COMPUTER LAW, Vol. 3, Rel. 67

The district court’s suggestion that Accolade could have avoided infringement by using a “clean room” technique is also “clearly erroneous.”289 A clean room procedure would not have avoided the need for disassembly here, because disassembly was necessary to discover the functional specifications for a Genesis compatible game.

Accordingly, disassembly was necessary to understand the functional requirements for Genesis compatibility. The interface procedures for the Genesis console are publicly distributed only in object code form. Thus, to understand these requirements it is necessary to decompile that object code. Decompilation necessarily entails copying. “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.”290 Accordingly the court concluded that the second statutory factor (nature of the copyrighted work) weighed in favor of Accolade.291 On the third factor (amount and substantiality of the portion used in relation to the work as a whole), Accolade disassembled the entire program, which weighs against Accolade, nonetheless, does not preclude a finding of fair use.

Summing up, the court concluded that the first, second and fourth statutory fair use factors favored Accolade, while the third favored Sega (and only slightly). Accordingly, the court held there was fair use. In determining whether a challenged use of copyrighted material is fair, the ultimate aim is to stimulate artistic creativity. The fact that programs are publicly distributed in object form often precludes public access to the embodied ideas and functional concepts, thus conferring to the copyright owner an effective monopoly over idea and concept. That defeats the fundamental purpose of the Copyright Act—to encourage production of original works by protecting the expressive elements, while leaving the ideas, facts, and functional concepts in the public domain for others to build on. “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.”292

The court moved on to discuss the trademark claim. This claim differed from the traditional “passing off” claim, and the parties agreed that there was both misuse of a mark and that confusion may result. The issue was which party should be held responsible. Sega adopted a security system governing access to its Genesis III console that displayed its mark and message whenever initialization code for the security system was utilized, whether the program was manufactured by Sega or a competitor. Accolade discovered how to gain access to the Genesis III console through the initialization code, and used that code, even though that use triggered the display of Sega’s trademark. Accordingly, Accolade’s inclusion of the TMSS initialization code had an effect beneficial to neither party.293

Because the TMSS regulates access to the Genesis III console, and because the record contained no indication of awareness of a feasible alternate method of access, the court of appeals held that Sega was primarily reasonable for the resulting confusion. Although the district court held that Accolade was responsible, the court of appeals noted that Accolade had no desire to display the Sega message, and used the TMSS initialization code only to gain access to the Genesis III console, being unaware of any other method for doing so. The false labelling was a result of the deliberate decision by Sega to include in the Genesis III console a device that would limit general access and cause false labelling. The decision to use the Sega mark as an essential element of the functional device that regulates access, and to cause the mark and message to be displayed when that device was triggered, compelled the court to place primary responsibility on Sega. Whatever Sega’s intent, it is the effect of the message that matters.

Sega argued that TMSS did not prevent competitors from developing Genesis III compatible cartridges that do not trigger a display of the Sega mark and message. Accolade disagreed, contending that the TMSS initialization is a functional feature that must be included in a program for it to operate on the Genesis III console. The court of appeals concluded that Sega did not meet its burden of establishing nonfunctionality.

Based on the Nagashima declaration, the district had court concluded that the TMSS initialization was not a necessary component, and that Accolade could have created a game cartridge lacking the initialization that would operate on the Genesis III console or could have programmed its games so that the false message would not be displayed. Either modification could have been accomplished at minimum expense.

The issue here is whether the TMSS initialization code is a functional feature of a Genesis compatible game. Whether a product feature is functional is a question of fact, but, determination of the correct legal standard for functionality is a question of law. The burden of proving non-functionality is on Sega. Here, the district court’s finding of non-functionality was based on its use of an incorrect legal standard. The record supports only one conclusion: the TMSS initialization code is functional.294 Part of the benefit the consumer desires is compatibility with the Genesis III console. The code provides that compatibility.

Sega argued that the Nagashima declaration established the actual existence of technically and commercially feasible alternative methods. But, Nagashima was familiar with the TMSS system, so that his affidavit established only that an individual familiar with the operation of the TMSS could discover a way to engineer around it. It does not establish that a competitor without such knowledge could do so. Nor is there evidence of any public or industry awareness of any alternative method for gaining access. Accordingly, the district court erred in concluding that the Nagashima declaration was sufficient to establish non-functionality.

Because the TMSS regulates access to the console, and because no way of access without the TMSS is known to competitors, there is no basis for finding non-functionality. Other than Nagashima declaration, the only evidence relating to Accolade’s ability to gain access (other than through the TMSS) is the affidavit of Miller, an Accolade employee who stated that Accolade’s software engineers were not able to discover such a method.

Sega contended that even if the code was functional, a copier was obligated to take the most effective measures reasonable to eliminate confusion. But, the court found that the measures adopted by Accolade satisfied this standard. Accordingly, the court reversed the district court’s grant of a preliminary injunction based on the trademark claim.

Another case that permitted reverse engineering was DSC v. DGI, where a district court held that it was fair use for a software developer to disassemble the code of a competitor in order to learn the functions and ideas embedded in that software.294.1 Defendant lawfully obtained a copy of plaintiff’s firmware and reverse engineered the code. The parties stipulated that defendant’s final product did not infringe. Noting that there was no Fifth Circuit authority on point, the court followed the Ninth Circuit in concluding that the copyright applied equally to intermediate copies. Inasmuch as the parties stipulated that defendant’s disassembled code was identical to plaintiff’s code, the court concluded there was infringement unless defendant’s fair use defense was valid.

Defendant’s purpose in disassembling was to discover the functionalities and ideas embedded in plaintiff’s firmware, so as to create defendant’s own firmware for use on a microprocessor that was compatible with plaintiff’s. One of defendant’s engineers disassembled the code and transmitted a copy of the disassembled code to another of defendant’s engineers. The second engineer created a flowchart from the code, and from that, then developed defendant’s firmware.

The court again followed the Ninth Circuit in determining that “when good reason exists for studying or examining the unprotected aspects of a copyrighted program, disassembly for the purpose of study or examination of the disassembled program constitutes fair use.”294.2 Looking to the four nonexclusive fair use factors, the court noted that on the first (character of the use), the copy was intermediate and its commercial exploitation was indirect or derivative. The only direct use was to study the functional requirements so as to manufacture a compatible product.

On the second fair use factor (nature of the work), the court found that firmware, because it has functional aspects that cannot be examined without copying, merited a lower degree of copyright protection than more traditional literary works. On the third factor (amount used), the court found that the entire work had been copied (but noted that this was the only way in which the defendant could learn the functionality). On the fourth factor (affect on market), plaintiff had conceded that defendant’s product would not interest actual or potential users of plaintiff’s product because defendant’s product lacked diagnostic, maintenance and serial port communications functions.

In Sony v. Connectix,294.4 plaintiff Sony marketed a “PlayStation” console, a small computer with hand controls that connects to a TV monitor and plays games, inserted in the form of CD’s into the PlayStation console. Sony owned the copyright on the basic input/output system (“BIOS”), the software that operates the PlayStation console. Defendant produced a software package called “Virtual Game Station,” that emulated on a personal computer the functioning of the PlayStation console. Thus, the owner of a personal computer, armed with a copy of the Virtual Game Station software, could play PlayStation console games on that personal computer. The Virtual Game Station software contained no Sony copyrighted material. In the process of producing the Virtual Game Station software, defendant repeatedly copied the BIOS software in its reverse engineering, undertaken to determine how the PlayStation console functioned. Sony sued for copyright infringement and the district court, concluding that the intermediate copying was not fair use, had granted an injunction precluding defendant from marketing its software or copying the BIOS code and developing other Virtual Game Station products.

In the PlayStation console, the BIOS was embedded in firmware. The Virtual Game Station software emulated both hardware and firmware. Software engineers designing a program that must be compatible with a copyrighted product frequently must “reverse engineer” that product to understand its functional elements. There are four methods of reverse engineering: reading about the program; observing it in operation by using it on a computer; performing a static examination of the instructions; and performing a dynamic examination of the instructions as the program is run. The first of these is often ineffective because written materials often misdescribe the program; moreover, in this case, such materials were not available. The other three methods required loading the program, which creates a copy. But none of the Sony code was copied into or appeared in defendant’s final product, the Virtual Game Station software. In emulating the PlayStation console hardware, defendant’s engineers purchased a PlayStation console and copied the code from the Sony BIOS firmware into the RAM of their computers. These engineers observed the functioning of the Sony BIOS in connection with the Virtual Game Station hardware emulation software as it was developed. In doing so defendant made additional copies of the Sony BIOS.

After defendant completed development of its Virtual Game Station software Sony filed a complaint. As noted above, the district court preliminarily enjoined defendant from copying or using the Sony BIOS in further development of the Virtual Game Station, and from marketing it. It also impounded defendant’s copies of the BIOS and of works based upon or incorporating the BIOS.

Defendant alleged that its copying constituted fair use. Because of the nature of software, the unprotected ideas and functions of the code are often undiscoverable in the absence of activity that may require copying it. In Sega the Ninth Circuit had held that, where disassembly is the only way to gain access to the ideas and functional elements embodied in code, and where there is a legitimate reason for seeking such access, disassembly is fair use. Sega had recognized that intermediate copying could constitute infringement, even where the end product did not itself contain copyrighted material. But such copying would be fair use if necessary to gain access to the functional elements.

The appellate court in Connectix examined the four statutory fair use factors. In connection with the nature of the copyrighted work, the appellate court recognized that some works are closer to the core of intended copyright protection than others. The BIOS lies at a distance from the core because it contains unprotected aspects that cannot be examined without copying. Consequently it merits a lower degree of protection than more traditional literary works. The only way to gain access to the functional elements of BIOS was through reverse engineering that required copying it onto a computer. The methods by which defendant reversed engineered BIOS were necessary to gain access to the functional elements. The district court had found that defendant’s copying and use of the BIOS to develop its own software exceeded the scope of Sega. The appellate court disagreed, finding that defendant’s use here did not exceed what defendant in Sega had done.

Sony argued that defendant here went beyond what was necessary because defendant repeatedly observed the BIOS in an emulated environment, making many copies. Sony contended these intermediate copies could not have been “necessary” because defendant’s engineers could have disassembled the BIOS and written their own BIOS, which they then could have used to develop the hardware emulation software. The appellate court disagreed noting that the “necessity” addressed in Sega was the necessity of the method (disassembly), not the necessity of the number of times that method was applied.

With regard to amount and substantiality of the portion used, the entirety of the BIOS was copied. However, in the case of intermediate infringement when the final product does not itself contain infringing material, this factor is of very little weight. As to the purpose and character of use, the question is whether and to what extent the new work is transformative. The Virtual Game Station software is modestly transformative. It creates a new platform (the personal computer) on which consumers can play games designed for the PlayStation console. This affords the opportunity for game play in new environments. Moreover, the Virtual Game Station software is a wholly new product, even though it may have a similarity of use and function with the Sony PlayStation console.

Defendant’s commercial use was an intermediate one, and thus was only indirect or derivative. Moreover, it reverse engineered the BIOS to produce a product that would be compatible with games designed for the PlayStation console. This is a legitimate purpose. Accordingly, this factor favors defendant. Although the district court had ruled that the Virtual Game Station software was not transformative on the rationale that a computer screen and TV screen are interchangeable, and defendant’s product therefore merely “supplants” plaintiff’s console, the appellate court ruled that defendant’s product was transformative and does not merely supplant defendant’s console. The district court had failed to consider the expressive nature of defendant’s software.

With respect to the effect of the use on the potential market, the appellate court noted that a transformative work is less likely to have substantial adverse impact on the potential market of the original work. Because defendant’s work is transformative and does not merely supplant plaintiff’s, the defendant’s work is a legitimate competitor in the market for platforms on which plaintiff’s games can be played. For that reason, some economic loss by plaintiff did not compel a finding of no fair use.

Accordingly, the appellate court found that three of the four statutory fair use factors favored defendant, and the one factor in favor of plaintiff was of little weight. The appellate court held that defendant’s reverse engineering of the BIOS extracted from plaintiff’s product purchased by defendant was a protected fair use. Other intermediate copies of plaintiff’s BIOS made by defendant, if they infringed, did not justify injunctive relief.

Ticketmaster v. Tickets.com294.5 dealt with two competitors who marketed tickets through their websites. Defendant used a “spider” or “crawler” to review the internal web pages of plaintiff’s website. This software extracted information in electronic form and loaded it into defendant’s RAM, where it remained for 10 to 15 seconds. Defendant then extracted from that information certain facts pertinent to events (e.g., event, date, time, ticket prices and other facts not subject as such to copyright), and discarded the remainder (plaintiff’s identification, logos, ads, and other information protected by copyright). The retained information was then organized in defendant’s format and displayed on defendant’s website. One issue in the case was whether storing the information taken from plaintiff’s pages (some of which was subject to copyright) for several seconds constituted copyright infringement. “Is this momentary resting of the electronic symbols from which [plaintiff’s] web page could be (but is not) constructed fair use where the purpose is to obtain non-protected facts?”294.6 The court answered in the affirmative, relying on Sony and Sega for the proposition that “temporarily” copying software to obtain non-protected “material” where that is the only way to get that material is fair use. Plaintiff argued, however, that here this was not the only way to obtain the unprotected information. Nevertheless, the court held that fair use applied here.

The court seemed impressed by the fact that the purpose was to extract unprotected information, and that the storage of the expression subject to copyright was of short duration. The court saw no public policy that would be served by prohibiting the conduct here. “In temporarily downloading [plaintiff’s] event pages to its RAM through the use of spiders, [defendant] was not exploiting [plaintiff’s] creative labors in any way; its spiders gathered copyrightable and non-copyrightable information alike but then immediately discarded the copyrighted material. It is unlikely that the spiders could have been programmed to take only the factual information from the [plaintiff’s] web pages without initially downloading the entire page.”294.7 The court went through a four-factor fair use analysis. It found that defendant’s site was commercial, tending to weigh against fair use; defendant’s data was only slightly transformative. The court seemed to believe that the nature of defendant’s work was analogous to the process of copying that the Sony v. Connectix294.8 court had condoned. The amount and substantiality of plaintiff’s work used was of little weight because the final product incorporated nothing from plaintiff. And the court found that there was no effect on the market value of plaintiff’s work. On balance, the court held there was fair use.

Nautical Solutions v. Boats.com294.9 presented another website interesting situation in which the fair use doctrine once more came to the assistance of an accused infringer. Each of the parties maintained a website on which yacht brokers advertised their products. For several months, plaintiff used an Internet spider (the “Boat Rover” program) to visit defendant’s site, extract facts regarding the listings contained thereon,294.10 and index those facts in a database searchable by plaintiff’s users. Plaintiff sought a declaration that its conduct did not constitute an infringement of defendant’s copyrights.294.11 The manner in which the spider functioned involved copying the hypertext markup language (the “html”), extracting the facts, storing them in a database, and immediately discarding the html. Defendant contended that this extraction of facts constituted copyright infringement. However, the court held that the “momentary copying of Yachtworld’s public pages in order to extract from yacht listings facts unprotected by copyright law constitutes a fair use,” and thus was not infringement. The court noted that, although plaintiff’s venture was commercial, the record contained no evidence as to the effect of plaintiff’s activity on the “potential market for or value of” defendant’s work. The court also ruled that, because plaintiff’s final product—the database—contained no infringing material, the “amount and substantiality of the portion use” was of little weight. The court found persuasive the rationale of Ticketmaster.294.12

In Galoob v. Nintendo,295 a court once more dealt with the issue of whether an add-on device, whose sole purpose was for use with a video game machine and its cartridge, infringed the copyright in the audiovisual works created by the video game. The device in question (the “Game Genie”) was sold by Galoob for personal use by consumers. Audiovisual work copyright owner Nintendo alleged that the consumer’s use of the device created an unauthorized derivative work, so that Galoob’s sale constituted contributory infringement.

A number of devices were marketed by various distributors for the purpose of interacting with Nintendo game play so as to speed it up, skip levels of play, etc. Some were marketed with the authorization of Nintendo; the Game Genie device was not. The Game Genie device had no use other than in connection with Nintendo’s game machine and Nintendo-compatible cartridges. The Game Genie device was inserted into the game machine, and the game cartridge was then inserted in to the Game Genie device. The Game Genie device altered play by permitting the user to change certain codes incorporated in the game program. The Game Genie device was used only at home by consumers (and not in any arcade), did not create a separate copy of the original game, made no permanent changes to that game, and could not be used except when attached to the original game.

The court held that no derivative work was created. Looking to the statutory definition of “derivative work,”296 the court reasoned that a derivative work must incorporate a preexisting work in a concrete or permanent form. The court reached this conclusion by virtue of the fact that the examples set forth in the statute did so, and the legislative history stated that such was required.297 The court endorsed the lower court’s finding that no independent work was created.298 Because defendant’s device could not independently produce a display, the court disagreed with Nintendo’s contention that the “Game Genie’s displays are as fixed in the hardware and software used to create them as Nintendo’s original displays.”299

Nintendo relied on Midway301 for the proposition that the speeded up display was a derivative work. but the court distinguished Midway because defendant’s chip there was not used in connection with plaintiff’s, but rather copied and replaced plaintiff’s; moreover, defendant’s product there supplanted sales of plaintiff’s. Here defendant’s product does not embody plaintiff’s, nor does it supplant demand for plaintiff’s. Stretching the definition of derivative work to embrace the Game Genie device would fail to protect society’s interest in the free flow of ideas, and would chill innovation.302 “The Game Genie is useless by itself, it can only enhance, and cannot duplicate or recast, a Nintendo game’s output. It does not contain or produce a Nintendo game’s output in some concrete or permanent form, nor does it supplant demand for Nintendo game cartridges. Such innovations rarely will constitute infringing derivative works under the Copyright Act.”303

The court also held that defendant engaged in fair use304 of plaintiff’s work, so that even if defendant’s displays constituted derivative works, defendant was not liable. A major issue on this point was whether the focus of the fair use inquiry should be defendant, or the users of defendant’s products (i.e., the would-be the direct infringers). The proper focus is on the direct infringers. The district court did not err when it concluded that the home users of defendant’s device engaged in non-profit noncommercial activity. Accordingly, their use of plaintiff’s work to produce derivative works would presumptively be fair use.

