout. The Federal Circuit said that the matter was fact-dependent and therefore up to the jury.
7. A symposium on issues under the SCPA, including several articles addressing the reverse engineering defense, is in the December 1985 issue of the Minnesota Law Review, beginning at 70 U. Minn. L. Rev. 263.
Comment on the following statement by the editors of the symposium issue of the law review (id. at 266):
[R]everse engineering [under the SCPA] strikes a balance between the exclusive reproduction right of an owner and the public interest in innovation, by imposing the condition on the copyist that an improved, or at least original, chip product must result. The concept of reverse engineering is thus a significant departure from the traditional [copyright] calculus of fair use criteria because the SCPA approach expressly balances a limitation on an exclusive ownership right to the extent that the copyist has contributed to the public interest.
Link to chapter 7 - Patent Protection of Software
Return to Table of Contents