Full opinion text
ORDER WALKER, District Judge. This order deals with a series of motions concerning ownership and protectibility against unauthorized copying of the visual displays associated with the Macintosh computer. Previous orders have described the earlier proceedings which led to the present motions. A brief summary of those proceedings will suffice here. I. Apple Computer, Inc. (“Apple”) filed this copyright infringement action on March 17, 1988, against Microsoft Corporation (“Microsoft”) and Hewlett-Packard Company (“HP”), claiming that Microsoft’s Windows 2.03 computer software and HP’s New-Wave computer software infringed seven copyrights held by Apple. The copyrights at issue protect the audiovisual works that Apple claims for the graphical user interface of its Macintosh computer. The litigation arose out of a dispute whether an earlier version of Microsoft’s software, Windows 1.0, infringed Apple copyrights. Microsoft and Apple sought to put that dispute to rest by an agreement on November 22, 1985 (“1985 Agreement”). By that agreement, Apple granted to Microsoft a non-exclusive license of the audiovisual displays in Windows 1.0. As the court found on March 20, 1989, however, the 1985 Agreement is not a complete defense to this action, because it was limited to the visual displays in Windows 1.0 and did not cover displays in Windows 2.03 that were not in the prior work. Apple Computer, Inc. v. Microsoft Corp., 709 F.Supp. 925, 930 (N.D.Cal.1989). To allow the court to determine which allegedly infringing Windows 2.03 and NewWave visual displays were not contained in Windows 1.0, to facilitate comparison of the works in suit and to give the court and the parties a means to determine the scope of copyright protection, Apple was asked to submit a list of alleged similarities between its works and the works of defendants. Apple filed this list on April 7, 1989. The list contained 189 alleged similarities between the Apple audiovisual works and Windows 2.03 and 147 similarities between the Apple works and New-Wave. The court then determined that under the 1985 Agreement 179 of the similarities claimed to be in Windows 2.03 were licensed, Apple Computer, Inc. v. Microsoft Corp., 717 F.Supp. 1428 (N.D.Cal.1989), and that the agreement would cover all but 54 of the similarities alleged to be in New-Wave, assuming of course that HP establishes that Microsoft in turn licensed these features to it, Apple Computer, Inc. v. Microsoft Corp., 759 F.Supp. 1444 (N.D.Cal.1991). After filing a supplemental complaint on June 28, 1991 that extended its claims to cover Windows 3.0 and NewWave 3.0, the updated versions of the Microsoft and HP works, Apple then filed two supplemental lists of similarities, incorporating the prior lists by reference. The second of the supplemental lists was filed on August 15, 1991 (“Second Supplemental List”). In the meantime, the parties engaged in lengthy and extensive discovery which concluded on January 31, 1992. There followed a deluge of summary judgment motions: two by Apple, seven from Microsoft and nine by HP. The court directed that these motions be taken up in three hearings on the basis of their legal and factual issues. The first was set to encompass all questions of originality, functionality, merger, indispensable expression and scénes á faire; these issues appeared in the Apple motion “as to Substantial Similarity and Certain Affirmative Defenses,” the separate Microsoft motions to dismiss claims against Windows 2.03 and 3.0, the HP motions on lack of originality, limited scope of protection, and items excluded from copyright protection, and Microsoft’s “Motion for Partial Summary Judgment Dismissing Apple’s Claims Against Windows 2.03 and 3.0 Based on Certain Copyrights.” At a second hearing, the court intended to hear argument on all questions of similarity contained in the motions, as well as the remaining issues in Apple’s motions, and the Microsoft motion to dismiss the contributory infringement claim, and at a third hearing, the court planned to take up argument on the damages issues: Microsoft’s motions concerning actual damages and indirect infringer’s profits, profits of foreign subsidiaries, and pre-judgment interest; and HP’s motions concerning profits from sale of NewWave, actual damages, unjust enrichment, and joint and several liability. But as often happens with the best laid plans, this scenario was upset when Apple refused to join the issues raised in defendants’ motions. Apple contended that its own lists of similarities are not exact descriptions of any infringing features, but merely examples of the overall similarity of defendants’ works. Sticking stubbornly to a “look and feel” or “gestalt” theory of this lawsuit, Apple was apparently of the belief that these passwords would automatically get its case around summary judgment motions and to a jury, regardless whether any of the visual displays that potentially comprise this “look and feel” are themselves protectible expression. Accordingly, Apple’s response to the detailed arguments against protectibility that defendants made on an item-by-item basis was that, “we do not attempt here to chase every rabbit loosed by defendants’ continuing focus on irrelevant detail.” Apple Mem in Resp to Def’ts Motions at 9-10. Unaided by any effective opposition from Apple to defendants’ motions, the court on April 14, 1992, after analytically dissecting the works in suit as best it could under the circumstances, determined that the 10 remaining items alleged against Windows 2.03 were not protectible under copyright law, and that 53 of the 54 alleged similarities in NewWave were subject to no or little copyright protection. Apple promptly moved for reconsideration and, for the first time, addressed the merits of defendants’ motions. II. As noted in earlier orders, Apple’s Macintosh microcomputer turned out to be one of the major commercial triumphs of the 1980s. Much of that success seemed to rest on the visual displays or images which the Macintosh generated on its computer screens. These proved highly intuitive, facilitated users’ learning of how to operate the Macintosh and introduced millions to the wonderful capabilities for useful tasks which computers offer. The combination of a computer monitor’s visual displays and the user command