The district court also properly concluded that the published nature of plaintiff’s work supports the notion of fair use. And the fact that any derivative works created by defendant’ device are comprised mostly of plaintiff’s displays does not render fair use inapplicable; in Sony305 the direct “infringers” copied the entirety of plaintiff’s works but the court found fair use. Further, the fourth fair use factor (effect of use on plaintiff’s potential market) is the most important factor. Here, the district court did not err in concluding that plaintiff failed to show any harm or likelihood of harm.306 The court found that plaintiff’s strongest argument against fair use was the fact that the work was not factual. Nevertheless, this was not dispositive.

In MTI v. Cams,307 plaintiff MTI created software for estimating the cost of machining manufactured parts. It secured a copyright registration for the program as a literary work, and a separate registration for selected screen displays as a literary work.308 Defendants functioned as plaintiff’s marketing representative, and saw the screen displays on demonstration and in users’ manuals. Defendants then developed the allegedly infringing software, which they marketed at about $2,000, whereas plaintiff’s package was marketed for $20,000.

The court was of the opinion that copyright in a program extends protection to its screen display,309 but did not believe that merely copying the display infringed that copyright.310 The court decided that the registration of the program accomplished the legal fiction of two interrelated registrations, so that copying the screen display would constitute infringement.311 This “recognizes that a computer program and its screen display are, for copyright purposes, fundamentally distinct.”312 Plaintiff claimed that copyright protected both the “external” aspects (the flow and sequence of screens) and the “internal” aspects (each screen).

Defendants contended that plaintiff’s displays contained no copyrightable subject matter, for a number of reasons. They argued that the screen displays were dictated by functional considerations that limited expression to screens similar to these; also that plaintiff’s displays were uncopyrightable forms conveying no information; and further, that there was no originality in the displays. Looking to the facts, the court determined that the flow and sequencing of screens was not dictated by functional considerations.

Plaintiff claimed that the formatting style313 of the screens was protected, but the court disagreed, finding that there were only a limited number of ways to handle that matter. Plaintiff also claimed protection for its internal methods of navigating the screen.314 The court found this also unprotectable, finding there were so few ways of simply moving about the screen that such protection would amount to unwarranted315 monopolization of easy-to-use internal navigation conventions.

Nevertheless, the court found that one aspect of plaintiff’s screens constituted copyrightable expression: the method of identifying the operation/department, tooling, and tooling material used.316 The court then went on to examine non-uniform aspects of several screens. It found unprotectable the two-column listing of shops through which a part might go.317 But there was sufficient originality to a screen seeking significant user input on job identification,318 because that screen and its components were not severely limited by functional considerations. That screen was also held to contain information, thus avoiding the proscription against copyright protection for blank forms containing none.319

Accordingly, the court held that in copying the sequence and flow, the convention for providing status information, and job information expressions, defendants infringed the screen display copyrights. Moreover, that infringement went beyond mere infringement of the copyright in the eleven screens deposited. Rather, defendants infringed also the copyright of the screen displays subsumed within the registration of the program.

There are also a number of videogame cases that can be viewed as having look and feel aspects.320 Most of these were decided in the early 1980’s during the height of the videogame craze. In some of these cases the copyright registration was on the audio-visual display, whereas in others it was on the underlying software. In a few cases, both types of registration existed. In most of these cases, plaintiffs succeeded in excluding defendants from using the work. Perhaps because these suits were directed at products viewed as toys, their significance may not be fully appreciated in the context of commercial software packages. And indeed, the focus was different, as those screen displays were regarded as audio-visual works, rather than textual works, such as spreadsheet screens. Nevertheless, the rationale underlying some of these opinions may have application in the present environment.

Another opinion of interest is Pearl v. Competition,321 which contains language that smacks of all three “waves.”322 It is difficult to determine from the opinion whether the decision went off on the copying of code, of structure or of look and feel. Competition Electronics developed a microprocessor-controlled shot timer device for competitive pistol shooting. It measured and displayed the number of shots fired in a pre-set time period (“par time”), the total number of shots fired, and the interval between shots (“shot review”). The device was controlled by pushing buttons. Competition claimed that Pearl’s competitive device infringed Competition’s copyright as to the shot review subroutine and the par time entry subroutine, both of which had competitive importance. There was similarity (though not identity) in the button sequences used by the two devices to display information.

The copyright owner’s first expert had been requested, without being shown a device, to design software meeting certain specifications: it would record shots, have par time and shot review capabilities, and be worn on a belt. He developed a software design with shot review and par time subroutines that varied from that of both devices.323 His hardware configuration also differed from both devices.

The copyright owner’s second expert had been asked to analyze similarities and dissimilarities among the three sets of shot review and par time entry subroutines, as well as a fourth set used in a product offered by a third party. He examined source code used in the parties’ devices, and developed flowcharts of the code “based upon the way the products operated.” He testified to identity of “systems level design” of plaintiff’s and defendant’s software in the case of each of the two subroutines.324 This conclusion was based on source codes (which were in different languages) and the “external operations of the products” and the witness’ analysis of “how the system-level software would have to function.”325 He further testified (apparently based upon the flowcharts he constructed) that the third party subroutines were quite different from those of the parties’ software, in part because the third party device used a keyboard, instead of buttons, for data entry. He also found substantial differences between the first expert’s device and those of the parties, which differences occasioned different software.

The accused infringer’s first expert testified that, while the source code of the parties’ subroutines was different because of differences in hardware and programmers, it was possible for the software to have been copied at a different level. The accused infringer’s second expert testified to differences at the code level and module level, but to similarity at the system level design.

The court credited the testimony of the copyright owner’s witnesses, and held that the similarities between the parties’ software was not “constrained by product function.” It found substantial similarity in the subroutines, and was unconvinced this was the result of independent effort. The court assumed that, because Pearl had access to Competition’s device, Pearl had access to the copyrighted work, although there is no discussion of whether (or how) Pearl obtained source code. The court held that “the Pearl Systems device captures the ‘total concept and feel’ of the Competition Electronics subroutines.”326

The court then went on to dispense of any problem that might be posed by the idea/expression dichotomy: “The par time entry subroutine was designed to provide a method for the user to set a par time. That is the idea … . The subroutines themselves are expressions of those ideas.”327 Another interesting observation of the court was that “the subroutines in both [devices] were nearly identical, and this was reflected by the fact that the subroutines were triggered by the same sequence of buttons. As there was ample testimony that alternative system level designs could have been used to avoid this similarity, we conclude that the idea did not have only one necessary form of expression, but many. The nearly identical structures of the subroutines … were not necessary to the tasks. Consequently, the idea and its expression did not ‘merge.’ ”328 The court cited authority for the proposition that the purpose or function of a utilitarian work is the work’s idea; whatever is not necessary to that purpose or function is expression.

It would appear that there is some confusion in this opinion, although it is difficult to identify its precise nature because the opinion in part is ambiguous as to rationale. Conceptually, it is difficult to assess whether the similarity on which the court focused is that embodied in code, flowchart, or purpose. In large part the confusion stems from not first separating that which is in the public domain, and therefore “fair game,” from that which may be protectable.

In Integral v. People Soft,329 the case turned on the battle of the experts. The parties marketed competing human resource management software (“HRMS”) packages. Plaintiff’s program was for a mainframe, with a license fee of $500,000. A group of plaintiff’s employees left to found defendant company, and to develop a functionally similar package for PC’s. After two years of effort and an expenditure of $1.2 million, it succeeded. Plaintiff alleged copyright infringement and trade secret misappropriation, and sought a preliminary injunction.

Plaintiff’s expert compared the two programs, and reported that there were similarities between the two programs, and a lack of a paper trail showing independent development of defendant’s. He concluded: “it can be reasonably presumed that the Integral system served as a model for the People Soft system,” and that the latter was derived from the former.

In connection with the copyright claim, the court noted that copyright does not protect functionality. Plaintiff conceded there was no copying in the code or user interface, but asserted copying in the “program design.” Specifically, it charged unauthorized copying, and making of a derivative work, in the structure, sequence and organization of the program portions that accomplished three particular general functions.330 Concentrating first on the copying aspect, the court found a reasonable jury could find access. Regarding similarity, plaintiff’s expert testified to similarities, and defendant’s testified that, using the methodology of plaintiff’s expert, the screen displays in an independent third party package was as similar to plaintiff’s as was defendant’s. The experts also gave divergent testimony on whether certain design similarities dealt with discretionary characteristics. Plaintiff’s case was not improved by the disclosure that the original conclusion of its expert’s report—that overall the designs were dissimilar—was stricken in the final version.331

The court concluded that substantial similarity had not been shown. The court faulted the main support for plaintiff’s charges—the expert’s report—as seriously flawed. The report intentionally ignored differences between the programs, and in a manner that was masked. Moreover, with respect to the particular type of software involved, defendant’s expert had more experience than plaintiff’s. Further, the court found that the defendant’s paper trail was consistent with that of other systems developed in the human resources management industry.

An issue that has been dealt with in surprisingly few cases is whether, for copyright purposes, a program is published when licensed under a confidentiality agreement. That issue was decided in PRC v. National, a 1991 district court decision, which held that it is not published.332 In that case the software in question was so licensed, and in at least one case source code was provided pursuant to the license agreement. The court was called upon to determine the issue of publication in order to ascertain whether 17 U.S.C. Section 410(c) applied.333 The court held: “The work was not made available to the general public without restriction. The only ‘giving away’ occurred under tightly restricted terms, e.g., confidentiality and restrictions on sublicensure to a competitor, and on release of source codes. This is not publication.”

There are also a number of videogame cases that can be viewed as having look and feel aspects.339 Most of these were decided in the early 1980’s, during the height of the videogame craze. In some of these cases the copyright registration was on the audio-visual display, whereas in others it was on the underlying software. In a few cases, both types of registration existed. In most of these cases, plaintiffs succeeded in excluding defendants from using the work. Perhaps because these suits were directed at products viewed as toys, their significance may not be fully appreciated in the context of commercial software packages. And indeed, the focus was different, as those screen displays were regarded as audio-visual works, rather than textual works, such as spreadsheet screens. Nevertheless, the rationale underlying some of these opinions may have application in the present environment.

In the second of the suits by a giant software vendor to be decided, Ashton-Tate v. Fox,340 the court issued a terse order dismissing the copyright claims with prejudice based on plaintiff’s inequitable conduct in its dealings with the Copyright Office: “The Court finds that Ashton-Tate, when it filed its original applications for copyright, repeatedly failed to disclose material information to the United States Copyright Office—that the dBase line of computer software was derived from JPLDIS, a public domain computer software program developed by Jet Propulsion Laboratory, and that dBase III was derived from dBase II. The Court further finds that Ashton-Tate’s repeated failure to disclose such material information was done knowingly and with an intent to deceive. The Court, therefore, finds that Ashton-Tate’s copyrights on its dBase line of computer software programs are invalid as a result of its inequitable conduct.”

Shortly after this ruling came down, a class action was filed against Ashton-Tate on behalf of all users of dBASE II and dBASE III software.341 Plaintiffs alleged that Ashton-Tate, knowing that it did not possess a valid copyright in the software, held itself out as holder of a valid copyright in the software, and used “fraudulent” copyright claims to induce users into license agreements that placed restrictions on them with regard to, and required payment for the use of, what was public domain matter. It has been reported that, even before this class action was filed, Ashton-Tate was plaintiff in some nine copyright infringement actions wherein defendants were likely to asset as a defense the Fox ruling.342

The trial court originally denied Ashton-Tate’s motion for reconsideration. Ashton-Tate sought leave from the Ninth Circuit to appeal the trial court’s ruling on inequitable conduct. The Register of Copyrights filed in the trial court a Declaration in which he noted that: no act of the Copyright Office is necessary to secure copyright; the Copyright Office examination does not include a search for prior art; if Ashton-Tate had disclosed in the application for the original copyright certificates information supplied by Ashton-Tate in certain supplementary certificates subsequently obtained by Ashton-Tate, the original certificates would nevertheless have issued; and where a copyrightable work is merely influenced or inspired by a pre-existing work, the Copyright Office does not expect or require that such pre-existing work be disclosed in the application. A few days after the filing, Fox Software filed a Declaration of the Register in which he stated that if the Copyright Office had originally known the facts stated in Ashton-Tate’s supplemental registrations, the Office would have required Ashton-Tate to amend its registrations.

On April 18, 1991, the district court filed a succinct order (1) granting Ashton-Tate’s motion for reconsideration of Fox Software’s summary judgment motion based on the allegation of inequitable conduct; (2) rescinding the order of December 11, 1990 that had granted that summary judgment motion; and (3) denying that summary judgment.

In 1991 Borland International executed an agreement to acquire Ashton-Tate Corp. Because both companies were large players in the market for relational database management systems, the U.S. Department of Justice Antitrust Division became involved, filing suit under Section 7 of the Clayton Act.343 That suit was terminated by consent decree wherein the two companies agreed to restrictions,344 including a restriction against certain assertions of copyright infringement.345

Then in March of 1988, in a suit that has attracted a good deal of ink, Apple sued Microsoft and Hewlett-Packard for copyright infringement on the basis of allegedly improperly copied data organization and screen displays.346 The copyrights were in audiovisual works consisting of the visual displays, and in literary works comprising the programs that embody the audiovisual works. Specifically, Apple alleged that Hewlett-Packard’s “New Wave” program and Microsoft’s “Windows” version 2.3 infringed. Although opinions have been handed down in this litigation, they have dealt with the license defense interposed by Microsoft, rather than with the copyright issues.

In March 1991 the trial court handed down a ruling on certain issues involved in Apple Computer, Inc. v. Microsoft Corporation.347 Plaintiff Apple claimed that defendant Microsoft’s “Windows” operating system (“O/S”), and defendant Hewlett-Packard’s (“H-P”) NewWave application infringed copyrights protecting the visual display of Apple’s Macintosh user interface. Specifically, Apple alleged that NewWave and Windows 2.03 infringed Apple’s copyrights in audiovisual works; and that Microsoft’s licensing of H-P comprised contributory infringement of those copyrights. In 1989 the court had determined that the 1985 agreement was not a complete defense, but had authorized Microsoft to use the visual displays in Windows 1.0 and the application programs; and that the displays in Windows 2.03 were in 1.0 and those application programs except for those ten displays relating to overlapping windows and icon manipulation.348

The Macintosh user interface consists of windows, icons, pulldown menus and other images on the display screen, and had quickly become quite popular with users, who found it unusually easy to manipulate. In 1985 Apple and Microsoft agreed that the displays in Windows 1.0 and certain application programs were derivative works of works of Apple; that Microsoft was authorized to use and sublicense those derivative works; that Microsoft would not offer a new program similar to its Excel prior to 1986; and that Apple was authorized to use any display created by Microsoft for five years as part of Windows. Microsoft then authorized H-P to use Windows in developing NewWave.

Each party then moved for partial summary judgment. Microsoft claimed that seven of those displays were licensed by the 1985 agreement, and that none of the ten comprised expression protectable under an Apple copyright. H-P claimed that each of the remaining items on Apple’s list of similarities as to NewWave as not original and was not protectable expression; that the scope of protection would be so narrow that only identical copying could infringe; and certain of the items were licensed. Apple claimed that the copyrights were valid and certain H-P affirmative defenses should be dismissed. In making its determinations, the court excluded licensed displays before examining substantial similarity.

In its motion, Microsoft argued that any display in Windows 1.0 and the five named application programs were licensed under the 1985 agreement, and that the seven displays were in one of those application programs. In opposition, Apple contended that each of the Microsoft application programs for the Macintosh owed its appearance to the Macintosh O/S, and that the displays were generated by the interaction between the O/S and the application program. Since not produced by the application programs alone, these displays were not, according to Apple, licensed.

The court noted (1) the California rule that ambiguous terminology is to be interpreted against the drafter (Microsoft), (2) the Ninth Circuit rule that copyright licenses are to be construed narrowly, and (3) that the purpose of the 1985 agreement was to settle a dispute involving Windows 1.0. It held that this agreement did not license displays generated by the interaction of application software with the O/S.349 Each of the seven displays owed its appearance to the O/S, and therefore was not licensed. Microsoft’s motion was denied.

In its motion Apple wanted a declaration that its audiovisual copyrights were valid, and dismissal of defendants’ affirmative defenses.350 For the most part the court determined there were factual issues so that summary judgment was inappropriate. However, the court dealt summarily with H-P’s defense of Apple’s fraud on the Copyright Office, and the originality of Apple’s displays. On the first of these, H-P contended that Apple’s failure to disclose to the Office that its works were based on preexisting works invalidates the copyrights. H-P offered evidence that the displays were strongly influenced by certain Xerox programs. However, the court saw that as merely a permissible borrowing of ideas. Further, H-P provided the court with no evidence of an intent to deceive the Office, a necessary element of the defense.351

The court saw as meritless H-P’s contention that Apple’s works were derivative, because there was no evidence that Apple infringed any Xerox copyright or that Apple substantially copied. Because there was no evidence of intent to commit fraud, and because it found Apple’s works were not derivative, the court held H-P had not shown fraud.

On the originality point, H-P contended that the works lacked originality. The court held that defendants misconstrued the nature of this requirement. The work need only be independently created, and embody a very modest amount of intellectual labor. The court saw no evidence of copying of expression (in contrast to idea). Accordingly, the court (1) granted Apples’ motion for partial summary judgment that Apple perpetrated no fraud, and Apple’s works were original; and (2) dismissed H-P’s affirmative defenses of fraud and lack of originality.

In an interesting suit, Xerox v. Apple,352 plaintiff sued on a number of claims predicated on the theory that Xerox had developed the Star workstation program as the first program to use a “mouse” and use fanciful images in a screen display, and that Apple had appropriated portions of this work to itself. The thrust of the suit was that Apple’s Macintosh user interface was based on Xerox’ Star and SmallTalk software, developed in the late 1970’s. Interestingly, copyright infringement was not among the claims alleged by Xerox.

Declaratory Judgment of Sole Ownership: Xerox sought a judgment acknowledging its sole ownership of the Star work. However, the court found there was no basis for declaratory relief, because Xerox had no real and reasonable apprehension of being sued by Apple. Apple had simply done nothing that could be construed as a threat of imminent litigation against Xerox.353

Order Striking Apple’s Copyright Registrations: In two claims, Xerox sought an order directing the copyright Office to strike registrations that allegedly violated several sections of the copyright law.354 Xerox alleged that two Apple copyrighted works unlawfully copied and modified pre-existing Xerox works, and that the copyright registrations Apple received intentionally failed to identify the works as derivative works of pre-existing Xerox works. The court held again that there was no basis for declaratory relief because no actual controversy existed. Further, the court held that the Copyright Act did not give the court the power to afford such relief, and that it lacked this power.

State Unfair Competition and Unfair Business Practices: Xerox alleged that Apple’s conduct constituted unfair competition and unfair business practice under California law. The court held that these state claims were pre-empted by the Copyright Act. Xerox’ claim here contained no “extra element” beyond that required for copyright infringement. Xerox argued that the confusion allegedly caused by Apple as to title to the work constituted the extra element, but the court disagreed. To state a claim under the state statute, Xerox had to show likelihood of confusion as to source, i.e., passing off. This is not the type of confusion alleged by Xerox (misappropriation). Such misappropriation must be redressed, if at all, under copyright law.

Federal Unfair Competition: Xerox sued for false designation of origin and false description under Section 43(a) of the Lanham Act. Xerox made no express allegations that Apple’s registrations were false or that Apple engaged in false advertising. Rather, Xerox’ position was that because Apple’s registration were allegedly improper, Apple’s use of the copyrights violated the Lanham Act. The court saw this as too far removed from the consumer to comprise a Lanham Act violation; it could not conclude that Apple’s alleged conduct had resulted in a likelihood of confusion as to the source of the copyrighted works.

In Secure v. Time and Space,355 both parties manufactured fax machines for secure transmission of sensitive documents. To function in connection with machines already in the market, defendant’s machine had to conform to a “handshake protocol,” specifying content, order and timing of the digital signals transmitted between machines. When a correct sequence of code, embodying correct timing, is received, compatible machines recognize each other and can communicate. However, there are a limited number of opportunities, within the protocol, to vary the signals. Plaintiff varied the content and timing of certain signals within these constraints and claims that these variations comprise both a trade secret and copyright-protected subject matter.

Defendant’s machine was originally not completely compatible with plaintiff’s, so the government lent defendant one of plaintiff’s machines to make necessary adjustments.356 Under California law, the court held there was no trade secret. The court ruled that plaintiff had not taken steps to retain secrecy, for it had sold machines to the government without a reservation of rights.

Plaintiff fared no better with its claim of copyright infringement, as the court ruled that the variations in the protocol contained no originality. The protocol sharply limited permissible variations, allowing insufficient choice and selection to qualify for originality.357 Defendant argued there was no “fixation” of the protocol variations, as required by section 102, and the court found a genuine issue of material fact on this point. Nevertheless, on the basis of lack of originality, the court granted summary judgment in favor of defendant.

In a multifaceted case, Allen-Myland v. IBM,358 defendant IBM counterclaimed that plaintiff AMI infringed the copyright in IBM’s 3090 computer system microcode. There were 42 models of the 3090, with 29,000 different model, memory and feature configurations possible. With each 3090 system it markets, IBM furnishes an archival copy of microcode. IBM published the original version of the microcode in 1985, and registered it with the Copyright Office in 1989. Copyright notice was displayed on copies distributed and on screen displays. Each split or reconfiguration involved altering the microcode, as the microcode for each system configuration was different. IBM supplied microcode to any customer who wanted to reconfigure, at a cost of roughly $1,000.359

AMI provided engineering services to users of IBM computers, and had performed numerous splits and reconfigurations360 of 3090 systems. AMI maintained in its library, and distributed, unauthorized copies of IBM microcode, maintained on five tapes. AMI copied and used 3090 microcode for reconfigurations in three ways: it copied code sent to it by IBM for reconfigurations; it modified tapes in its library to implement configurations that were close to the code in question; and it created “rainbow” microcode by combining portions of microcode in its library for several different systems. AMI asserted a number of defenses to the copyright infringement charge, including: non-copyrightability of microcode; permissible repair; statutory authorization; estoppel; permissible self-help; copyright misuse; and implied license.

Copyrightability—AMI argued that the microcode on one tape comprised no more than a parts list, and lacked sufficient originality to be copyrightable. The court rejected the argument because it found that the five tapes contained a single work, and the tape in question could not be viewed in isolation. The tape in question contained a table that was used in connection with material on the other tapes. AMI also claimed that protecting microcode would go beyond protecting expression and would protect idea because the system could not be reconfigured without making a copy of microcode for the new configuration. The court rejected that assertion because it found that the microcode for the new configuration did not have to be the same as IBM’s.361

Idea/Expression—AMI contended that in designing the system so that it could not be reconfigured without changing microcode, IBM improperly used its copyright to control the ability of owners to use and change their machines. The court disagreed, noting that Apple v. Franklin362 held that where other programs could be written to perform the same function, then the program in question is expression and subject to copyright. The microcode here could have been written in a number of ways. Whether AMI could feasibly have created alternative microcode is irrelevant.

Repair—AMI analogized to the patent repair doctrine, which provides that the owner of a patented product may repair (but not reconstruct) it. The court found the analogy wanting.363 The patent repair doctrine permits replacement or repair of unpatented parts of a patented combination. Here, the component being replaced was itself the subject of copyright.

Statutory Permission—AMI argued that 17 U.S.C. § 117 authorized the copying it had done. But the court held that AMI’s conduct went beyond that permitted by the statute.364 Section 117(1) permits making a copy as an essential step in using the program with a computer, where it is used in no other manner. This did not apply in the instant case because AMI copied to accumulate a library and to make copies of microcode for reconfigured or split systems. Section 117(1) also permits adaptation where necessary to get the program to run on a particular computer. But the court held that was not present here, either, because the adaptations were all made to code supplied for a computer different than the one on which the adapted program was run.

Consent Decree Violation—AMI charged that IBM was estopped from asserting its copyrights for violations of a consent decree entered in an antitrust suit brought by the United States.365 AMI contended that IBM violated provisions that enjoined IBM from (1) controlling modifications to or attachments to computer systems; (2) engaging in lease-only of computer systems because IBM freely provided code to lessees to support reconfiguration, but refused to provide free code to owners); (3) refusing to sell microcode at reasonable, non-discriminatory prices; and (4) refusing to furnish a copy of code at cost to computer system owners. The court held that to establish such an estoppel AMI had to show not only that IBM engaged in conduct prohibited by the decree, but also that IBM caused AMI’s unauthorized copying.

As to reconfiguration, the court was not convinced that IBM brought caused AMI’s infringing activity. Although AMI contended it was impossible to reconfigure without copying and changing the microcode, the evidence was that for roughly $1,000 anyone could quickly obtain from IBM the microcode to support a reconfiguration. AMI made a similar argument regarding microcode for splits, and the court agreed, holding that IBM did impermissibly attempt to control the alteration of computer systems by prohibiting reconfigurations or subjecting them to its control through pricing of the microcode necessary for the splits.366 By pricing microcode for a split to eliminate arbitrage possibilities, IBM violated the consent decree provision enjoining IBM from “prohibiting, or in any way subjecting to IBM approval or control, alterations in or attachments to” a computer.367 IBM failed to establish that it required several hundred thousand dollars per split to recover development costs.368

Feist did indeed hold that copyrightability for a work did not mean every element of it was protected; protection did not extend to elements not original. But, the court distinguished the instant case because tape 2 was a substantial and necessary portion of a single work—the 3090 microcode. This distinguished it from the white pages, which the court found to be a discrete, severable element of a telephone book.

Nevertheless, after concluding it was unnecessary to determine the originality of tape 2, the court went on to assess the originality of tape 2. First, it concluded that tape 2 should not be analyzed under the Feist standard, because that standard related only to compilations of fact, and tape 2 did not qualify as such. Feist held that facts were not original and therefore not protectable by copyright. But tape 2 contained instructions and software tools as well as data. All of this was created by IBM employees, rather than pre-existing matter taken from the public domain. The court analyzed the material on tape 2 and concluded that some of it was original, as having been independently created by IBM, and requiring more than de minimis creativity.

The court saw the idea of the work as a horse race video game where players can bet on the outcome of a prerecorded, actual horse race. The court noted that where expression is inseparable from idea, where it necessarily follows from the idea, or where as a practical matter it is indispensable or standard in the treatment of the idea, copying does not constitute infringement.372 Here some of the features of plaintiff’s game were inextricably linked to the idea,373 while others were not.374 The issue with which the court grappled was how to compare works containing some protected, and some unprotected, expression.

The court noted that in the simpler case of two works where all expression is protectable, the Ninth Circuit uses the extrinsic-objective/intrinsic-subjective test. In the first test (extrinsic/objective), ideas are compared using analytic dissection. Since defendant here conceded similarity of ideas, the extrinsic/objective test was satisfied. The intrinsic/subjective test asks whether the ordinary observer, without using analytic dissection, would find the total concept of look and feel substantially similar. The question the court addressed was how to implement this test in the present case.

An issue arose also in connection with precisely what it was that should be factored out. Defendant argued that in the application of this test, the Los Alamitos location,376 the disk titles,377 and the horse names378 should also be factored out. The court disagreed. Looking to precedent, it concluded that the only appropriate exclusion was for expression inseparable from idea. Expression unprotectable for other reasons379 should not be excluded from the determination of total concept and feel.

Although even unprotectable material should be considered when determining if there is substantial similarity of expression, no substantial similarity may be found under the intrinsic test where analytic dissection demonstrates all similarities arise from the use of common ideas… .

The reason for this rule is commonsensical and sound: part of what makes a work uniquely the author’s own is how she or he arranges the materials within it. A work’s originality can stem as much from an author’s arrangement of component pieces as from the originality of the pieces themselves. It is for precisely this reason that the intrinsic, total concept and feel test forbids element-by-element comparison.

Accordingly, while it was proper to use analytical dissection to factor out the scenes a faire in Rodesh’s disc, the remainder of the work must be compared as a whole—even if it is comprised of individually uncopyrightable elements.380

Applying that test, the court held that no reasonable fact finder could find substantial similarity. The court’s rationale was that horse race video games sharing the unprotectable scenes a faire would be similar, no matter how different their non-scenes a faire elements. “This shows that the scenes a faire elements account for the expressive similarity of the rival discs. Accordingly, the remaining elements cannot be said to make the two works substantially similar.”381

Plaintiff also alleged trade dress infringement, i.e., that the total image of the product was copied. The elements of such a claim are that the dress be non-functional; that it be inherently distinctive or have secondary meaning; and that imitation be likely to cause confusion among purchasers. A feature is functional where it is essential to the product’s use or affects its cost or quality. On this point, the court saw an issue of fact regarding the names of the horses, the track and the disc. Similarly, the appellate court saw a fact question as to whether the claimed dress was distinctive382 or had acquired secondary meaning.383

Healthcare v. Lippany384 involved an ex-employee of a hospital management consulting firm that had developed a number of computer programs. Years before leaving plaintiff’s employ, defendant began developing for his own account programs functionally similar to some of plaintiff’s. Defendant’s programs embodied plaintiff’s confidential information and methodologies.

In deciding plaintiff’s motion for a preliminary injunction, the court found the issue of substantial similarity to be controlling. The court followed Whelan v. Jaslow385 for the proposition that the “idea” is the purpose of the copyright-protected work, and “the expression of the idea in a software computer program is the manner in which the program operates, controls and regulates the computer in receiving, assembling, calculating, retaining, correcting, and producing useful information …”386 Plaintiff here showed that other programs were functionally similar to those of plaintiff and defendant, but employed different methodologies than did the programs of the parties.

“The critical determination is whether these methodologies constitute ‘expression’ and hence are entitle to copyright protection, or whether, despite the fact that they have been characterized by plaintiffs as the programs’ structure, sequence and organization, they are more akin to ‘idea,’ which must remain in the public domain.”387 The Whelan court compared the organization of the modules and subroutines of the two programs. For example, that court noted that defendant’s program performed certain report/sales analysis functions in the same way except for reversing the order of certain steps. However, in the instant case there was very little evidence of “sso.”388 The evidence showed that “certain choices were made among factors at a gross level, e.g., the scope of the system, the number of variables to be used or the portion of the workforce to be included in calculation of labor hours. But this, held the court, was not so.

Plaintiff introduced evidence of choices among alternative factors in the development of the program.389 Such choices influence the structure of the program. But there was no evidence to show how these choices would translate into a program, or how they would “operate, control or regulate the receipt, assembly, calculation, retention, correlation and production of information.” The court therefore concluded that plaintiff had not shown that its methodologies comprised the sso of a program.

Johnson v. Phoenix390 presented a situation where plaintiff’s program, registered for copyright, controlled waste water treatment plants. Plaintiff sued a number of ex-employees and their new company for copyright infringement, trade secret misappropriation, and other causes. The district court granted plaintiff a preliminary injunction. On appeal, defendant argued that the program was not original. The court of appeals held that the fact that other programs were “similar” was insufficient to rebut the prima facie effect of plaintiff’s copyright registration.

Defendant also argued that the district court erred in concluding that a copyright could be infringed without copying literal elements. “Whether the non-literal components of a program, including the structure, sequence and organization and user interface, are protected depends on whether, on the particular facts of each case, the component in question qualifies as an expression of an idea, or the idea itself.”391 Here the district court found that the sso was expression, and the appellate court was not prepared to find this was clear error. Likewise, the court of appeals viewed the master’s report as describing sufficient similarity as to indicate there was no clear error in the lower court’s finding of copying.392

In Foresight v. Pfortmiller,393 defendant altered plaintiff’s program, under contract to a party owning a copy of the program, by adding files. Plaintiff sought a preliminary injunction precluding modification and distribution. The court denied the motion, holding that 17 U.S.C. § 117 permitted the owner of a copy of a program to authorize the adaptation of a copy as an essential step in its use of the copy. Adding features, according to the court, constituted making an adaptation. So long as the adaptation was for the sole use of the party owning the copy, the court saw the conduct as within Section 117. However, the court did enjoin sale of the enhanced product.

In Service & Training v. Data General,394 one issue was whether defendant’s program had sufficient originality to sustain copyright. Plaintiff argued that the program was a derivative work of plaintiff’s program. In determining that defendant’s program was not a derivative work, the court found it probative that while plaintiff’s program had 200 files, defendant’s had some 1,200 files. Defendant had secured a copyright registration, and the court held that plaintiff failed to rebut the presumption of validity that attached to the registration.

For years an issue hotly debated, yet barren of direct authority, was the copyrightability of “microcode.” The first decision on this point came in a suit wherein NEC was accused of copying Intel’s microcode.395 In 1989 the U.S. District Court for the Northern District of California ruled, after a bench trial, that microcode was copyrightable subject matter.396 The court also held that Intel had forfeited its copyright and that NEC did not infringe.

Copyrightability. The court found that microcode, generally, fell squarely within the definition of “computer program” set forth in Section 101 of the Copyright Act.397 The court noted that microcode, like other program codes, is afforded copyright protection as a “literary work,” and that a particular literary work is copyrightable if “fixed in any tangible medium of expression” (undisputed in this case), and “original.”

NEC contended that many microsequences of this particular microcode were not copyrightable because each consisted of only a few obvious steps, and thus lacked the necessary originality because fragmentary and dictated solely by functional considerations. The court disagreed. While it conceded that several microsequences in the code, considered alone, might be directed solely by functional considerations, these were only small segments. Moreover, any work could be chopped into parts, each of which could be said to have very few creative steps. Originality, ruled the court, means only that the work was independently created and that its creativity exceeds that required for a fragmentary work or short phrase. This microcode met the test.

NEC argued also that the microcode was an element of the computer itself, and therefore could not be used “in” a computer, as required by the Copyright Act definition of “computer program.” But the court saw nothing in any statute suggesting that copyrightability for a program depended on what type of program it was, or upon the function it served within the machine.

Finally, NEC claimed that only expression of an idea was copyrightable (rather than the idea) and that when the two are inseparable, copying is not infringement. Accordingly, NEC contended that if there were substantial similarities, they resulted from constraints that severely limited the ways in which the ideas could be expressed. The court chose to treat this argument as bearing on whether there was infringement, rather than on whether the code was copyrightable.398

Forfeit. The court went on to hold further that Intel had forfeited its copyright because notice had been omitted from more than a relatively small number of copies, because it failed to take reasonable efforts to add notice after the omission had been discovered, and because the copies were distributed absent a written requirement mandating such marking.