functions on the keyboard or other input devices is called the computer’s user interface. On the Macintosh, the screen displays include icons or symbols to represent programs or information, pull down menus or lists of commands or information, use of windows to display information and the ability to move, re-size, open or close those windows to retrieve, put away or modify information, and a display of text by a proportionally spaced font in all menu items, title bars, icon names and text directories for a consistent and distinctive appearance. When a computer’s visual displays incorporate significant graphic elements, as does the Macintosh, it is referred to as a graphical user interface (GUI or, in the argot of the trade, “gooey”). Apple came into existence in the 1970s. By late in that decade, Apple had enjoyed some success with its Apple II home computer and was seeking to develop new products. On at least two occasions in December 1979, Apple personnel (including its founder Steven Jobs) visited Xerox Corporation’s Palo Alto Research Center (“PARC”). That facility had for several years done extensive work to make the operations of and images generated by computers more appealing and comprehensible to those lacking technical training or bent. An underlying assumption of Xerox’s product development program was that the users of its products would not be particularly interested in computers themselves, but instead in what tasks or functions computers could perform. An overriding objective of Xerox’s efforts was to make computer images and usage appealing and intuitive. During the visits, Xerox’s Larry Tesler demonstrated to the Apple people Xerox’s Smalltalk software. Smalltalk had moveable rectangular overlapping windows, window tabs that represented files and information, a muted background to make the windows and tabs stand out, point and click manipulation of windows and tabs, and pop-up menus, among other graphical features. Many of the features which this order will discuss at length were developed at Xerox PARC and shown to the Apple personnel on their visits. Smalltalk also made extensive use of a hand operated device called a mouse which functioned as a cursor pointer. The mouse was not original to Xerox, having been developed at Stanford Research Institute, but Xerox— standing on the shoulders of those who preceded it as others would stand on Xerox’s shoulders — extensively employed the device in its user interfaces. A few months after the visits to PARC, Tesler went to work for Apple on its Lisa project, one of the new products Apple was developing. Additional Xerox personnel followed. In 1981, Xerox introduced its 8010 Star Information System, a workstation comprised of even higher performance machines than those on which Smalltalk ran. The Xerox Star made extensive use of icons to organize its functions. The features of Xerox’s graphical user interface, such as use of a mouse-driven cursor, overlapping windows, iconic representation and so on had been made possible by development of bitmap graphics. A computer’s processing unit generates, and a computer monitor displays, images by composing dots on the monitor screen called pixels which are illuminated, shaded or colored to achieve the desired effect. Bitmap graphics allow particular bits of the computer’s memory function to manage each pixel, permitting programmers great flexibility in creating the visual displays actually generated. Programs written for computers having only limited amounts of such memory are forced to base their visual displays on vertically or horizontally defined lines and thus generally are confined to blinking cursors, letters, numbers and similarly limited characters. Such a visual interface is generally termed an alphanumeric or character-based interface and lacks the ability to generate the more complex shapes, forms, artistry and animation associated with a graphical user interface. A user wanting to start a word processing application, for example, in an alphanumeric or character-based interface might type the command “Exec WS” or to create a new file might type “Dup. F. Txt.” A graphical user interface allows the user to see, point to and manipulate graphical images, symbols or words to instruct and interact with the computer to perform these same functions. The ability to employ bitmap graphics and thus the graphical versatility of the Xerox PARC products were at least in part attributable to the configuration of the Xerox systems. The Xerox Smalltalk program demonstrated to the Apple groups in December 1979 ran on Xerox’s proprietary computers which were linked on an Ethernet network. Similarly, in the somewhat later Xerox Star, individual monitors and keyboards were connected via an Ethernet network to file servers, printers and data base servers. By sharing resources in this manner, the Star workstations of that era were able to devote a significant amount of memory to screen displays and user commands. Stand alone personal computers, to at least the early 1980s, were incapable of mustering the operating memory resources necessary to generate visual displays of the complexity and animation of the Xerox products. But, in 1979, Motorola introduced its 68000 microprocessor, a chip that took its title from the number of its components or transistors and could accommodate a vastly greater number of instructions than other chips used up to that time for personal computers. The Motorola 68000 chip could, in a stand alone microcomputer, support the high resolution display and highly interactive graphical interface necessary to emulate the ease of use of the Xerox products, Apple based its new products, the Lisa and Macintosh, around the Motorola 68000 chip. Before Apple could get either of these products to market, however, IBM entered the personal computer field. Theretofore a producer of large mainframe computers, IBM sought to produce an economical product. Introduced in 1981, IBM’s original PC used chips produced by an outside supplier, Intel, whose early microprocessors possessed less memory capability than the Motorola 68000 chips. IBM turned to another outside supplier, Microsoft, to produce system software for the PC. The MS-DOS system software Microsoft devised for the IBM PC was confined by the limited capabilities of the hardware