Infringement: The court held there was no infringement because there was no substantial similarity. Two experts testified, with conflicting conclusions. In deciding the issue, the court admonished against permitting dissection of the accused work into pieces, and viewing each piece in isolation. The court found that none of the ninety microroutines in NEC’s code were identical to Intel’s although some of the shorter ones were substantially similar. But most of these short substantially similar routines involved straightforward operations in which close similarity was to be expected in the absence of copying. Moreover, some of NEC’s shorter microroutines were substantially different from Intel’s. The court concluded that an ordinary observer considering the accused microcode as a whole would not recognize it as having been taken.

Because several of the shorter microroutines were substantially similar, the court did not end its infringement analysis here, but acknowledged an obligation to make a qualitative rather than quantitative judgment, and looked to whether actionable copying was responsible for those similarities.399 Intel offered as evidence of copying the relative inexperience of NEC’s programmer with microprograms, the small time frame involved and the relatively few notes NEC’s programmer made (as compared to notes he made in his work on other micro-programs). The court did not view this as probative, finding that the programmer was very talented, had significant assembly language experience, was able to accomplish the task within the two month time frame, and had no reason to make multitudinous notes since he was working alone and the task was relatively simple.

Interestingly, although the programmer had previously disassembled Intel’s microcode, and did not contend that he had not been influenced by that experience, the court acknowledged only that this experience may have been “a source of ideas” that he utilized. Finding also that he did not undertake to copy, the court was not convinced that disassembly of Intel’s code by the programmer formulating the NEC code required a finding of substantial similarity.

Similarly unimpressive to the court were Intel’s evidence of an alleged admission of substantial similarity made by NEC in the Japanese language, and evidence of a program “patch” found in the same portion of plaintiff’s and defendant’s work.400 On the first point, the court found insubstantial probative evidence, and on the second it found no evidence probative of copying. Further, Intel alleged copying because in a six-line microsequence the NEC instructions used the same registers401 as did the Intel instructions. But the court held that circumstances other than copying could explain this and that, at any rate, this similarity did not exist in the final version of NEC’s code. And in dealing with two other pertinent microsequences the court was willing to assume arguendo that in preparing NEC’s initial version NEC relied on its disassembly of Intel code.402 But because this initial NEC code was substantially changed by NEC, the court held that any such copying was irrelevant since because where copying disguised as to be unrecognizable there is no infringement.

Intel argued that three other alleged uses of disassembled code were “strongly suggestive” of copying. First, it argued that the NEC microcode’s use of the “dual-bus” characteristic403 of the microprocessor far less frequently than dual-bus used by the “clean room” code404 subsequently created by NEC indicated copying. The Intel code did not use the dual-bus feature at all. The court saw this as indicative of nothing pertinent. Second, Intel noted that twelve of thirteen “groupings” of subroutines in NEC’s code were the same as in Intel’s. The court negated the importance of this because logic, to some extent, determined subroutine grouping.405 Also, NEC used the same disfavored method of treating “illegal” instructions as Intel did. The court was willing to assume that this was evidence of idea copying, but was unable to find in the final version of NEC’s code any expression of that idea.

Another Intel argument relied on the fact that NEC’s microprocessor had twice as much ROM available to it as Intel’s. Accordingly, Intel argued that there was no tight size constraint imposed on NEC and in the absence of copying NEC should have had better microcode. But the court noted that NEC elected to use the additional ROM for additional microinstructions. “This also was a legitimate constraint; NEC was not obliged to avoid the similarity that other constraints imposed by creating a larger microcode.”406

Intel tried to show there were many alternate ways in which NEC could have written its microprograms, but the court held otherwise, noting that Intel failed to consider the constraints that limited NEC’s choices. NEC had a license from Intel to duplicate the Intel microprocessor hardware, and NEC’s use of that hardware limited substantially the choices available to it in creating microcode.

Finally, the court felt that the “clean room” microcode created by NEC constituted compelling evidence that the similarities between the NEC code and the Intel code resulted from constraints. The court stated that there was no infringement in NEC’s creation of the clean room software and concluded that similarities between the clean room microcode and Intel’s microcode were attributable largely to the constraints of hardware, architecture and specifications. The court found further, that the similarities between the clean room microcode and NEC’s microcode were at least as great as the similarities between NEC’s microcode and Intel’s. Accordingly, it concluded that those similarities also resulted from the constraints. This is a curious comparison. It would seem that if comparisons among the three sets of code are to be used as a basis for exculpating the NEC code, the appropriate comparison is similarities between clean room and Intel code, versus similarities between NEC and Intel code.

The court also found probative the testimony of Intel’s programmer that he would expect independently created microcode to have fewer similarities in longer sequences than in shorter sequences because “there is more opportunity for the longer sequences to be expressed differently.”407 Accordingly, the court found that the constraints, rather than copying, were responsible for the principal similarities between NEC’s microcode and Intel’s.

Finally, the court looked to those shorter simpler microroutines of NECs that closely resembled Intel’s (although none were identical). In determining an idea’s range of expression, constraints were relevant factors to be considered. NEC’s microcode was constrained in its expression by use of the microinstruction set in the hardware of the Intel microprocessor. Because of those constraints, these simpler microroutines could be expressed in only a few limited ways. Because NEC was free to use the embodied ideas, and since the underlying ideas of each of these microroutines were capable of only a limited range of expression, the expressions could be protected only against “virtually identical” copying, which did not transpire here.

In what once was a leading case in the area, Whelan v. Jaslow 408 presented a situation where defendant was a dental laboratory, and plaintiff a developer and marketer of computer programs. Defendant’s principal believed his operations could be made more efficient by computerization. He hired plaintiff’s predecessor to develop custom software for plaintiff’s IBM computer under an agreement providing that plaintiff’s predecessor would design a system for defendant’s needs and could market it to other dental laboratories, for which defendant would receive a royalty. Plaintiff’s principal wrote a program (the “Dentalab” program) in the EDL language, suitable for the IBM Series One. Defendant was plaintiff’s sales representative for that program. Defendant’s principal subsequently realized that the marketability of the Dentalab program was somewhat restricted because some dental labs used computers which would not accept EDL. He then created a corresponding program in the BASIC language, which was the allegedly infringing program.

Defendant terminated the distributorship agreement, stating it considered itself the exclusive marketer of the Dentalab program, which contained its “valuable trade secrets”. Defendant subsequently marketed both programs, advertising the Dentcom program as “a new version of the Dentalab computer system.” Plaintiff alleged that defendant’s marketing of the Dentalab and the Dentcom programs infringed plaintiff’s copyright in the Dentalab program.

Plaintiff’s expert testified that the Dentcom program was not a translation of the Dentalab program but that the programs were similar in three respects: most file structures and screen inputs were virtually identical, and five particularly important sub-routines performed almost identically. Defendant’s expert compared the source and object codes of the two programs and testified about the many ways in which they differed. He concluded that “substantive differences in programming style, in programming structure, in algorithms and data structures, all indicate that the Dentcom system is not directly derived from either of the other systems.”409

The district court found that plaintiff was the sole author of the Dentalab program and that the contract between plaintiff’s predecessor and defendant’s principal stated plaintiff’s predecessor would retain full ownership of the software. The lower court held the copyright valid and infringed by defendant’s marketing of the Dentalab program.

The court enumerated the steps in program creation as:410 identifying the problem; learning about the problem so as to outline a solution; creating a flowchart; breaking down the problem into modules; determining the nature and position of input; structuring the data files; coding (only 20% of cost of program development); debugging; documentation; and maintenance. In this case plaintiff’s principal spent much time studying the lab, organizing the modules and working out data arrangements, and relatively little time coding.

The traditional test of substantial similarity is bifurcated: the fact finder decides (with expert testimony) whether there is sufficient similarity to conclude that the copyrighted work was used in making defendant’s; if so, the fact finder decides (without expert assistance) from the perspective of the lay observer whether the copying was illicit.

The “ordinary observer test” (developed for cases involving novels, plays, etc.), which does not permit expert testimony, is of doubtful use where subject matter is complex. Likewise, the distinction between two parts of the bifurcated test may be of doubtful use where the finder is the same person exposed to the expert in the first step. The appellate court rejected the ordinary observer test and adopted the single substantial similarity test (wherein expert testimony is available). That test was applied by the district court.

Defendant had two arguments: copyright does not extend beyond the literal elements (source and object code) of a program, so that similarity in overall structure (the basis for the district court’s finding of infringement) is irrelevant; and even if copyright extends to non-literal elements, there was not sufficient evidence of substantial similarity.

It is well settled that copyright extends to computer program source and object codes. However, the district court found no such copying here. Rather, it held there was infringement because the overall structure was substantially similar. On appeal the issue was whether mere similarity of overall structure can be the basis for infringement, i.e., whether copyright covers the structure of a program, or only its literal elements.

The court noted that section 102(a)(1) extends copyright to literary works, and computer programs are literary works. Copyright in other literary works can be infringed without substantial similarity of literal elements. One can violate the copyright in a play or book by copying the plot or plot devices. By analogy it would appear that copyright in a program can be infringed absent copying of literal elements. Defendant argued that, while true of other literary works, this was not true of computer programs because of (1) the Section 102(b) dichotomy between idea and expression; and (2) the legislative history embraced in the CONTU report.

Section 102(b) of Title 17 states that copyright does not extend to any idea or method of operation, and defendant argued that the structure of a program is the idea and not the expression, so that structure cannot be protected by copyright. The court disagreed, promulgating a rule for distinguishing idea from expression in the context of programs. Admitting that the distinction may be elusive, the court stated it must note the pragmatic considerations underlying the distinction: the balance between protection (incentive) and dissemination of information. In Baker v. Selden the dispute was whether Selden’s blank forms (identical to Baker’s) were part of the method (idea) of Selden’s book, and hence non-copyrightable, or part of the copyrightable text (expression).411 Selden held that where the method cannot be used without employing the diagrams, such diagrams are to be considered as necessary incidents to the art and within the public domain. Selden held the blank forms were necessary incidents to the method and not entitled to copyright protection.

The Whelan court drew from this a rule for distinguishing idea from expression: the distinction may be drawn with reference to the end sought to be achieved by the work in question: “… the purpose or function of a utilitarian work would be the work’s idea, and everything that is not necessary to that purpose or function would be part of the expression of the idea … where there are various means of achieving the desired purpose, then the particular means chosen is not necessary to the purpose; hence there is expression not idea.”412 The court viewed the “idea” in the Dentalab program as the efficient management of a dental laboratory;413 because that could be accomplished in a number of ways with a number of different structures, the structure of the Dentalab program was deemed part of the program’s expression, and not its idea.

Scenes a faire are incidents, characters and settings indispensable in the treatment of the given topic, and therefore not protectable by copyright. Granting a copyright would give the first author a monopoly on commonplace ideas. This is a restatement of the hypothesis that the purpose or function of a work is part of its idea; anything necessary to effect that function is also necessarily part of the idea. Likewise, fact-intensive works are given limited copyright coverage because there are only a limited number of ways to express factual material and those devices therefore belong to the idea and no the expression. Moreover, this rule, permitting copyright protection beyond literal computer code, was viewed by the court as providing incentive for programmers by protecting their most valuable efforts, while not giving them a stranglehold over development of new programs to accomplish the same end.

Defendant raised two arguments against this position. First, it stated that programs are so intricate they are almost impossible to copy except literally, and that anyone who attempts to copy structure without copying literal elements must expend much effort and creativity. The court rejected that argument because first, it is not true that approximation of a program, short of perfect reproduction, lacks value. Such copying would yield a significant advantage over competitors. Secondly, the fact that it may take much effort to copy does not mean the copier is not an infringer. Defendant also argued that in programs the concept of structure was too vague to be useful in copyright cases. The court admitted that limiting copyright to literal codes would be simpler; but ease of application was found not to be a sufficient counter-weight to the considerations espoused by the court.

Defendant also argued that the developmental process is different in computer programing than in other fields, requiring a particularly restricted application of copyright. Under this theory, progress in programming is achieved by using stepping stones, which requires plagiarizing; giving programs too much copyright protection retards progress. But the court was unconvinced that progress in computer programs was different.

Applying its general rule to the facts, the court noted that the purpose of the Dentalab program414 was to aid in the business operations of a dental laboratory. The court took pain to deny any implication that the idea or purpose behind every utilitarian work would be whatever it accomplishes, so that structure and organization would always be part of the expression. The court recognized that the idea or purpose behind the utilitarian work may be to accomplish a particular function in a certain way, although the court saw no suggestion that the purpose of the Dentalab program was so refined. The structure of this program was not essential to the task; there were other programs performing the same function with different structures and designs.

The district court stated that the mere idea of a computerized program for operating a dental lab would not be subject to copyright; different programs may functionally serve similar purposes without being copies of each other. It stated that expression of the idea in a program is “the manner in which the program operates …”415 The appellate court agreed, concluding that “the detailed structure of the Dentalab program is part of the expression, not the idea, of that program.”416 The court found support for this conclusion in SAS v. S&H, discussed infra.

The court also held that the 1976 Copyright Act indicated that Congress intended that structure and organization of a literary work be part of its expression, protectable by copyright. The court supported this conclusion by noting that the definitions in the Act of “compilation” and “derivative work” spoke to assembling materials coordinated or arranged in a certain way, and to abridgement, condensation and recasting of a work.

The court further noted that its decision would be at odds with Synercom,417 which held that input format forms were to be construed as ideas, not expressions. Synercom concluded that the organization and structure of input formats was inseparable from the underlying idea. That court acknowledged that in some cases structure and sequence might be part of expression rather than idea but held that in the case of input format forms, structure and organization were inherently part of the idea. Synercom pondered: “if sequencing and ordering [are] expression, what separable idea is being expressed?” In the instant case the court concluded that the idea is “efficient organization of a dental laboratory.”418 Because there are a variety of program structures through which such an idea can be expressed, structure is not a necessary incident to that idea.

A second argument of defendant’s is that the CONTU Report is the legislative history for the 1976 Copyright Act and precludes copyright protection of structure. The court disagreed, first noting that the CONTU Report is pertinent only to one portion of section 101, and to section 117 of the Act, and did not affect any other section. Neither of those sections was at issue in this case. Nor, held the court, did the CONTU Report suggest that copyright be limited to literal code.

Defendant alleged there was insufficient evidence of substantial similarity to support the finding of infringement. The experts testified as to three areas. First, as to file structures, defendant’s expert analogized computer files to blank forms containing no information, and asserted that blank forms cannot be protected by copyright. While the court viewed the analogy as correct, it held that blank forms can be copyrighted where sufficiently innovative that the arrangement is itself informative. It appeared to the court that the structures in question were sufficiently complex and detailed to sustain copyright. The court believed there were many ways in which the same goal (organization of the business aspects of a dental laboratory) could be accomplished, and several approaches might use very different file structures or might require different information.

Secondly, as to screen output defendant claimed that screen output is covered by a different copyright than is the underlying program, and that screen output bears no relation to the underlying program. The court acknowledged that screen outputs are audio-visual works covered by a different copyright, and that plaintiff asserted no claim of copyright infringement with respect to screen outputs. However, the court felt that screen output nevertheless might be relevant to the question of whether the copyright in the program had been infringed. While it could not be direct evidence, the court believed it could be inferential evidence, and that the risk of undue prejudice did not outweigh its probative value. Judge Rosenn believed that screen outputs here would have no probative value.

Also, as to the five important sub-routines in question, defendants claimed they failed “to understand how a substantial similarity in structure can be established by a comparison of only a small fraction of the two works.”419 Thus, defendant asserted that one cannot prove substantial similarity without comparing the greater part of the two works in question. The court rejected that contention, citing numerous cases involving other types of works, and failing to see a difference in the case of computer programs. “Because all steps of a computer program are not of equal importance, the relative inquiry cannot therefore be the purely mechanical one of whether most of the program’s steps are similar. … We must ask whether the most significant steps of the programs are similar.”420 The court found that this is just what plaintiff’s expert had done.

Finally, defendant claimed the district court erred in evaluating the testimony of the two experts. Reviewing the testimony, the court saw no clear error. Defendant’s expert did not compare the two programs that were the basis for the appropriate determination of substantial similarity. Also, defendant’s expert concentrated on a comparison of source codes, whereas plaintiff’s expert spoke extensively about structures.

In SAS421 defendant created a statistical analysis program package for use on IBM equipment. Plaintiff established design goals for a similar program for use on DEC equipment. In connection with that project and without communicating its intentions, plaintiff obtained a license to use defendant’s software and received some of defendant’s source code.

After plaintiff had done substantial development work, it explored with defendant the possibility of joint marketing, and defendant explained the copyright and otherwise proprietary nature of its software. Subsequently, defendant terminated the software license for alleged violations. This declaratory judgment action involved issues concerning breach of contract and copyright infringement. Both parties moved for summary relief.

Defendant sought partial summary judgment on its copyright infringement claim. A licensee infringes a copyright by exceeding its license. Accordingly, using defendant’s program on a non-designated CPU, making unauthorized copies, and post-termination use of defendant’s program all constitute copyright infringement, so that summary judgment on this issue was granted.

There is no longer doubt that computer programs are copyrightable. However, plaintiff alleges that as a matter of law its program cannot be a “copy.” It relies on section 102(b), which states that copyright protection does not extend to any idea or process. The court ruled that defendant here sought not to protect ideas but rather their particular expressions in the form of specific programs. And plaintiff wanted to use and market those expressions without investing the capital and manpower necessary to develop them.

It is clearly an infringement to translate a program from one computer language to another, as it is to translate a novel from English to French. In each case the substance of the expression is the same between original and copy, with only the external manifestation of the expression changing. There would probably be copyright infringement in taking a detailed description of a particular problem solution, i.e., a flow chart or a step-by-step set of prose instructions, and programming that in computer language. But there would be no infringement in preparing a program from a mere general description of the problem to be solved because there the program would require substantial imagination and independent creativity. Hence it could no longer be merely a copy or version of the problem statement. (The court cited Synercom for these propositions.)

But if one adopted an existing program and made very simple changes involving mere rearrangement of existing expression to disguise copying, there would be an absence of worthwhile imagination and independent thought. These simple changes would involve even less independent thought than translating a novel from English to French. The copyright “cannot be limited literally to the text, else a plagiarist would escape by immaterial variations.”422

The fundamental question concerned what level of similarity short of verbatim copying constitutes infringement. Essentially this is an ad hoc determination. One issue here was what if anything was directly copied or altered insignificantly. A genuine issue of material fact existed on this,423 so that on that matter summary judgment was denied.

At trial, the testimony of the experts of both parties was that two-thirds of the coding in the present version was new, which the court ruled sufficient for it to constitute an original work of authorship.424 The SAS program has three parts: statistical analyses, linkage, and supervisor. Under its license, S&H received source code for the statistical analyses and linkage portions.

To load SAS source code onto its DEC equipment, S&H had to (and did) translate it from the IBM coding scheme (EBCDIC) to the DEC coding scheme (ASCII).425 The result (which could not be run on the DEC computer) was viewed by the court as a copy, was not for backup and was not authorized. However, once it was loaded S&H programmers were able to manipulate SAS source code as in a word processor.

SAS source code was extensively and systemically used by S&H in the preparation of the S&H product. After the dispute arose, S&H made cosmetic changes to its source code so as to mask and disguise evidence of copying. Plaintiff’s expert, relying in part on some forty-four examples, testified that the S&H source code was substantially similar to that of SAS. Defendant’s expert did not disagree substantially as to these examples but claimed they were not evidence of copying. S&H argued that the examples were trivial in comparison with the 186,000 lines of source code.

Plaintiff argued that in preparing its program, S&H made an unknown number of unauthorized (and therefore infringing) exact copies, and also that the S&H program itself was an infringing copy or derivative work. The court held that S&H’s use of the SAS program on an unauthorized computer to make copies for unauthorized purposes, and use following plaintiff’s lawful termination of the license constituted infringement.

The court ruled defendant had copied, based on these forty-four examples and other evidence. The lack of design documentation for the S&H product was viewed by the court as further evidence of copying: S&H adopted the SAS program “wholesale and in detail as the design document for its product.”426 In addition, the S&H software contained evidence of “undocumented options,” such as vestiges of options once existent, and eliminated by plaintiff except as to certain statements overlooked. Further evidence of copying was found in the organizational patterns used by S&H.

The unrebutted testimony of plaintiff’s experts also established that defendant’s program followed the organizational structure of SAS down to a detailed level. S&H argued that it essentially adopted the SAS organizational scheme but wrote the actual program code independently of SAS code. Even if so (as discussed above, the court found otherwise) the court believed there would be infringement, citing Meredith Corp. v. Harper & Row Publishers, Inc.426.1 Meredith found a process, by which a textbook was outlined and a number of individuals wrote “original” text based on the outlines, to result in duplication of expression and not merely idea. The court also saw as supporting its position Synercom’s 427 statement that programming in source code the detailed English language solution of a problem would duplicate expression and infringe.

On the issue of whether S&H appropriated only ideas, the court noted it was particularly difficult in the case of complex software to make such a determination. S&H presented no evidence that functional abilities, ideas, methods and processes of SAS could be expressed in only very limited ways. To the extent similarities existed, the court held they represented unnecessary and intentional duplication of expression. The court found as a matter of fact that the expression and not merely the ideas of SAS were copied, and that the copying was substantial. The author of a complex computer program

“is faced with virtually endless series of decisions as to how to carry out the assigned task. Beginning with a broad and general statement of the overall purpose of the program, the author must decide how to break the assigned task into smaller tasks, each of which must in turn be broken down into successively smaller and more detailed tasks. At the lowest levels the detailed tasks are then programmed in source code. At every level, the process is characterized by choice, often made arbitrarily, only occasionally dictated by necessity. Even in the case of simple statistical calculations, there is room for variation, such as the order in which arithmetic operations are performed. S&H itself presented evidence that there are at least two substantially different ways of programming the ‘MEANS’ procedure, one of the simplest statistical analyses. As the sophistication of the calculation increases, so does the opportunity for variation of expression.”428

Further, the court held that defendant’s argument that it could cure the infringement by simple excision was flatly inconsistent with the statute: “The argument essentially is that S&H can misappropriate every internal and external feature of building, can then cure any impropriety by changing the tint of the windows and the color of the siding.”430

As for the breach of contract claim, S&H violated the terms precluding use on an unauthorized CPU, creation of unauthorized copies, and use subsequent to proper termination. S&H also violated the implied duty of good faith and fair dealing by creating a new work based on the same idea, theme, or title (even though a stranger might do so without infringing). In every contract there is an implied covenant that neither party will do anything to destroy or injure the right of the other to receive the fruits of the contract.

Nor was the court limited to ordering excision of the specific lines of source code referred to by plaintiff’s expert. Programs constantly evolve, and alterations can as easily introduce infringing elements as remove them. Defendant continued to alter its program during the litigation, and the court doubted that all results of copying had been or ever would be excised. Nor could the supervisor portion of defendant’s program be surgically separated and said not to infringe. The boundary between supervisor and non-supervisor portions was different in the two programs. Code for handling input/output in plaintiff’s program was outside the supervisor (so that defendant received source code), but in defendant’s program such code was inside the supervisor. “This court is not a computer programmer, and is simply not able to determine, out of 186,000 lines of integrated product code, which of them reflect misappropriation. Nor is the court willing to rely on adjudged plagiarists, whose courtroom testimony the court found not worthy of belief, to make that determination.”431 The court enjoined the marketing of defendant’s program or any other product copied or derived from it or from plaintiff’s program.

In Goodpasture,432 plaintiff cooperative developed software (“Telcot”) to provide members with information regarding cotton prices, availability, accounting services, and the capability to consummate sales electronically. The system comprised “functional specifications outlining what the software can accomplish, design specifications determining how the software accomplishes those functions, programs implementing the design specifications, and documentation detailing the program and specifications so that future programmers can understand and modify the software.” Users could gain access to the software, residing in a central computer, but not to the documentation, programming, design or functional specifications of the system.

The software was developed by plaintiff’s team of employees, and was first marketed in 1975, since which time it has continuously been refined. In 1979, plaintiff’s general manager left to start CXS, which then agreed with plaintiff to develop and market a PC version of the Telcot software; any modification, developments or enhancements of the Telcot software would be owned jointly.

In 1984 CXS hired the development team away from plaintiff and employed them on the PC project. One member, upon leaving Plains, made and took a tape copy of the Telcot source code. Subsequently, Plains and CSX terminated their agreement. After about a year at CXS, the development team designed a PC version of the Telcot software, and produced a document setting forth design specifications for it. No programming was written. At about that time defendant Goodpasture hired the development team away from CXS.

The development team signed employment agreements with Goodpasture agreeing not to breach any confidences of former employers. In violation, one member brought to Goodpasture a diskette containing the Telcot software. Twenty days after arriving at Goodpasture, the team had completed design of a PC version of the Telcot software designated “GEMS.” That software was not fully operational at the time this action was brought.

The GEMS software is very similar to the Telcot software in terms of “functional specifications, programming and documentation levels.” Several pages of the GEMS design manual appear to be direct copies of pages from the design manual created at CXS. The main difference between the two seems to be that the Telcot software is designed to work on a mainframe, whereas the GEMS software is designed for a PC. Defendants allege that with one exception they did not copy programs, but instead “drew on their knowledge of the cotton industry and expertise in computer programming and design gained over a number of years.”

Plains sued Goodpasture and the development team for misappropriation of trade secrets, unfair competition, conversion, breach of confidential relationship, trade dress infringement, and copyright infringement. Plains moved for a preliminary injunction on its copyright and trade secret claims. The district court denied the motion, concluding that appellant failed to show irreparable harm or substantial likelihood of success. Plaintiff contended on appeal that the district court abused its discretion by using the wrong legal standard in analyzing the copyright infringement claim, and by not addressing the trade secret misappropriation claim.

Plaintiff asserted the district court failed to consider evidence of substantial similarity, because it had inquired solely into direct copying, whereas the central issue was organizational copying, i.e., reproduction of the structure of the system as outlined in the design specifications. Plaintiff asserted further that the evidence of substantial similarity in organization was uncontroverted. One member of the team did copy a sub-routine from the system, and when Goodpasture discovered this the sub-routine was replaced and that employee discharged.

The court of appeals did not accept the contention that the lower court ignored evidence of organizational copying. Rather there was testimony by several witnesses for defendant that similarity existed “on a level not protected by appellant’s copyright.” One witness testified that he compared the “overall structure of these programs,” and stated he had “expanded his study” of the program from the narrow search for verbatim copying. The district court could rely on these statements in interpreting the expert’s opinion that the GEMS software was not copied from the Telcot software. There was also testimony on behalf of defendant that it did not rely on any material belonging to plaintiff or CSX, and that a mainframe program such as the Telcot could be altered to run on a PC only with enormous changes (so that rewriting the program would be faster than modifying it). The court viewed the testimony as sufficient to support a tentative finding that the defendant did not copy plaintiff’s software; the ruling below was not clearly erroneous.

“The legal finding of the district court ultimately rests on a judgment about the extent of the protection offered by plaintiff’s copyright. “The court of appeals followed Synercom,433 which held that “input formats” of a computer program were ideas and not expressions, and thus not protected by copyright. “To the extent that input formats represent a level of computer software design more specific then functional design and more general than line-by-line program design, the issue of their copyrightability is relevant to the issue of whether GEMS infringes on protected Telcot designs.” Plaintiff urged reliance on Whelan,434 which rejects the Synercom premise that “there is a difference between the copyrightability of sequence and form in the computer context and in any other context.” In Whelan the structure, sequence, and organization of programs were held copyrightable.

But the Fifth Circuit declined to follow Whelan for two reasons: first, the record here was only partially developed; second, defendant presented evidence that many similarities between the two programs were dictated by “externalities of the cotton market.” “… market factors play a significant role in determining the sequence and organization of cotton marketing software …” To that extent, the Fifth Circuit felt the facts of the case fit squarely within Synercom’s analogy to the hypothetical development of gear stick patterns. The court therefore declined to hold that those patterns could not constitute “ideas” in a computer context.

The preliminary injunction based on misappropriation was also denied, with the district court finding that plaintiff could not demonstrate a likelihood of success on whether defendant used the trade secrets. “The critical issue is that misuse of these implementations can occur only through copying the particular software designs on a sufficiently specific level. These are not alleged marketing or mode or doing business trade secrets. They are matters of design, where the issue of misuse boils down to evidence of copying. If no copying occurred on any level, appellant cannot demonstrate that appellees misused the trade secrets they allegedly possessed.” The court held that the district court properly relied on the opinion of defendant’s expert that no copying of plaintiff’s software occurred. “If appellees did not in any way ‘copy’ any part of appellant’s protected idea or expression, then appellant cannot demonstrate trade secret misappropriation any more than it can show copyright infringement.”

In Bull v. American Express,435 plaintiff developed a compiler program, “TPS6,” for compiling its proprietary computer programming language “Screenwrite,” for use in processing a high volume of transactions quickly on Bull computers. Plaintiff licensed the use of TPS6 to defendant and assisted defendant in writing, in Screenwrite, certain application software. Plaintiff alleges that without consent defendant had a third party develop an IBM version of the TPS6 compiler. Defendant’s software allegedly copied the structure, sequence and organization, and look and feel, of plaintiff’s.

Plaintiff alleged, inter alia, infringement of the copyright in the TPS6 program and of the “Screenwrite language” and “certain functions of TPS6.” Defendant moved to dismiss as to the claim relating to the Screenwrite language and the functions. Plaintiff did not claim that it had a separate copyright registration for “Screenwrite” apart from the TPS6 registration, or that the TPS6 registration covered “Screenwrite.” Holding that a plaintiff cannot sue for copyright infringement without a valid registration, the court dismissed the claim.436

The court further held that structure, to the extent not necessary for function, was protectable expression. The court adopted the Whelan437 view that the function of a program is its idea, and everything unnecessary to that function is expression. Because there might be other operating systems that achieved the same function as TPS6, the court held that resolution of whether defendant’s software infringed must await trial.438

In Johnson v. Phoenix,439 both parties marketed programs designed to control waste water treatment plants. The issue was whether the structure, sequence and organization (“sso”) of defendant’s program infringed plaintiff’s copyright. The district court granted a preliminary injunction, which the court of appeals affirmed. The appellate court held that sso is protectable by copyright, and that idea here did not merge with expression.

In Dynamic Solutions v. Planning & Control440 plaintiff sought a preliminary injunction precluding copyright infringement. Defendant provided business training programs using game simulation software developed by plaintiff. The programs originally used by defendant were written by plaintiff for a mainframe time-sharing computer. The parties signed an agreement in 1975 specifying that plaintiff would serve as defendant’s time-sharing broker; that plaintiff would provide consulting services for creating new simulation software; that plaintiff would make available certain proprietary software; that software subsequently developed by plaintiff would be the property of defendant; and that defendant owned the software already created. Plaintiff struck the provision transferring ownership of its proprietary software. The court distinguished property rights in the games themselves (concededly defendant’s) from the rights in the programs running the games. “That distinction is possible because different computer programs can be written that produce the same output … ,” citing Apple v. Formula.441

The programs at issue here were written by plaintiff in 1983 and 1984 for microcomputers, and operate two of the same games as do programs plaintiff wrote in 1975 for use on time-sharing mainframe computers. Plaintiff conceded that the property rights in the 1975 programs belonged to defendant. Looking to the 1975 agreement and the conduct of the parties, the court ruled that plaintiff owned the property rights in the microcomputer programs at issue.

The court next considered the validity of the plaintiff’s copyrights on those programs as derivative works. Defendant first claimed that plaintiff’s use in the microcomputer programs of the pre-existing 1975 software was unauthorized. But the court found that plaintiff’s use of the earlier works in preparing the microcomputer programs was authorized by defendant either implicitly or by way of ratification.

Defendant next claimed that the copyright was invalid for plaintiff’s failure to deposit a copy of its works in registering them. In its copyright registration statements, plaintiff stated the date of creation and publication as 1983 for one program and 1984 for the other. Those statements were accompanied by copies of the first and last 25 pages, pursuant to 37 C.F.R. Section 202.20(c)(vii)(A). However, the program versions deposited by plaintiff were not versions as they existed in those years, but were somewhat revised versions as they existed in 1986. The earlier versions no longer existed and could not be reproduced, nor could plaintiff specify how the programs changed, or whether there were any changes in the pages deposited.

The court held that this deficiency neither invalidated the copyright nor rendered it unenforceable. Although defendant characterized this as a failure to deposit the source code for which protection was sought, the court ruled it could just as easily have been an error in the dates listed on the registration form. Such errors do not affect plaintiff’s right to sue unless they are knowing and might have caused the Copyright Office to reject the application. The court saw no reason to believe that either of these things was so. Accordingly, the court found plaintiff likely to succeed on its claim that it owned a valid copyright.

Defendant raised an equitable unclean hands defense. The court noted that for such a defense to succeed, plaintiff’s wrongdoing must be “of serious proportions.” One basis asserted for the defense was plaintiff’s knowing deposit of 1986 code rather than 1983/1984 code, but the court rejected defendant’s contention that this deficiency was intentional. “The most reasonable explanation for the date discrepancy is that plaintiff entered the wrong dates on the registration forms.”442

As a second misuse defense, defendant pointed to the fact that in the space for identifying pre-existing works on the copyright registration form, plaintiff stated “computer generated simulations of same title published in 1975,” without stating that plaintiff itself had no rights in the pre-existing work. But the court did not see plaintiff’s silence on ownership there as evidence that it misled the Copyright Office.

Finally, defendant argued that plaintiff misused the copyright by using the 1975 software in training seminars in Europe without authorization or payment. While finding that the factual contentions might be correct, the court did not believe they gave rise to copyright misuse. First, the seminars were not sufficiently related to the subject matter of the action (copyright in the microcomputer programs). Nor did the seminars affect the equitable relations between the parties. Further, the court was not convinced that this conduct was sufficiently serious. Finally, the court believed it may not have occurred to plaintiff that the 1975 agreement forbade that conduct.

Moving on to the issue of infringement, the court noted that plaintiff asserted infringement in two ways. First, defendant conducted seminars in 1986 using the programs, after plaintiff’s cease and desist letter. Defendant argued that under the 1975 agreement it was entitled so to use the software in question. However, the court construed the agreement differently, found that it did not permit such use, and held that defendant’s use of the microcomputer software in these seminars infringed valid copyrights of plaintiff.

The court ruled that where the subject is complex, as are computer programs, expert as well as lay evidence may be used. The court believed plaintiff likely to prevail on its claim as to the 1986 Operations Management program, even though it controlled a different game than the 1984 Project Management program. Works may be substantially similar without being literally identical. Plaintiff’s expert testified that portions of the two works were almost literally identical, and that those similarities were not functionally required.

The court also found a likelihood of infringement, based on the testimony of plaintiff’s expert to the effect that the two programs shared structural similarities. “I agree with the Third Circuit’s careful analysis [in Whelan ] that to the extent structural similarities in a computer program are ‘not necessary to [the] purpose or function’ of the program, they constitute a protectable expression of an idea.” The court was of the opinion that plaintiff likely would show both “fragmented literal similarities” and “comprehensive non-literal similarities.” This despite the fact that plaintiff’s expert compared only a small part of the two codes. Because defendant did not yet have an opportunity to offer its own expert testimony regarding this issue (this motion having arisen on a proffer by defendant), the court permitted defendant such an opportunity.