for which it was designed and thus its visual displays employed an alphanumeric or character-based user interface. IBM nonetheless brought enormous financial resources and marketing capabilities to the personal computer field which fairly exploded in the first few years of the 1980s. In 1983, Apple introduced its bulky Lisa which, among other features, had a bitmap video display capable of a dot resolution of 364 lines of 720 dots each, detached keyboard and mouse driven cursor pointer. Apple priced Lisa at about $10,-000, compared to the IBM PC then selling in a range beginning at under $2000. Despite her significantly more sophisticated graphics than the IBM PC, Lisa was a flop. The following year Apple introduced the Macintosh. This small machine had a high resolution bitmapped built-in monitor, screen displays based on the Motorola 68000 microprocessor, detached keyboard and mouse. With a now classic advertising campaign introduced during a half-time break of the 1984 Super Bowl and a significantly more competitive price than the Lisa, the Macintosh got off to a good start despite the heavy competition from IBM. Apple also took a quite different tack to software than had IBM. Apple not only owned the system software usable on the Macintosh but it required third party software developers to employ the Macintosh graphical interface, so that a user could readily go from one application to another keeping the same basic and familiar visual displays and input commands in all applications. By shortening the learning process for users of new applications, this requirement contributed greatly to overall user ease and satisfaction associated with the Macintosh. By contrast, the screen displays and user commands of an MS-DOS based application program designed to run on an IBM PC or an IBM compatible may bear virtually no resemblance to another application program with which the user is already familiar, increasing learning difficulty and user frustration. While Apple’s policy of maintaining uniform applications interfaces for the Macintosh has limited the number of Macintosh applications written, this policy has also helped Apple’s marketing efforts to meld the images of Macintosh hardware and software into a single and “user friendly” whole. In time, more sophisticated microprocessors were developed and increased the memory capacity of microcomputers. Microsoft developed and, in 1985, introduced its Windows 1.0 software. Windows operates on top of DOS and extends its visual or graphical capabilities. The later versions of Windows took advantage of the greater operating memory resources available in more recent non-Apple hardware. These could generate increasingly complex visual displays and animation and thus, with each new version, the visual displays of Windows seemed more and more to resemble those in the Macintosh. IBM and especially IBM compatible hardware has consistently been less expensive than comparable Macintosh hardware. There developed a huge base of IBM and IBM compatible hardware users interested in protecting their investments in such hardware and DOS based software. Because each version of Windows seemed increasingly to offer that which Apple had touted as the special advantage of the Macintosh, Windows proved a hugely formidable force in the software market. Apple responded with this lawsuit. III. The Ninth Circuit employs a two-part test to determine whether a work infringes the copyright in another work. First, the “ideas” of the works in suit are compared for substantial similarity, using an “extrinsic test” or “objective analysis of expression.” Sid & Marty Krofft Television v. McDonald’s Corp., 562 F.2d 1157, 1164 (9th Cir.1977); Shaw v. Lindheim, 919 F.2d 1353, 1357 (9th Cir.1990). Analytic dissection, employing a list of criteria of comparison informed by expert testimony, is a part of this exercise, which makes this well-suited for determination as a matter of law. Krofft, 562 F.2d at 1164. If the ideas are substantially similar, then an “intrinsic test” or “subjective analysis of expression” is used. Shaw, 919 F.2d at 1357. In suits involving literary works, Shaw requires that the intrinsic test be performed by the trier of fact, and not by the court upon a motion for summary judgment. Id at 1359. It should be emphasized that Shaw confines its holding to suits involving literary works, or arguably the literary aspects of works, see Brown Bag v. Symantec, 960 F.2d 1465, 1476 (9th Cir.1992), and does not extend to visual displays. Moreover, the “intrinsic test” entails a comparison of the portions of a work that can be the subject of copyright protection. See Pasillas v. McDonald’s Corp., 927 F.2d 440, 443 (9th Cir.1991). Because the “intrinsic test” is thus limited, a finding of similarity of ideas does not complete the “extrinsic test.” See Melville B. Nimmer and David Nimmer, 3 Nimmer on Copyright, § 13.03[B] at 13.57 (Matthew Bender, 1992). If the similarity of the works in suit stems solely from unprotectible features, then the plaintiffs case is missing an essential element of infringement. Feist Publications, Inc. v. Rural Telephone Service Co., — U.S. —, 111 S.Ct. 1282, 113 L.Ed.2d 358 (1991); Data East USA, Inc. v. Epyx, Inc., 862 F.2d 204, 208 (9th Cir.1988). Under such circumstances, summary judgment is plainly appropriate before even reaching the “intrinsic test.” See Frybarger v. IBM, 812 F.2d 525, 530 (9th Cir.1987). The “non-literal components” of a computer program, including its user interface, are protectible if, “on the particular facts of each case, the component in question qualifies as an expression of an idea [not] an idea itself.” Johnson Controls v. Phoenix Control Systems, 886 F.2d 1173 (9th Cir.1989). In affirming a preliminary injunction against copyright infringement, the Second Circuit observed that the sequence of images on a computer screen “might contain so little in the way of particularized form of expression as to be only an abstract idea portrayed in a noncopyrightable form [citation omitted].” Stern Electronics, Inc. v. Kaufman, 669 F.2d 852, 857 (2d Cir.1982). To determine whether similarities result from unprotectible ideas, analytic dissection of similarities may be required. “If this demonstrates that all similarities in expression arise from use of common [i.e., non-protectible] ideas, then