In Arndt, Plaintiff Williams, an investor/advisor and commodities trader, created a booklet (on which he spent about 1,000 hours) embodying and explaining his trading method, and printed a dozen copies.444 He sold a copy to each of eleven preferred customers for $2,000 a copy, apparently under a confidentiality agreement. That booklet (“FTM”) bore neither copyright notice nor his name. He then corrected some typographical and textual errors, inserted copyright notice, and printed 75 copies. This edition was called “FTCM,” and he distributed them to customers for $3,000 each. One of the eleven initial customers sought plaintiff’s authorization to have the method described in the work embodied in a computer program; plaintiff gave permission on condition the method be treated confidentially. This customer then sent a copy of FTM to defendant Arndt and contracted with him to embody the method in a program. Apparently Arndt was not placed under a confidentiality obligation, and he began selling copies of the program for $1,000 each.

FTM embodied a unique system containing unique features. Other than defendant’s, no other system contained these features. The output of plaintiff’s manuals and defendant’s programs was substantially the same, “even though one was written in English and one was written in a computer language or source code. * * * It is the presence of the seven special features in both which is significant. * * * In overall structure, the two competing systems are substantially similar.”445 Defendant claimed that its source code is but a new and different expression of the idea of the market trading system.

The court found that while there may be creativity in devising source code, “source code is not an entirely new, unique expression of ideas. It is simply a translation from one language to another. The computer programmer writes in computer language the commands necessary to implement the direction provided in the narrative. The computer language can be compared to any foreign language.”446 The court concluded that in this case the source code was not a different and unique expression of a well-known idea. While the quality of a program is determined by the quality of the programmer, the program here still amounted to a translation. Otherwise, every expression would be open to copying in source code form.

The court concluded that notice was not required on FTM because the 11 copies were not publicly distributed. “They were limited to a definite, very selective group for a limited purpose on the condition that the contents not be disclosed or further disseminated.”447 Further, only a relatively small number of copies were distributed, and plaintiff made a reasonable effort to add copyright notice to those.

As to substantial similarity, the question was whether defendant’s work could be recognized by a reasonable observer as having been copied from plaintiff’s. The court stated that the most graphic evidence of substantial similarity was the comparison of FTM with defendant’s program. Plaintiff’s witness went through a step-by-step process comparing both systems, convincing the court that the works were substantially similar.

In Uniden448 plaintiff sought a preliminary injunction for infringement of the copyright on its computer program. Plaintiff and defendant each manufactured mobile radio equipment, controlled by software. Plaintiff developed its own software over a period of years. Defendant then commenced development of a mobile radio compatible with plaintiff’s. In the course of this development, defendant disassembled plaintiff’s software by translating the machine or object code into human-readable assembly language. Defendant then prepared and studied flow charts of plaintiff’s program. Plaintiff’s radio used an Intel 8049 micro-processor; defendant’s radio used a Hitachi micro-processor.

In determining the motion, the court ruled that in copyright infringement cases reasonable likelihood of success is the most predominant of the four tests. The copyright registration is prima facie evidence of validity and the facts stated therein. Moreover, defendant admitted access. As to substantial similarity, the copying need not be slavishly detailed, so long as the accused work has captured the “total concept and feel” of the copyrighted work. Atari v. North American Philips, 672 F.2d 607, 614 (7th Cir.), cert. denied, 459 US 880 (1982). Application of the ordinary observer test in a computer software context has proven problematical. Some commentators suggest an “iterative” approach to substantial similarity, focusing on an analysis of similarities and differences between the two programs. The court held that under either the ordinary observer test or the iterative test the defendant’s programs were substantially similar.

The court found a number of important instances of copying. To make compatible radios, defendant had to copy the “Barker code,” a pattern of one’s and zero’s in a prepatterned sequence. Both the sending and receiving unit must identify it for communication to be established. The system takes eight samples and compares them with an internal table to determine sample error. Both defendant and plaintiff use 56 bit samples, viewed by the court as significant for three reasons. First, while plaintiff chose 56 because its micro-processor cannot sample at a greater rate, defendant’s micro-processor could, and would function better if it sampled at a higher rate. Secondly, defendant’s error table exactly duplicated plaintiff’s, which is not necessary for compatibility. Finally, both systems used an error threshold of eight, even though subsequent work by plaintiff had indicated that six would be better.

Both systems used an “H-matrix” to detect transmission errors. Although there are 32 ways in which such a matrix can be configured, defendant’s was an exact duplicate of plaintiff’s. The court also viewed as significant the fact that both systems had inverse H-matrix’s, even though plaintiff’s program included one only because of mis-communication among its developers. Further, plaintiff’s H-matrix was loaded into the software code in reverse order because plaintiff’s micro-processor more efficiently decrements than increments. Defendant’s micro-processor is equally efficient at incrementing and decrementing; nevertheless, defendant’s H-matrix was loaded in reverse order.

In addition, plaintiff’s software contained three lines of code that were completely unnecessary; precisely these same three lines were found in defendant’s program at exactly the same location. And an error in the “select call prohibit” feature was written into plaintiff’s program; defendant’s program featured the same error.

As evidence of dissimilarity, defendant offered side-by-side, line-by-line comparisons of its code and plaintiff’s, both translated presumably into Hitachi hexadecimal language. In such form the programs clearly were not line-by-line duplications. Because of differences in addressing features between the micro-processors, a given step may take a different number of commands in one language than in the other. Accordingly, literal translation of plaintiff’s Intel instructions into Hitachi language would skew the program so that line-by-line comparison became meaningless. In Whelan, the court compared plaintiff’s program written in EDL language with defendant’s written in BASIC language. Despite finding that these were not line-for-line duplicates that court found copying, noting it would be very difficult to translate literally a program written in one of those languages to the other. Although Whelan dealt with source language, the Johnson court saw the same considerations applying to micro-processor hexadecimal representations.

Defendant further argued that plaintiff’s program was not protectable by copyright because derived in large part from pre-existing public domain material. In particular, it contended that the H-matrices, Barker Code, error look-up tables and Barker correlation tables had long been in general scientific use. Nevertheless the court found that the degree of originality was far from trivial. Plaintiff contributed something “recognizably its own.” The test of originality is a “modest” one. Further, the mere fact that component parts of a work were not original to plaintiff does not preclude originality in the combination of those components as a separate entity.

The court viewed as a threshold question the issue of what characteristics of plaintiff’s work were entitled to copyright protection. Defendant contended it copies only the idea of a mobile radio compatible with the plaintiff’s system, and not the physical embodiment. Defendant argued that to make its radio compatible, it had to conform its software to certain aspects of plaintiff’s and had to use the same Barker code, decoding Matrices, overall program structure and data tables. Defendant argued that this compatibility requirement limited the number of expressions of the idea so that idea and expression merged.

The court stated that the “bottom line” “in a computer software context was whether the copyrighted instructions were the sole means of accomplishing a given task …”449 The court also noted that had defendant contented itself to observing the general outlines of plaintiff’s programming and then doing its own coding through its own imagination and independent thought, there would be no infringement. It adopted as typical the procedure for writing a computer program explained in SAS Institute,450 noting that at some point the broader general statement merges into the expression but that it need not pinpoint the exact line of demarcation because the facts here clearly fall on the side of expression.

The court found that the record demonstrated that compatible programs could have been written without verbatim duplication. For example, the H-matrix could have been configured in any of 32 different ways. The Barker code had to be identical, but identity of Barker word correlation techniques and sampling rates were not necessary. Moreover, defendant did not have to duplicate plaintiff’s inverse H-matrix or sample error table. Defendant’s compatibility with plaintiff’s radio did not excuse its copying because copying plaintiff’s code was not the only means to create a compatible program.

The first case to deal squarely with the issue of infringement of a copyright on a computer program in software or firmware form seems to be Data Cash.451 Plaintiff hired a third party to create a program for a computer chess game; the ROM embodying the program was then manufactured for plaintiff, who electrically integrated the ROM into the chess game and, beginning in 1977, marketed the game. Plaintiff subsequently learned that its manufacturer was manufacturing an identical ROM for another chess game, first marketed in late 1978. Plaintiff placed no copyright notice on its ROM, game, packaging, or instructions, but notice did appear on all copies of the source program (used internally by plaintiff), for which a copyright registration was secured in November 1978. In early 1979 suit was commenced for copyright infringement and unfair competition. Plaintiff moved for a preliminary injunction, and defendant moved for summary judgment on the copyright claim.

Hence, a misunderstanding of the technology may have led the court to confuse programming phases with programming devices.454 Specifically, the court confused the object program with the ROM in which it was embodied, although the court correctly concluded that it is only in object form that the program interacts directly with the hardware.455

“Thus, at some point in its development, a computer program is embodied in material form and becomes a mechanical device which is engaged in the computer to be an essential part of the mechanical process. At different times, then, a given program is both ‘source’ and ‘object.’ The ‘source program’ is a writing while the ‘object program’ is a mechanical tool or machine part.”456

The district court decided that the action was properly brought under the 1976 Act,457 but that by virtue of § 117458 the “1976 Act itself does not apply.”459 Accordingly, the court held that if, as of the time of the alleged infringement, the program had been neither published nor registered, then the common law rule should apply; otherwise, the 1909 Act would control. The court then held that under either standard a ROM is not a copy.460 Under common law, the court held that a copy must appeal to the eye.461 Likewise, the court held that under the 1909 Act, a 1908 case462 implicitly adopted by that Act precluded the ROM from being a copy for lack of appeal to the eye. “In its object phase, the ROM, the computer program is a mechanical tool or a machine part but it is not a ‘copy’ of the source program.”463

The court also read dicta in Synercom, supra, to the effect that a program was not a copy or version of a general statement of the problem which it solved, as supporting its conclusion that a ROM could not be a copy. “The program and statement are so different, both in physical characteristics and in intended purpose, that they are really two different expressions of the same idea, rather than two different versions of the same expression.”464

The court of appeals affirmed the denial of plaintiff’s motion for a preliminary injunction, and the grant of defendant’s motion for summary judgment on the copyright claim, but (at least purportedly) disagreed with the trial court’s reasoning.465 Although adopting the lower court’s definition of an object program, the appellate court stated it was unnecessary to reach the issue of whether the ROM was a copy, ruling that when it entered the public domain the work lacked proper notice and, under the applicable 1909 Act, copyright was forfeit.

The game was sold without restriction to the general public beginning in 1977, in which year over 2500 were sold. Readout from the ROM would contain no notice, although plaintiff could have caused such a notice to appear; and placing notice on the game packaging, or instructions (on none of which notice appeared) would have posed no difficulty.466

Although plaintiff argued that the 1976 Act controlled on the issue of publication, the court found otherwise. Although the 1909 Copyright Act, and the common law, would apply if publication occurred prior to 1978, plaintiff argued that the 1976 Act applied because there was no evidence anyone saw a printout of the program prior to 1978. However, the court held that the 1909 Act applied because under that Act publication occurred when the copyright proprietor “sacrificed control of the work such that anyone who wished to view the work could do so, not … when the public actually viewed the work.”467 In selling over 2500 copies in 1977, plaintiff surrendered control. Under the 1909 Act, publication without notice forfeited the copyright, unless it was only a “limited publication.”468

The court held that, had it been marketed in the form of a manuscript, the program would have entered the public domain. Accordingly, it stated that if a limited publication were to be found, it would have to be because of restrictions inherent in the nature of the game. Plaintiff argued that (1) it did not know the ROM could be copied; (2) on learning of the possibility, it took all reasonable steps to prevent duplication and to inform others of its claim; and (3) defendants were not misled by the absence of notice. But the court found a publication, responding that: (1) dedication is a question of law, not intent; (2) the efforts to lock the door came after the horse could have been stolen; and (3) there is no evidence defendants were not misled, and absence of notice is fatal even if no one was misled.

One implicit aspect of the appellate opinion bears mention. While purporting not to consider the issue of whether, under the 1909 Act, a ROM constituted a copy of the program embodied in it, the court—apparently without considering that issue and in the face of its statement that it need not reach the issue—implicitly decided it in the affirmative. In determining that publication occurred prior to 1978, the court looked to Section 26 of the 1909 Act, which defined “date of publication” as “… the earliest date when copies of the first authorized edition were placed on sale, sold, or publicly distributed by the proprietor … .”470 In holding that publication occurred in 1977, the court apparently assumed that the ROMs were “copies.”

Plaintiff also asserted a claim for unfair trade practices and unfair competition by virtue of defendant’s marketing and advertising, as its own, copies of plaintiff’s ROM. In denying defendant’s motion for summary judgment on this claim, the district court noted that misappropriation for commercial advantage was unfair competition, and that INS471 was still the law.472

Franklin473 was ultimately decided on the basis of the patent/copyright dichotomy. Plaintiff alleged copyright infringement474 as to some twelve system programs.475 En route to its ultimate conclusion—that object code embodied in a silicon chip ROM476 might not be protectable by copyright as not constituting a fixed expression of an original work—the district court discussed a number of concepts.477

One major issue was to determine the “original work.” The court noted the distinction between the original work (e.g., a literary work), and the range of copies which may embody it (e.g., book, periodical, punch cards, microfilm, tape recording). “Of signal difficulty in this case is the elasticity of the word ‘program’.”478“It is not clear whether the program-designer’s idea of the operating system program, the source program, or the ROM is the ‘original’ work of authorship. It is not surprising that this should be hard to determine, because at each stage major transformations in the structure of the ‘program’ take place.”479 In the court’s view, plaintiff’s best argument was that the idea of the program was the original work. But then, asked the court, did the designer envision the “architectural structure” of the chip, or did he envision the flow chart?480 In the court’s view, in the former case, the programmer designed a utilitarian aspect, whereas in the latter, he did not design the chip “architecture.”

The court seemed to believe the object code embodied in a ROM might not to be protectable by copyright because of the Constitutional dichotomy between the writings of an author, and the discoveries of an inventor. It saw that division as embodied in 17 U.S.C. § 102,483 and noted that § 101 precluded copyright protection where utilitarian aspects existed and pictorial features could be separately identified and could exist separately.484

Plaintiff claimed that (1) its programs were forms of expressions; (2) an object program embodied in ROM or disc is a form of expression and a work of authorship; (3) a ROM is a tangible medium of expression, and not a mechanical device; and (4) the intent of the 1976 Act and 1980 Amendment was to protect programs in all their forms as a species of literary property. The court stated that CONTU had not concluded whether programs embodied in ROMs were protectable by copyright, that judicial opinion was mixed, and that the legislative history of the statute was unclear.

In dealing with the practice/expression dichotomy, the court looked to Baker v. Selden.485 In Selden, plaintiff held a copyright on a book which described a method of accounting, and included forms necessary to practice the method; plaintiff sought to preclude defendant from copying those forms. That court distinguished practice of the art from its expression: “The object of one is explanation; the object of the other is use.”486 The Franklin court saw Selden as interpreting the division embodied in the Constitution. The court saw the key to subject matter protectable by copyright as whether it teaches, and can be read by, a human. It was unclear, held the court, whether a program in binary code, or one embodied in a ROM, teaches and is meant to be read. “For these reasons, it may be more accurate to say that operating systems487 are an essential element of a machine.”488“If a ROM is found to be a mechanical device, it loses the protection reserved for writings and expression under copyright.”489

On the basis of this analysis, the court concluded that “the question must be: is the expression directed to a human audience? * * * It is the fixed expression of [an ability to create human interaction] that the copyright law protects, and only this.”490 Noting that, pursuant to § 102(a), copyright protection subsists in any work fixed in a medium from which it can be perceived with a machine, the court interpreted this section to mean that the work must be “perceptible to a human audience and not merely perceptible to a machine. Its plain meaning suggests that machine aid is recognized in this section only if it is capable of aiding human perception.”491 Finally, the court distinguished this case from the video game situation, where the program was created for the purpose of creating visual images perceptible to humans. There, the court believed there was an “expression” perceptible with the aid of a machine:

“The ‘original work’ is the visual display that has been fixed, presumably, first in source code and later in object code. No matter how it has been fixed, its purpose from the outset was to generate an image that could be perceived and its goal was to attract and engage a human audience. Such purposes and goals satisfy our conventional expectations of expression.”492

Thus, the test applied seemed to be whether the purpose of the original work was to communicate with a human. If so, that machines may be necessary to render the copy directly perceptible to a human will not adversely affect whether copyright protection is available. But if the purpose of the original work was not to communicate with humans, copyright protection is inappropriate.493 While this distinction probably falls short of the “bright shining line” which lawyers seek, but which may exist only in lawyer heaven, it appears to be workable. The distinction is not unlike that suggested by Professor Melville Nimmer, Vice Chairman of CONTU and author of a definitive work on copyright.494

Concluding that the district court’s doubt as to the copyrightability of programs embodied an erroneous view of the law, the Court of Appeals reversed the denial of a preliminary injunction. Defendant did not dispute copying, and its factual defense was that it was not feasible for it to write its own operating system programs.495 Defendant’s principal legal defense was that system programs were not susceptible to copyright protection.

As to the first point, the Third Circuit found no basis in the statute for distinguishing copyrightability of source code from that of object code. It also found that its own opinion in Williams v. Artic dealt with many of the doubts expressed here by the district court. The 1976 Act states two requirements for a work to constitute copyrightable subject matter: It must be (1) an “original work of authorship”, and (2) “fixed in a tangible medium of expression.” Section 102(a) enumerates seven types of “works of authorship,” and the court found that the legislative history suggested that programs were within the section 102(a) definition of literary works.496

The CONTU Final Report recommended that the law be amended to make it explicit that programs, to the extent they embody an author’s original creation, are copyrightable. The 1976 and 1980 versions of section 117 dealt only with the scope of protection. “The language of the [1980] provision, however, by carving out an exception to the normal proscriptions against copying, clearly indicates that programs are copyrightable and are otherwise afforded copyright protection.”497

The district court’s suggestion that copyrightability depends on a communicative function to individuals emanated from the White-Smith holding that a piano roll was not a copy of a musical composition, because not in a form others generally could perceive. But the appellate court found it clear from the language and legislative history of the 1976 Act that it was intended to obliterate the White-Smith distinction. Moreover, the CONTU majority took the position that object code was appropriate for copyright; this, in the face of a dissent based on the theory that the machine control phase of a program is not directed to humans. The Williams defendant likewise argued that a copyrightable work must be intelligible to humans. The answer here is the same as the answer there: the statute states otherwise.498

Turning to the copyrightability of a program embedded in a ROM, the court found that issue also decided in Williams v. Artic. Defendant there argued that copyright would not apply to a ROM because it was a utilitarian object or machine part. This court held there that “fixation” was satisfied through embodiment of the expression in the ROM. Accordingly, the court reaffirmed that “a computer program in object code embedded in a ROM chip is an appropriate subject of copyright.”499

The court next turned to what it viewed as defendant’s main contention: that system programs500 are excluded from copyright protection under section 102(b), and under Selden. Defendant here viewed Selden as standing for three propositions: (1) use of a system does not infringe a copyright on the description of the system; (2) copyright does not extend to purely utilitarian works; (3) copyright may not be used to monopolize an idea. Defendant made two arguments: (1) an operating system program is a process, system, or method of operation and therefore uncopyrightable; and (2) in the case of system software, idea merges with expression, and therefore no copyright protection is available.

The court agreed with defendant’s contention that, underlying section 102(b) and Selden, is the distinction between property subject to the patent law (which protects discoveries), and property subject to the copyright law (which protects writings). But the court believed that defendant misapplied the distinction. Plaintiff here did not seek to copyright the method which instructs the computer to perform its operating functions, but only those instructions themselves.

Moreover, the court saw defendant’s contention as inconsistent with its concession that application programs are appropriate for copyright protection.501“Since it is only the instructions which are protected, a ‘process’ is no more involved because the instructions in an operating system program may be used to activate the operation of the computer than it would be if instructions were written in ordinary English in a manual… .”502

Defendant’s argument that “an operating system program is part of a machine mistakenly focuses on the physical characteristics of the instructions. But the medium is not the message… .” The mere fact that the operating system program may be etched on a ROM does not make the program either a machine, part of a machine, or its equivalent.503 The court agreed with the CONTU majority which stated that programs should no more be considered machine parts than videotapes should be considered parts of projectors.504

“… where the art [taught by a book] cannot be used without employing the methods and diagrams used to illustrate the book, or such as are similar to them, such methods and diagrams are to be considered as necessary incidents to the art, given therewith to the public; not given for the purpose of publication in other works explanatory of the art, but for the purpose of practical application.”505

The court refused to accept what it saw as the expansive reading given this language by some other courts.506 Defendant’s interpretation of this language precludes copyrightability where the work is put to use. However, the court saw that interpretation as rejected in Mazer: “We find nothing in the copyright statute to support the argument that the intended use or use in industry of an article eligible for copyright bars or invalidates its registration.”507

Another basis for the court’s conclusion was that the statutory definition of a computer program did not distinguish between application and operating programs. The court noted that the district court in Apple v. Formula found no distinction, and that Tandy had upheld the copyrightability of an operating system without discussing the issue.

On the idea/expression dichotomy, the court noted that under Selden, as interpreted in Mazer, “a copyright gives no exclusive right to art disclosed; protection is given only to the expression of the idea—not the idea itself.”508 The dichotomy is embodied in section 102(b), which the legislative history indicates was intended “to make clear that the expression adopted by the programmer is the copyrightable element in a computer program, and that the actual processes or methods embodied in the program are not within the scope of the copyright law.”509 Noting that courts have found difficulty in drawing the line between idea and expression, the court stated that in the case of an operating system, the line must be a pragmatic one, balancing competition on the one hand against protection on the other.

“If other programs can be written or created which perform the same function as an Apple’s operating system program, then that program is an expression of the idea and hence copyrightable. In essence, this inquiry is no different than that made to determine whether the expression and idea have merged, which has been stated to have occurred where there are no or few other ways of expressing a particular idea.”510

“The idea which may merge with the expression, thus making the copyright unavailable, is the idea which is the subject of the expression… . If other methods of expressing that idea are not foreclosed as a practical matter, then there is no merger. [Defendant] may wish to achieve total compatibility with independently developed application programs written for the Apple II, but that is a commercial and competitive objective which does not enter into the somewhat metaphysical issue of whether particular ideas and expression have merged.”511

In Formula513 the same plaintiff filed a similar complaint and sought a preliminary injunction enjoining copyright infringement by another defendant of several system programs embodied in ROMs and diskettes. The copyrights were registered, and plaintiff’s ROMs and diskettes had a copyright notice on, or immediately next to them. Plaintiff asserted that the programs had been copyrighted and published in “either source or object code versions, or both.” Defendant’s programs embodied in ROMs and diskettes were “virtually identical” to plaintiff’s.514

Many of the significant issues were not decided because defendant did not contest them. Defendant did not contend that the extent of copyright protection afforded a program depended on the form in which it was copyrighted, and did not argue that converting a program from source to object code deprived it of any protectibility. Defendant also conceded that: some programs were copyrightable; a program was a “work of authorship”; a ROM was a “tangible medium of expression”; the program embodied in a ROM is “fixed in a tangible medium”; and the expression of a program in source or object code was a “copy.” However, relying on Section 102(b),515 defendant contended, consistent with the doubt expressed by the Franklin district court, that “those [programs] integral to the operation of the machine which do not directly produce visual communication with the user of the machine” are not copyrightable.516

Characterizing copyrightability of programs before 1976 as unclear, the court believed that the 1976 Act broadened the scope of copyright protection. Original section 117 froze the law as to program copyrightability.517 The court saw section 102(b) as retaining the basic dichotomy between idea and expression, which divides the domain of patent from that of copyright. Defendant argued that a system program was an idea, procedure, system, or method of operation. The court rejected that contention, holding that

“[e]ssentially all computer programs as embodied in ROMs and diskettes are designed to operate a machine in such a way as to ultimately produce some useful communication to the user—that is their purpose. It is difficult to understand how they can be classified into two categories for copyright purposes, with protection afforded to one category and not the other, based on whether they directly generate that communication or whether they merely direct certain machine functions which eventually result in that expression. Either all computer programs so embodied are within the terms ‘idea, procedure, system, method of operation’ and are excluded, or all of them are outside those terms and thus protectable. There is nothing in any of the statutory terms which suggest a different result for different types of computer programs based upon the function they serve within the machine.”518

Because of the legislative history of the Act,519 the court had no doubt that the present Act protected all types of programs, however fixed. “… the copyright law … includes protection of all computer programs, regardless of their function or purpose.”520 The court concluded that plaintiff here sought to protect against copying not ideas, but expressions.521

Finally, the court distinguished the Franklin district court opinion denying preliminary relief to the same plaintiff on an identical complaint. At about the time that Franklin was decided, the Third Circuit affirmed a permanent injunction in Williams v. Artic,522 where, according to the Formula court, defendant argued that since a ROM was used to control a machine, it was utilitarian and not subject to copyright protection.523 The Formula court adopted the Third Circuit’s response: The statute defines “copy” to include a material object in which a work is fixed and from which it can be perceived directly or with the aid of a device.524 Accordingly, a preliminary injunction issued.

The court of appeals held that the trial judge had not abused his discretion or relied on erroneous legal premises, or clearly erroneous findings of fact. Defendant contended that system programs, unlike application programs, were ideas or processes, and that a program should be protected only if it embodied expression which is communicated to the user when the program is run.525 The court found the argument contrary to the language of the Act,526 the legislative history,527 and the existing cases. The court stated that the expression/idea dichotomy was intended to “prohibit the monopolization of an idea when there are a limited number of ways to express that idea.”528 The evidence showed that numerous methods existed for writing the programs in question. “Apple seeks to copyright only its particular set of instructions, not the underlying computer process.”

In Australia, this same plaintiff, suing for infringement of copyright on some of the same programs, ultimately was successful.529 Defendant marketed in Australia a computer, manufactured in Taiwan, and containing allegedly infringing ROMs.530 Plaintiff sued for infringement of the copyright on the source (assembly) version;531 also, on three alternative theories plaintiff sought to enjoin copying of the object (hexadecimal) code.

First, plaintiff alleged that the object version was a new and original literary work, distinct from the source version, and entitled to its own copyright. Alternatively, the object version was alleged to be a “reproduction” in material form of the source version. Finally, the object version was said to be an adaptation or translation of the source version. Among the issues found to exist as to the source version, and the object version, were the following: was it an “original literary work” within the meaning of the Australian Copyright Act; and, if so, would the making of ROMs, by defendants in Australia, constitute infringement?

The trial court found it necessary to treat only the first issue, as it concluded that in either source or object form, a program was not a literary work—“something intended to afford ‘either information or instruction or pleasure in the form of literary enjoyment.’ ”532 “The function of a computer programme is to control the sequence of operations carried out by a computer… . a contrast may properly be drawn between something which is merely intended to assist the functioning of a mechanical device and literary work so called. The position is even stronger in the case of the object programme … .”533 The trial court dismissed contrary United States cases as inapplicable because based on the 1976 and 1980 amendments to United States copyright legislation “specifically dealing with copyright protection for computer programmes.”534 However, the appellate court reversed, and found for plaintiff.535

In Tandy,536 plaintiff alleged that defendant’s duplication of a ROM embodying plaintiff’s copyrighted computer program—in object form—constituted infringement.537 Plaintiff manufactured the Radio Shack TRS–80, a computer designed for home use and embodying in a ROM an “input-output routine”: “a computer program which tells the computer how to take the information which is put into the computer by operator in one computer language and translate that information into a more simplified ‘machine language’ which the computer can understand.”538 Plaintiff claimed that defendants copied this program and used it in their own computer designed for home use, the PMC–80.539 Defendants moved to dismiss the claim of copyright infringement, contending that a ROM did not constitute a copy of the original program pursuant to the copyright law, and that, accordingly, duplicating a ROM would not be infringement.

There was no dispute that the 1976 Act was the applicable law540 and the court held, pursuant to Sections 101 and 102, that (1) a program is a “work of authorship” subject to copyright,541 and (2) the ROM is a “tangible medium of expression” such that an embodied program is subject to the copyright laws. In reaching this first conclusion, the court looked to the legislative history, which it read as indicating that “Congress understood that computer programs were subject to copyright protection under the law as it existed prior to the 1976 Act, as well as under the new statute.”542 The court also noted that § 102(a) provides that works can be “fixed” in “any tangible medium of expression, now known or later developed, from which they can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device.” The court held that imprinting a ROM with a program fell within this definition.543

The court next addressed the effect of the original section 117 of the 1976 Act. Defendants contended that this section required use of the pre-1978 standard to determine whether a ROM was a “copy.” Plaintiff argued that (1) a post-1978 standard applied; (2) defendants may have made their ROM, not by duplicating plaintiff’s, but by copying plaintiff’s printed program statements and imprinting those on a chip; and (3) pre-1978 law would permit a holding that a ROM is a “copy.”544

Finally, the court noted that “unauthorized duplication of a visually displayed copy of the program” would fall within the copyright laws. Plaintiff’s suggestion that the chip may well have been created by first copying a printout, and then imprinting that on a chip, provided an independent basis for denying the motion to dismiss.

GCA v. Chance547 presented an instance where a company sued two ex-employees for infringing copyrights on its diagnostic system548 and its operating system programs. On plaintiff’s motion for preliminary injunction, defendants contended that, because plaintiff had registration certificates only on its source code (which defendants characterized as unpublished), the copyrights did not protect his object code (allegedly published). Defendants also alleged that plaintiff deceived the Copyright Office by failing to disclose the publication of the uncopyrighted object code.

The court rejected these arguments, noting that in the registration process the Copyright Office directed plaintiff to furnish information on source code only. The court also asserted that the source code was copyrightable as a work of authorship fixed in a tangible medium from which it can be perceived. The court declined to follow the Data Cash district court’s ruling that a ROM is not a copy. “Because the object code is the encryption of the copyrighted source code, the two are to be treated as one work; therefore, copyright of the source code protects the object code as well.549

A prima facie case of infringement having been established, the burden shifted to defendants to rebut validity. Defendants alleged prior publication and lack of notice, but as a factual matter the court saw no merit to these defenses. Nor did the court accept defendant’s contention that plaintiff committed fraud on the Copyright Office; plaintiff complied with the regulations in filing the application materials on source code. Finally, defendants contended that Section 117 gave them a license to copy. The court rejected this because defendants were not rightful “owners” of the program, and the copies they made and used were not for the purposes set forth in Section 117. Accordingly, the court enjoined defendants from copying the operating system programs.550

Hubco551 was a suit in which defendant charged copyright infringement, inter alia. Defendant manufactured computer hardware, and developed software (including the operating system, known as “BOSS”) for its hardware. By means of certain “governors” embodied within it, BOSS controlled the amount of disc storage, and the number of peripheral devices, available to the user. When a user desired to increase (“upgrade”) the capability of his system, defendant “upgraded” the user’s version of BOSS to permit this. BOSS included a coding algorithm which “scrambled” the output when BOSS was printed out. This made it more difficult for others to determine how to remove the governor and effect an upgrade.

Plaintiff had a procedure (“Nilsson Method II”) for identifying and removing the governor. Alleging that plaintiff’s Nilsson II was obtained improperly, and relying in part on its copyright claim, defendant sought a preliminary injunction restraining plaintiff from using Nilsson II.552 In Nilsson II the location of the governors is ascertained by comparing object code printouts of operating systems embodying different governors;553 once located, the governors can be removed. Although defendant apparently created the high capacity printout, the opinions are unclear as to how plaintiff came into possession of the means for doing so. Initially, plaintiff performed its upgrades by sending employees to effect the Nilsson II procedure at the user’s site; subsequently plaintiff developed a software package which itself performed the procedure.

Defendants had certificates of copyright registration, apparently on the object code of a high capacity version of BOSS. Relying in part on its copyright claim, defendant sought a preliminary injunction restraining plaintiff from using its procedure.554 Specifically, defendant alleged that in achieving the scrambling algorithm, plaintiff infringed the copyrights.555 The court assumed arguendo that the code embodying the scrambling algorithm was copyrightable and properly copyrighted. But it found there was not a reasonable probability that defendant would be able to show copying.556

Subsequently, relying in part on its copyright claim, defendant once more sought a preliminary injunction restraining plaintiff from using Nilsson II, in either an on-site procedure or a program package, to upgrade BOSS. Characterizing its earlier decision as directed to whether plaintiff copied the scrambling algorithm, the court here directed its attention to whether plaintiff’s conduct concerning Nilsson II infringed defendant’s copyright in the entire operating system. The court saw this as embracing three questions: (1) was the operating system object code copyrightable?; (2) was the object code of BOSS properly copyrighted?; and (3) did plaintiff’s conduct infringe?557

Plaintiff challenged the originality of that portion of the BOSS object code which controlled the size of the memory and the permissible peripherals, averring that defendant’s techniques were taught academically, and used in industry, for many years. The court rejected this as irrelevant; there was no evidence of defendant’s copying, and much testimony concerning its independent development of object code.

Plaintiff argued that under § 102(a) the object code was not copyrightable because: (1) “it is a process;” (2) it was part of a machine; (3) it was an internal process not communicated to humans;558 and (4) it was not fixed in any tangible medium of expression. The court’s opinion intertwines these arguments, and, following its view of Williams v. Artic, supra, and the CONTU Report, it held object code copyrightable under § 102.

Plaintiff also alleged any copyright had been lost by public distribution without copyright notice. The court distinguished general (and possible divesting) publication from a limited (non-divesting) publication, defined as one for limited purposes, to a selected group. By virtue of the use restrictions in defendant’s license agreements, the purpose was found limited; because BOSS was offered only to owners of defendant’s hardware, there was a reasonable probability it was published only to a selected group.

Finally, plaintiff argued his use of Nilsson II did not involve copying object code, and therefore did not constitute infringement.559 The court disagreed, finding, as to the on-site modification, a reasonable probability that in making a printout of the high capacity unscrambled object code, and reproducing a portion of that in the user’s system, there was copying. As to the program package, the court found there was likewise a reasonable probability defendant would show copying. The court relied on Professor Nimmer, who saw the 1976 Act as reading out of the law the requirement that a “copy” must be perceptible without the aid of a machine, and who viewed inputting to a computer as creating a copy. The court stated that plaintiff’s software package contained560 a copy of some object code from a high capacity version of BOSS.561

Plaintiff’s game had four “instruction” ROMs used by the microprocessor to make game decisions (e.g., where the characters move); it also had two “character” ROMs which contained information that caused the character images to appear on the screen. Plaintiff had a copyright registration for the audiovisual effects of the game, and another for the programs embodied in the six ROMs.563

Defendant’s kit contained four ROMs to be substituted for plaintiff’s instruction ROMs, and one ROM to be substituted for one of plaintiff’s two character ROMs. Defendant’s characters bore no resemblance to plaintiff’s, and the nature of defendant’s maze, especially at advanced stages, was different from plaintiff’s.