no substantial similarity exists.” Data East, 862 F.2d at 208. In performing this analysis, the Ninth Circuit has instructed district courts to focus on the claimed similarities between the works in suit because a focus on dissimilarities would “distract a reasonable observer from a comparison of the total concept and feel of the works.” Id. Further, in order for the trier of fact to compare the expression in an “intrinsic test,” there must be something to compare. If a work contains no protectible expression, then the intrinsic test is mooted. See Aliotti v. R. Dakin & Co., 831 F.2d 898, 901 (9th Cir.1987). Analytic dissection, “not for the purposes of comparing similarities and identifying infringement, but for the purposes of defining the scope of plaintiffs copyright,” Brown Bag, 960 F.2d at 1475-76, is required under those circumstances. Analytic dissection for this purpose must be guided by the doctrines that draw the somewhat hazy boundary between idea and expression. A work, of course, can be comprised of many basic ideas, which cannot themselves be protected. See Frybarger, 812 F.2d at 529-30; 17 U.S.C. § 102(b). If the idea and the author’s particular way of expressing that idea cannot be separated, under the concept of merger only identical copying of the expression is barred. See Krofft, 562 F.2d at 167-68. This has been said to occur, “[i]f, in describing how a work is expressed, the description differs little from a simple description of what the work is.” Id at 1168 n. 10. Merger means there is practically only one way to express an idea. But if technical or conceptual constraints limit the available ways to express an idea, even though there is more than one avenue of expression available, copyright law will abhor only a virtually-identical copy of the original. Telemarketing Resources v. Symantec Corp., 12 USPQ2d 1991, 1989 WL 200350 (N.D.Cal.1989), aff’d in part as Brown Bag Software v. Symantec Corp., 960 F.2d 1465 (9th Cir.1992). Sometimes this is described as “indispensable expression,” a term which is also used as a synonym for the scénes á faire doctrine. Scénes á faire originated in stock characters and features of dramatic works, See v. Durang, 711 F.2d 141 (9th Cir.1983), and now encompasses stereotyped expression, standard or common features in a wide variety of works, including audiovisual works generated by computers. See Frybarger, 812 F.2d at 530. The functionality of a feature may deprive it of protection under the copyright laws. This exclusion is codified at 17 U.S.C. § 102(b): In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work. Blank forms are one example of a feature that is unprotectible because functional. 37 C.F.R. § 202.1(c); see also Baker v. Selden, 101 U.S. 99, 11 Otto 99, 25 L.Ed. 841 (1880). Additionally, turning the tables on a copyright plaintiff, the lack of originality of a feature or a work will deprive it of copyright protection. Apple Computer, Inc. v. Microsoft Corp., 779 F.Supp. 133, 135 (N.D.Cal.1991); Midway Mfg. Co. v. Bandai-America, Inc., 546 F.Supp. 125, 140 (D.N.J.1982). These doctrines inhere in the source of copyright, U.S. Const., Art. I, § 8, cl. 8. Hence, copyright protection may not serve as an invulnerable shield at the behest of any who invokes its powers. Copyright protection must serve a designated purpose; namely, “[t]o promote the Progress of Science and the useful Arts[.]” Id. Copyright affords an incentive to authors, the guarantee that free-riders will not be able to appropriate the revenues needed to recoup the author's investment in creativity. But while copyright protection increases the expected revenues of authors by restricting competition, it also can have the effect of raising the costs of creation by making a useful building block of creativity the exclusive property of a prior author. Courts have developed copyright’s limiting doctrines as an implicit response to the problem posed by these revenue and cost effects. Certain features of artistic works are so common, or so obvious, that they require virtually no creative effort to conjure and thus do not depend on the revenue effect of copyright protection. The concepts of ideas, scénes á faire, and lack of originality fall into this category of limiting doctrines. On the other hand, placing some features of artistic works in the exclusive domain of one author would so raise the costs of creation for others as to impede the progress of the arts. This accounts for the doctrines of merger, limited number of ways, and functionality. To apply these doctrines, courts must analytically dissect the works of an author seeking copyright protection to determine the metes and bounds of his property interest in the works. The features of works that are covered by these limiting doctrines are protected only from virtually identical copying, for this is the province of the ultimate free-rider, who makes a zero investment in creativity. The court has on an earlier occasion in this case expressed its belief that the various doctrines that limit copyright protection are often barely distinguishable from one another. 779 F.Supp. at 134. Nevertheless, in April the court made such a distinction in ruling on the protectibility of individual items. IV. Apple contends that to “understand” the distinctive appearance of the Macintosh interface, “one needs to consider not only the individual elements that make up the appearance of the interface but also the way those elements are arranged and interact with one another to create the consistent and distinctive Macintosh interface.” Schneiderman 4/24/92 Suppl Decl ¶ 3. The claimed unifying idea of the Lisa and Macintosh works is “an interface suggestive of an office environment with a desktop background, implementing through animated graphical images and fanciful symbols what has been referred to as a ‘desktop metaphor.’ ” Apple Support Mem, Motion for Part SJ re: Subst’l Sim at 8:9-13. Making the point literally, Apple’s counsel has gone so far as to exhibit in the courtroom a chart purporting to show an alternative portrayal of this metaphor, a computer screen containing the image of a desk! Apple’s expert put such hyperbole aside and attempted to isolate the general concepts or ideas that make up the “desktop metaphor.” Although the expert’s descriptions are less than categorical, they principally include: multiple windows that serve as separate workspaces or mini-screens unto themselves; the ability to open and close these windows; icons that represent programs, files and documents; the use of a mouse to manipulate directly these visual displays; menu bars dedicated to an array of choices. Schneiderman 5/24/90 Dep at 243-44; Schneiderman 4/24/92 Suppl Decl; see also Apple Mem Recon, Exh B, listing fourteen “Macintosh Components;” Foley Decl II11. These concepts arguably were used in the Macintosh because — to some degree — they emulate the familiar environment of the desktop and thus render the computer a somewhat less forbidding apparatus. But Apple goes further than simply claiming that the Macintosh interface expresses these ideas. To the extent the individual features of the Macintosh interface are licensed or are unprotectible they are together, or in conjunction with the .protectible features, claimed as a copyrightable arrangement — a “look and feel” which constitutes protectible expression apart from its individual elements. Apple’s theory is necessitated here because the actual arrangement of displays on a computer monitor running any interactive program is largely the product of the user’s efforts, negating any claim of the programmer to original authorship. This arrangement or “look and feel” theory is further necessitated because use of graphic imagery of office objects in computer interfaces is indisputably unoriginal to Apple. See, e.g., Dep Exhs 222, 237, 279, 313. But, more importantly, use of such objects or an arrangement of them denotes the “desktop metaphor,” not as an idea unifying the expressive elements of the Macintosh interface but simply as a collection of visual displays and user commands designed to render use of the computer, as Apple’s expert concedes, more “utilitarian.” Schneiderman 2/26/92 Dep at 558-59. The elements of such an arrangement serve a purely functional purpose in the same way that the visual displays and user commands of the dashboard, steering wheel, gear shift, brakes, clutch and accelerator serve as the user interface of an automobile. See Synercom Technology, Inc. v. University Computing Co., 462 F.Supp. 1003, 1013 (N.D.Tex.1978). Purely functional items or an arrangement of them for functional purposes are wholly beyond the realm of copyright as are other common examples of user interfaces or arrangements of their individual elements— the dials, knobs and remote control devices of a television or VCR, or the buttons and clocks of an oven or stove. Of course, the elements of these everyday user interfaces are seldom conflated into metaphoric images, but that does not mean that the user interface of a computer is less functional. Under the law of this Circuit, an article which has “any intrinsic utilitarian function” can be denied copyright protection “except to the extent that its artistic features can be identified separately and are capable of existing independently as a work of art.” Fabrica Inc. v. El Dorado Corp., 697 F.2d 890, 893 (9th Cir.1983) (emphasis in original). The very purpose of a computer user interface “lies in its ability to help people prepare and analyze their work quickly and flexibly.” Comment, Lotus Development Corp. v. Paperback Software International: Broad Copyright Protection for User Interfaces Ignores the Software Industry’s Trend Toward Standardization, 52 U Pitt L Rev 689, 705 n 65 (1991). Copyright protection can attach only to such a product’s separate artistic features or can afford only such limited protection as appropriate when its features are the product of a compilation. See Harper House, Inc. v. Thomas Nelson, Inc., 889 F.2d 197, 205 (9th Cir.1989). The similarity of such functional elements of a user interface or their arrangement in products of like kind does not suggest unlawful copying, but standardization across competing products for functional considerations. Standardization of the visual features in a computer’s interface helps to achieve its purpose, a point which Apple learned early on when it insisted on interface uniformity for Macintosh applications and which has also been implicitly recognized by this court. Ashton-Tate Corp. v. Ross, 728 F.Supp. 597, 602 (N.D.Cal.1989), aff'd 916 F.2d 516, 522 (9th Cir.1990). Some visual displays are or become so closely tied to the functional purpose of the article that they become standard. If “market factors play a significant factor in determining the sequence and organization” of a computer program, then those patterns may well be termed ideas beyond the ownership of any one seller. Plains Cotton Co-Op v. Goodpasture Computer Serv., 807 F.2d 1256, 1262 (5th Cir.1987). No better evidence of “market factors” (i.e., expectations of users) accounting for the features of computer user interfaces can be found than the almost invariable incorporation of those features in most graphical user interfaces. Microsoft’s videotape Exhibits A and B, seénes á faire (“videotape A” and “videotape B”), establish that graphical user interfaces almost always incorporate the basic elements of the Macintosh interface; even some character-based interfaces possess some of these features. Based on what is observable in Microsoft’s videotapes A and B, the five basic features of graphical user interfaces are found in the following combinations: Features 1. windows. SCENES Á FAIRE TABLE 2. Iconic representation. 3. Object opening/closing. 4. Menus. 5. Iconic manipulation. Computer interface, release yrs. 1. 2. 3. 4. 5. XX XX Xerox Smalltalk, 1979 X Perq/ICL, 1980-85 X X X X X Xerox Star, 1981 X . X Symbolics Release 4.0 and 4.5, 1981-82 X Apollo Domain 3.0, 1981 X X Lilith System, 1982 X DRI Concurrent CP/M, 1983 XX XX Tektronix Smalltalk, 1984 X X * IBM Top View, 1984 XX X Apollo Domain 8.0, 1984 X X X X X Metaphor, 1984-89 X X * Quarterdeck DESQview, 1985-90 X X X X X Xerox Viewpoint, 1985 X X Symbolics Genera, 1986 XX X Commodore-Amiga Workbench 1.3, 1987 XX X Digital Equipment DECwindows, 1988-89 XX XX IBM/MS OS/2 1.1, 1988 XX X Apollo Domain 10.3.5, 1988-91 XX XX Sun Microsystems/OSF Looking Glass, 1988-90 XXX X Open Software Foundation Motif, 1989-91 X X X X X NeXT 2.1, 1989-91 X X X X X IBM/MS OS/2 1.2, 1989 XXX X Consortium Tab Window Manager, 1990 X X X X X GeoWorks Ensemble, 1990 X X X X X Atari GEM, 1990 XXX X X.Desktop, 1990 XX XX Sun Microsystems Sun View, 1990 XX XX Sun Microsystems Open Windows, 1990 (a/k/a “Open Look”) XX XX Commodore-Amiga Workbench, 2.0, 1991 To be sure, a few judges have said in relation to computer software that “the purpose or function of a utilitarian work would be the work’s idea, and everything that is not necessary to the purpose or function would be part of the expression of the idea.” See, e.g., Whelan Associates v. Jaslow Dental Laboratory, 797 F.2d 1222, 1236 (3d Cir.1986); Broderbund Software, Inc. v. Unison World, 648 F.Supp. 1127 (N.D.Cal.1986); Digital Communications v. Softklone Distributing, 659 F.Supp. 449, 462 (N.D.Ga.1987). But this surely is too facile a distinction as it obscures rather than illuminates the first part of the analysis commanded by the Ninth Circuit, comparing the similarities of ideas. The Whelan court dealt with claims that the defendant unlawfully copied plaintiff’s dental laboratory record keeping program to create a competing program and determined that “the idea is the efficient organization of a dental laboratory.” 797 F.2d at 1240. Apple relies heavily on Whelan's dichotomy to support its “desktop metaphor” line of attack and strenuously contends that it is consistent with Ninth Circuit authority. See Apple Reply Mem, Motion for Part SJ re: Subst’l Sim at 8-9. As noted above, however, the Ninth Circuit has recognized that a single program may contain a number of nonprotectible ideas. See Frybarger, 812 F.2d 525. A single program need not limit its expression to one idea, a point that would seem particularly true in the case of software which controls a wide variety of computer functions. Furthermore, the parties’ own understanding as shown by the negotiations leading to the 1985 Agreement demonstrates that Apple and Microsoft accepted specified visual displays as the protectible expression, not the entirety of the interface. 717 F.Supp. at 1431. Still further, the extensive criticism by courts and commentators of Whelan and kindred decisions has recognized that “Whelan’s general formulation that a program’s overall purpose equates with the program’s idea is descriptively inadequate,” the product of “the opinion’s somewhat outdated appreciation of computer science,” and its heavy emphasis on “metaphysical distinctions” instead of “practical considerations.” Computer Assoc. Int’l, Inc. v. Altai, Inc., 1992 WL 139364, 92 Daily Journal DAR 10110, 10117 (2d Cir, June 26, 1992). Most importantly, the Whelan rule distends copyright protection, placing off-limits alternative and improved means of expression and thereby upsetting the uneasy balance which copyright attempts to maintain by preventing free riders from ripping off creative expression while not stifling others from improving or extending that expression. Id at 10120-21. Copyright’s purpose is to overcome the public goods externality resulting from the non-excludability of copier/free riders who do not pay the costs of creation. Peter S. Menell, An Analysis of the Scope of Copyright Protection for Applications Programs, 41 Stan.L.Rev. 1045, 1059 (1989). But overly inclusive copyright protection can produce its own negative effects by inhibiting the adoption of compatible standards (and reducing so-called “network ex-ternalities”). Such standards in a graphical user interface would enlarge the market for computers by making it easier to learn how to use them. Id. at 1067-70. Striking the balance between these considerations, especially in a new and rapidly changing medium such as computer screen displays, represents a most ambitious enterprise. Cf Lotus Dev. Corp. v. Paperback Software Int’l, 740 F.Supp. 37 (D.Mass.1990). While the Macintosh interface may be the fruit of considerable effort by its designers, its success is the result of a host of factors, including the decision to use the Motorola 68000 microprocessor, the tactical decision to require uniform application interfaces, and the Macintosh’s notable advertising. And even were Apple able to isolate that part of its interface’s success owing to its design efforts, lengthy and concerted effort alone “does not always result in inherently protectible expression.” Computer Assoc. Int’l, 1992 WL 139364 at *19, 92 Daily Journal DAR at 10121. By virtue of having been the first commercially successful programmer to put these generalized features together, Apple had several years of market dominance in graphical user interfaces until Microsoft introduced Windows 3.0, the first DOS-based windowing program to begin to rival the graphical capability of the Macintosh. The Macintosh still to this day offers graphical features that translate into competitive advantages. See Walter S. Moss-berg, PC Shoppers May Find It’s Wise to Develop a Taste for Apples, Wall St Journal, Aug. 6, 1992, at Bl. To accept Apple’s “desktop metaphor”/”look and feel” arguments would allow it to sweep within its proprietary embrace not only Windows and NewWave but, at its option, also other desktop graphical user interfaces which employ the standardized features of such interfaces, and to do this without subjecting Apple’s claims of copyright to the scrutiny which courts have historically employed. Apple’s copyrights would hold for programs in existence now or in the future — for decades. One need not profess to know for sure where should lie the line' between expression and idea, between protection and competition to sense with confidence that this would afford too much protection and yield too little competition. The importance of such competition, and thus improvements or extensions of past expressions, should not be minimized. The Ninth Circuit has long shown concern about the uneasy balance which copyright seeks to strike: What is basically at stake is the extent of the copyright owner’s monopoly— from how large an area of activity did Congress intend to allow the copyright owner to exclude others? Herbert Rosenthal Jewelry Corp. v. Kalpakian, 446 F.2d 738, 742 (9th Cir.1971). The court declines Apple’s invitation to use the advent of the microcomputer and its interface to abandon traditional standards which govern copyrights and invent some new law based on highly indefinite constructs such as “look and feel.” As a result, if “desktop metaphor” is