The court rejected defendant’s argument that the visual display and the computer program were so “intertwined” as to preclude their separate consideration. The court found the program to constitute a distinct creation. The court noted it was possible to design a game that would infringe Midway’s audiovisual copyright, but would use a different program. The court also noted it was possible for a program to infringe without the audiovisual display infringing. Moreover, the skill required to design the program was different from the process of conceiving and designing the distinctive characters.564

The court stated that availability of copyright protection for ROMs registered in 1981 was governed by the Copyright Act amendment of 1980. The court stated that even prior to the 1976 Act, high level instructions written in human readable source code were eligible for copyright protection, and that the legislation was intended to protect object code as well as source code. Hence, the reference in § 101 to instructions to be used “directly” or “indirectly.” Only object code can be used directly. According to the court, the legislative history for the 1980 amendment confirms that protection for object code was intended.

And several courts have held that since copyright protection subsists in original works of authorship which can be perceived with the aid of a machine, § 102(a), machine readable object code is protectable. “Whatever its merit as an original matter, the 1976 Copyright Revision Act has foreclosed the argument that object code—meant to be read by machines rather than humans and incomprehensible to any but highly trained computer specialists—is not a proper subject for copyright protection.”565 The object code is nothing other than a direct transformation of the program composed in source code. To allow protection of source but not the “mechanical implementation” of the same program would not be logical.

Defendant argued that object code embodied in a ROM is distinguishable from object code embodied in disc or tape in that the ROM is a form of computer circuitry or hardware, and therefore a utilitarian object for which copyright protection is not available. The court rejected this argument, noting that a ROM is not strictly hardware (“a purely mechanical component of a computer”) or software (“a set of instructions external to the machine”). It is “firmware.”

The court saw no reason why object code stored on a chip might freely be copies, while the same code stored on a tape or disc might not be. Plaintiff’s copyright of the object code stored in ROM was held valid. Access being indisputable, the court further found that overwhelming evidence supported the conclusion that defendant’s ROMs were substantially similar to plaintiff’s.566 “The high degree of identity of the two programs and the virtual impossibility that two programmers working independently could write so nearly identical programs compelled the conclusion that” defendant’s ROMs infringed the programs stored in plaintiff’s ROMs.

Defendant argued that since the full program of plaintiff was stored on six ROMs, in order to find infringement, the court must find substantial similarity of all five of defendant’s ROMs to six of plaintiff’s ROMs. But this was unnecessary, for it is infringement to copy one chapter of a novel. The fact that defendant did not copy the data stored in two of the ROMs does nothing to mitigate his piracy of the portions of the program stored in the other ROMs.

An amicus raised the argument that plaintiff’s first sale exhausted its right to control use or alteration by owners of the “copyrighted portions of those machines.” Support purportedly came from a line of patent cases holding that a patentee cannot prevent a lawful owner from repairing his machine. According to the court, the only copyright act provision arguably pertinent is § 109, which gives the owner of a copy the right to resell or display that copy. But § 109 does not authorize adaptation or reproduction of a work.567

The court felt that the reasoning of the Seventh Circuit in Midway v. Artic, supra, (which also concerned speed-up kits) was also on point. The Seventh Circuit reasoned that, because of increased demand for speeded-up games, the sale of the ROM should not be construed as conveying the right to speed it up. That court held the speed-up kit to be a derivative work and held that licensees who installed it would be direct infringers, while the manufacturer was a contributory infringer.

BPI568 presented a situation where, in developing an accounts receivable system, plaintiff contracted with defendant to produce some twelve programs embodied in the system. Plaintiff alleged that defendant infringed its copyright (1) on these twelve programs, and (2) on other programs developed by plaintiff as part of its system.

In Williams v. Artic570.1 the district court held that a ROM embodying the program was copyrightable subject matter under the 1976 Act, as amended.570.2 On appeal, defendant argued ROMs did not qualify for copyright protection because they were utilitarian objects or machine parts. But the court distinguished between the ROM itself (not copyrightable) and the “artistic expression in original works which have met the statutory fixation requirement through their embodiment in the ROM devices” (copyrightable)571 Pointing out that this was the first video case to deal with copyright in the program (as opposed to solely the display), the court noted that the defendant conceded a program could be the subject of copyright as a literary text. However, defendant contended that a ROM was not a copy, and that a program copyright was not infringed when a program was loaded into a ROM, which is a utilitarian device. Defendant argued there is a distinction between the source code version (copyrightable) and the object code version because a “copy” noting that perception via machine is expressly countenanced572

There are now cases interpreting Section 117. In Formula573 in a decision affirmed by the Ninth Circuit, the district court had previously enjoined defendant from copying defendant from copying plaintiff’s copyrighted programs or distributing or selling any chips, ROMs, or diskettes containing a copy of plaintiff’s copyrighted programs.

Section 117 states it is not an infringement for the owner of a copy of a program to create another copy as an essential step in using the program to create another copy as an essential step in using the program in conjunction with a machine, provided it is used in no other manner. Defendant argued that as the rightful owner of the diskette containing a copy of Apple’s programs, defendant is the owner of a copy of the program within the meaning of 117. It further argues that making the chips on which plaintiff’s copyrighted programs are recorded constituted activity permitted by 117 in that the chips are “a new copy … created as an essential step” in using the program in conjunction with a computer.

Many companies have developed original application programs which they record on diskettes and sell. Apple has licensed certain of these concerns to copy its system programs and record those on the same diskettes. The programs are meant to be used together. The diskettes in question here contain both original programs of a third party and Apple’s copyrighted programs. The third party clearly has a right under its license with Apple to sell these diskettes to defendant, and defendant had a right to buy and resell them. Apple’s contention is that defendant commits contempt because it copies Apple’s programs in permanent form onto chips and thereafter sells the chips. Accordingly, as a matter of first impression the court was called upon to construe section 117.

The court looked to the legislative history, and in particular the CONTU Final Report which points out that programs written on paper cannot be used in a computer without being translated, and that programs on diskettes cannot as a practical matter be used without first being loaded (i.e., “copied”) into a memory device within the computer. The Report stated that section 117 was recommended to facilitate use of programs by allowing such copying.

Stating that the rightful possessor of a copy of a program should be permitted to load it into a computer, CONTU went on to stress the limited nature of the right to copy granted by section 117: only “to that extent which will permit its use by that possessor … .” This limited copying permit embodied in section 117 was recommended to CONTU in a report by NSF.

Accordingly, the court held: only an owner/user of a computer who rightfully owns a copy of a copyrighted program is authorized to make another copy and even then only to the extent necessary to enable his use of that program in his computer; the new copy must be made only for the owner/user’s internal use and must be destroyed when the original copy is resold; and the new copy cannot be made accessible to others.

Here defendant is not an owner/user of a computer. Nor can its copying of plaintiff’s programs be construed as “internal use” by defendant. Nor has defendant confined its use of the copy of its own purposes, but rather has made those programs available to others. Accordingly, defendant is not within section 117.

As an alternative ground for finding contempt, the wording of section 117 implies that the type of authorized copying must be no more permanent that reasonably necessary. Here, defendant’s method of copying is more permanent than necessary because the diskette could be used in defendant’s computer by use of a copying method less permanent than the one employed. The diskette could simply be copied into RAM, which would only be a temporary fixation. When the computer is turned off, that RAM copy is lost.

Moreover, defendant’s legal justification for its manufacture and sale of the chips was mere subterfuge. And the court did not believe that section 117 should be construed so as effectively to emasculate the protection for programs contained in other sections of the Copyright Act. Accordingly, it found defendant to be in contempt.

Synergistic v. IDB573.1 was another case which raised the issue, under § 117, of whether the defendant owned a copy of the computer program in question. The defendant had contracted with plaintiff to develop that software, and the plaintiff had delivered a copy to defendant. In defendant’s motion for summary judgment on a copyright infringement claim, the court found nothing dispositive of the ownership issue in the contract between the parties, but did find the answer in the Uniform Commercial Code.

The court first held that, because “goods” “in the form of computer hardware and software” predominated over services in this transaction, the transaction was covered by the UCC. Section 2-401 of the UCC provides: “Unless otherwise explicitly agreed[,] title passes to the buyer at the time and place at which the seller completes his performance with reference to the physical delivery of the goods.” In this instance, there was no explicit agreement between the parties on this point. Accordingly, the court held that title to a copy of the software passed to defendant when plaintiff delivered the copy to defendant, and that defendant therefore had a right to engage in the limited copying and adaptation permitted by 17 U.S.C. § 117. Because plaintiff failed to submit evidence creating a genuine issue of fact that defendant’s conduct fell outside the bounds of propriety prescribed by that statute, the court granted summary judgment.

The court found that in determining whether the mileage data was copyrightable the most important question concerned that effort involved in compiling it. Although the data base was completely different in form from plaintiff’s tables and map mileage segments, this drastic change in format would not alone prevent a finding of infringement. While courts often offer similarity of arrangement as the basis for finding compilation infringement, cases are rare in which a copier escapes liability for substantial copying solely because he changed the form of the compilation. The small significance of arrangement has been largely recognized in compilations. The creativity or effort that engages the machinery of copyright lies not in the arranging but in the compiling. “Nowhere are deficiencies inherent in the traditional approach more apparent than in the context of computer data bases Data bases are clearly permissible subjects for copyright under the 1976 Act, yet it is often senseless to seek in them a specific, fixed arrangement of data.”575

The court must here also assess the effect of copying the data into a computer data base. The question is whether such input constitutes copying. Nimmer states that, at least after 1980 amendment, input of a work into a computer results in copying so as to infringe. Before the 1980 Amendments inputting may not have been copying under the 1976 Act. At a pretrial conference the parties here agreed that inputting a copyrighted work into a computer data base could be an infringement under the 1976 Act, and that former section117 did not limit the protection against such inputting to that provided by the 1909 Act. [Query as to whether present section 117 deals at all with data bases, inasmuch as its express terms speak of computer programs, a term whose definition in the act would not seem to include data bases.]

“Under current law, it is clear that inputting a copyrighted work into a computer would violate the copyright holder’s exclusive rights … as the tabular data may be perceived by way of machine … it would seem to be within the definition of copy. The data from the maps can be analyzed similarly… . The way that information must be formatted in order to be of use by a particular computer or program should not prevent a finding of infringement. The irrelevance of the arrangement of data in the context of a computer data base is highlighted by the fact that the [mileage guide] is marketed as a computerized version” of plaintiff’s mileage guide.

The Eleventh Circuit has recently held that production, promotion and distribution of pirate computer chips for the purpose of allowing the unauthorized descrambling and viewing of satellite pay TV signals constitutes copyright infringement.576 The court found that since the defendant demonstrated and promoted the pirate chip, there was irrefutable evidence of direct copyright infringement.

(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.577

Regarding the first factor, the court found that not only did the defendant have a commercial purpose in producing and selling the pirated chips, but he encouraged others to make a profit by selling the chips. Since the purpose of the first factor is to prevent the user from profiting from the exploitation of copyrighted material without paying for it, this factor worked against the defendants.

The fourth factor, the effect on the potential market for or value of the copyrighted work, was termed as “undoubtedly the single most important element of fair use.”578 The unauthorized reception of scrambled pay TV without payment clearly injures both the present and future customer markets and deprives the plaintiffs of deserved profits. Finding that none of the four factors supported the defendants, the court held that the fair-use defense did not preclude summary judgment for the plaintiffs on the claim of direct copyright infringement.579

There are two decisions of interest on whether distribution of software under a confidentiality agreement absent required copyright notice constitutes publication so as to divest copyright. In UNIX System v. Berkeley Software,580 the plaintiff forfeited copyright in a valuable computer program. Plaintiff was an AT&T subsidiary and the assignee of AT&T’s rights in the allegedly proprietary UNIX operating system. After creation of UNIX software by AT&T, AT&T and the University of California (“Berkeley”) agreed to foster the development of UNIX software. AT&T granted Berkeley a right to prepare derivative works. The agreement provided that AT&T had no ownership of any portion of such a derivative work that was not part of the unmodified program, while Berkeley was free to license derivative works that contained no AT&T proprietary information.

Berkeley created derivative works and licensed them to other AT&T licensees; later Berkeley licensed redacted versions of the derivative works more widely. Berkeley alleged that the redacted versions (“Net2” software) contained no AT&T proprietary information because it had been developed without any AT&T proprietary code.581

The court found that the Net2 software contained a very small proportion of UNIX code.582 The court explained that similarities in names were functionally irrelevant to the internal program. However, in interfacing with other programs, the names could be important for compatibility. Defendant contended the similarities resulted from publicly available code, copied comments, or compatibility requirements. Plaintiff’s expert disagreed.

Plaintiff sued, inter alia, for copyright infringement and trade secret misappropriation, and moved for a preliminary injunction on both of those claims. In analyzing the likelihood that plaintiff would prevail on its copyright claim, the court focused on the issue of publication. Published in 1978, the UNIX software was subject to the copyright notice provision originally embodied in the Copyright Act of 1976. Publication constituted “the distribution of copies … of a work to the public by sale or other transfer of ownership, or by rental, lease or lending …”583

The software had been distributed by plaintiff without copyright notice to thousands of licensees. The applicable statutory savings provision listed three alternative bases for preserving copyright in the event of publication absent copyright notice: that a relatively small number of copies lacked notice; registration within five years of publication plus reasonable effort to add notice to copies already distributed; or proof that the omission was due to the failure of a third party who thereby violated an express written agreement. Since plaintiff’s conduct fit within none of these, plaintiff would have been deemed to lose its copyright if it had published.

The court turned to the common law doctrine of limited publication, which provided that where an author “communicates the contents of a manuscript to a definitely selected group and for a limited purpose, and without the right of diffusion, reproduction, distribution or sale …”584 there is no publication so as to divest copyright. The court held that this doctrine survived the Copyright Act of 1976. On the instant facts, the most important of these criteria was that of communication to only a “selected group.”

The court held that a “selected group” had to be more restrictive than simply customers who desired to use the work; otherwise all publications would be limited. Conceding that courts had expanded this concept in applying it to software, the court held the expansion applied only to ancillary software that supported the sale of hardware. The court further held that the UNIX software did not qualify.

The final question on the publication issue was whether plaintiff would license all applicants. The court found that the plaintiff’s screening filtered out few potential licensees, and that the group to which the software was offered was not “selected.” Accordingly, the court denied plaintiff’s motion for a preliminary injunction based on copyright infringement.

In another case dealing with publication, D.C.I. Computer Systems, Inc. v. Pardini,585 the Ninth Circuit affirmed the district court’s dismissal of a copyright infringement claim because the work was “published more than five years prior to the date of its copyright registration.”586 The court recognized the doctrine of limited publication where (1) the work is distributed to only a definitely selected group of people; and (2) only for a limited purpose, without right to reproduce, distribute or sell. However, plaintiff had not selected a particular group and offered to distribute the work to them. Moreover, it did not meet the limited purpose prong, since its purpose in distribution was financial. Since the work was published, without copyright notice on any copies, more than five years before registration, it passed into the public domain under 17 U.S.C. § 405(a).

Data General v. Grumman587 dealt with whether a copyright infringement plaintiff can obtain relief if it does not furnish to defendant, in response to a discovery request, a copy of the source code of the program in which copyright is allegedly infringed. Plaintiff Data General alleged copyright infringement and trade secret misappropriation. Before trial, defendant moved to dismiss the copyright claim because plaintiff did not comply with defendant’s request to produce the source code of all versions of the program that were released. Plaintiff maintained that it did not maintain such code in the ordinary course of business, and that it would be unduly burdensome to reconstitute the code. Defendant argued that without the original source programs plaintiff could not establish a link between the object code copied by defendant and the “source code programs registered with the Copyright Office.”

The trial resulted in judgment for plaintiff on the issues of copyright infringement and misappropriation of trade secrets. In a post-trial motion, defendant asserted that plaintiff had not proved that the programs concededly copied by defendant were the programs in which plaintiff held copyrights. In particular, defendant argued that, since plaintiff had not produced source code for any of the versions, plaintiff could not prove infringement of “the source code programs registered with the Copyright Office.”588

But the court saw this as irrelevant, explaining that the deposit does not define the substantive protection afforded by the copyright. It is the work that may not be copied, not simply the particular expression embodied in the deposit. Source code and object code are two representations of the same work. Plaintiff could show infringement by showing copying of either form. The court found there was abundant evidence of, and an admission of, identical copying of the object code. Copying of the work could be shown by copying of object code or of source code.

Defendant also argued that, even conceding that there was copying, (1) there was no proof that copyright-protected aspects of the program were copied, and (2) it would be impossible to show copying of copyright-protected aspects without the source code.589 The court held that such an exercise was necessary only to show substantial similarity, and that this issue was not present here. Here, there was direct copying of the content of code, and no evidence of defendant’s independent creation.

The court here seemed to confuse the two connotations of “substantial similarity.” Noting that plaintiff must show actual copying either by direct evidence thereof, or by the combination of access plus substantial similarity, the court viewed the Altai test of substantial similarity as applicable to the issue of actual copying. However, the Altai test is meant to distinguish protectable expression from matter not protectable by copyright, a different issue from whether the similarity is such that actual copying must be inferred. Thus, the court confused the issue of (1) whether anything was actually copied, with the issue of (2) whether any copyright-protectable aspects of plaintiff’s program were present in defendant’s program.

Defendant also contested admissibility of the deposition testimony of its employee, who was plaintiff’s ex-employee, wherein he repeatedly invoked his Fifth Amendment privilege against self-incrimination on whether he provided copies of plaintiff’s software to defendant. The jury was permitted to draw an adverse inference against defendant as a result of that invocation. But the court concluded that where defendant was aware of the activity and profited from it, the jury could properly be permitted to draw such an inference.