to have any meaning in the context of a traditional copyright analysis, it should serve merely as a label for that group of “ideas” embodied in the Macintosh interface devoted to utilitarian uses of that computer, or as a shorthand way of describing the purpose or object of the panoply of ideas of multiple windows, iconic representation and manipulation, menus and object opening and closing functions to assist computer users in operation of their machines. “Desktop metaphor” does not describe the single unifying idea of the Macintosh interface, but is simply another name for the type of interface used on the Macintosh and is by no means exclusive to it. V. The Apple lists of claimed similarities and the submissions of its own expert make evident the following groupings of standardized features or “ideas”: (1) use of windows to display multiple images on a computer screen and facilitate interaction with the information contained in the windows; (2) use of icons to represent familiar objects from the office environment and facilitate organization of information stored in the computer’s memory; (3) manipulation of icons to convey instructions and to control operation of the computer; (4) use of menus to store information or functions of the computers in a place that is convenient to reach, but saves screen space for other images; and (5) opening and closing of objects as a means of retrieving, transferring or storing information. For the reasons which follow in detail, these are common to all the works in suit, and must be deemed “ideas” and thus placed beyond the lone province of Apple or any other programmer. Having decided that a graphical user interface may be composed of many ideas, the court turns to deciding what elements of the Macintosh interface constitute protectible expression. Microsoft moved for partial summary judgment that copyright protection does not extend to any of the 10 individual items on the first list that have not previously been deemed to be covered by the 1985 Agreement (Al, A8, Bl, B2, Dl-3, G4-6). The court granted the motion as to each of the items on three different grounds: merger (all but DI and G6); scénes á faire (all); due to limited number of ways to express idea (Al, DI, G6). The court also ruled that 25 of the 54 items in the lists of similarities alleged against HP’s NewWave products were not original to Apple, that 36 of the items were excluded from copyright protection under § 102(b), and that limited protection was afforded 17 items under seénes á faire and 20 items because of merger of idea and expression. Only item H2, the use of a trash can to represent the discard folder, was determined not to be covered under any of the separate motions. Apple in its motion for reconsideration has finally addressed the individual items on the lists. The court recognizes that the artistic expression that is associated with each item is the actual ground for the protection claimed, this being a suit over visual displays. The law of the Ninth Circuit makes plain, however, that when any of the various doctrines that limit the scope of protectibility are in operation, the affected expression may provide a basis for a claim of infringement only if the alleged copy is virtually identical to the plaintiffs version. See Frybarger, 812 F.2d at 530; Krofft, 562 F.2d at 1168; Rachel v. Banana Republic, Inc., 831 F.2d 1503, 1507 (9th Cir.1987). A. Overlapping Windows. The first and probably most notable idea associated with the Macintosh interface is the use of overlapping windows to display multiple images on a computer screen in order to facilitate organization of information and the user’s interaction with it. Overlapping windows are central to Apple’s claimed similarities and, as shown in the Seénes á Faire Table, supra, invariably present in graphical user interfaces. Al The design and layout in Macintosh includes overlapping rectangular windows in front of a muted background pattern. The court previously ruled that this item was not protectible (1) because the expression described is merged with the idea, (2) under the doctrine of seénes á faire, and (3) because of the limited number of ways to express the idea due to technical or design constraints. Microsoft contends that the main element of Al, the use of overlapping windows, is a functional idea that is necessary for the “desktop metaphor” to be implemented. As evidence of the merger of idea and expression, Microsoft points to use of overlapping rectangular windows by Xerox’s Smalltalk in 1979, by Lilith in 1981, and by the Perq systems as early as 1980. Microsoft’s seénes á faire argument is supported by an exhibit showing overlapping rectangular windows to have been utilized in twenty-six other commercial systems over the past decade. Microsoft videotape A. This, Microsoft asserts, is ample evidence that the feature is standard treatment in the industry, common to this particular genre of work. Microsoft also asserts that the algorithms used to create the shape of windows generate rectangles as a quicker, easier shape than the alternatives, one that does not use too much computing power. Foley Dec! HU 18, 33. Additionally, the design of a program based on a desktop metaphor requires a “muted” background to avoid distracting the user from his work. Apple, on the other hand, argues that because Windows 1.0, as well as Xerox Star and Cedar use tiling instead of overlapping windows, this feature is not indispensable to a graphical user interface. Further, Apple offers deposition testimony that there are many ways to maximize the use of screen space without overlapping windows, such as showing only one large window at a time (i.e., “switching,” infra). Apple also claims that the details of the windows are substantially similar, although those details, such as scroll bar and grow and close boxes, have been determined to be licensed. Two parts of Apple’s claim are easily decided. First, the rectangular shape of the windows is not a copyrightable feature in a computer graphical user interface or, indeed, any other medium of expression. 37 C.F.R. § 202.1(a). Second, the use of a muted background may arguably be a feature of the Macintosh interface not present in Windows 1.0, but there can be no serious contention of its protectibility. A muted background is a default image which appears on any portion of the screen not occupied by a window or menu bar and hardly can be said to represent the sort of creative achievement or expression which the law should exert itself to protect. The overlapping feature of the windows in the Macintosh interface is also unprotectible (and conceivably licensed by the 1985 Agreement). This determination requires somewhat more explanation. The advent of overlapping windows in computer interfaces was largely the product of the operating memory capacity of computers outpacing the capacity of monitor screens to display all the information the computer was able to generate. The computer memory’s ability to handle greater amounts of information and even at times simultaneously to run more than one application program gave rise to the need to accommodate resultingly complex visual images on the limited two-dimensional surface available on a ■ computer screen. Only two methods or processes of displaying two or more images presented themselves to programmers: (1) switch from one image to another, one at a time, each image occupying the entire screen, or (2) split the screen, giving some screen space to each image. All methods of displaying more than one image on the monitor screen use one, the other, or some combination of these means of expression. Switching is illustrated in Figures 1 and 2 which depict use of a screen to switch from image 1, a blank screen, to image 2, a screen of slanted lines. The entirety of the screen can be switched from one image to the other in the sequence or frequency desired. Figure 3 illustrates the alternative means of expression, splitting the screen between the two images noted. The tiled window approach of Windows 1.0 obviously bears close resemblance to a split of the screen of the type illustrated in Figure 3; in most, but as will be seen not all, visual displays of that work, the screen is split between or among the window images generated without switching from one window to another. The Macintosh interface, on the other hand, more plainly represents a combination of both approaches, with the muted background as the default image for any portion of the screen not occupied by one or the other of the windows (or the menu bar which constantly appears, but changes with the application running). The screen is split among the window images displayed and the muted background default image. See Figure 4 in which the blank image is the overlapping (or active) window. Because with Macintosh overlapping windows, the exposed portions of each window are shown simultaneously, the screen may be said to be split among or between the windows displayed; but the active window or topmost one covers the unexposed portions of the remaining or inactive window and thus the screen to that degree shows only one image at a time or employs a switch between or among those portions of the windows. For example, Figure 6 shows the split/switch configuration associated with the screen appearance depicted in Figure 4; the screen is split between images 1 and 2 in the portions of the screen so noted in Figure 6, but on portion 1A the slanted line image is switched off and the blank image switched on. Similarly, when Figure 5 represents the screen split/switch configuration, the images marked 1 and 2 are split between the portions of Figure 5 so denoted but in portion 2A, blank image 1 is switched off and slanted line image 2 is switched on. The user, of course, retains the ability to switch quickly among the exposed windows by clicking on the window desired to be made active or brought to the top. But the means of expression of multiple images are limited to splitting the screen or switching the images. Although the court earlier determined that the overlapping windows of the Macintosh interface was unlicensed by the 1985 Agreement, focusing on the ability to overlap completely and obscure windows, 709 F.Supp. at 930, the tiling function in Windows 1.0 contains the ability to overlap. In Windows 1.0, tiling sometimes resulted in the information in certain windows, such as the clock, being squeezed or stretched to accommodate its surroundings, as other windows encroached or receded. Compare Figures 7 and 10 in which the slanted line image 2 is stretched out to occupy space formerly occupied by blank image 1 which is squeezed down to less screen space. But certain other windows, such as the calculator, calendar, or notepad, merely have more of their contents revealed or obscured by changes in size. Thus, as in Figure 8, a portion of the other image “overlaps.” Note the portion of image 2 occupying area 1A in Figure 8 is overlapped and thus, as in the Macintosh interface, switched off while image 1 is switched on in this area. Further, in Windows 1.0, the user can switch the screen spaces occupied by the images by clicking the mouse on a title bar to generate a mini-icon, and moving the mini-icon into the window whose space will be swapped. The effect is that the window moving to the larger space shows more of its contents, and the window moved to the smaller space shows less — just as if one was on top of the other, and they were flipped. Compare Figures 9 and 10. To be sure, the overlapping windows of the Macintosh are considerably more versatile and aesthetically pleasing than the tiled windows of Windows 1.0 because the tiled windows extend in a cumbersome fashion completely along one axis or the other, making either width or heighth uniform. Moreover, in a tiled window system, the windows must shrink as more and more of them are added to the screen whereas the ability freely to re-size windows in the Macintosh interface without affecting the size of others allows the user to size them in relation to the data they contain or for any other reason. Because a programmer must choose between switching images or splitting the screen, the means of expression are limited, and merger applies at least to the basic arrangements claimed in similarity Al. Moreover, of course, one of these means — splitting the screen — was plainly licensed by the 1985 Agreement and arguably the “overlapping” feature of Windows 1.0 provides at least a basis for Microsoft to contend that this feature is also licensed. Notwithstanding Microsoft’s possible contention, the court believes that overlapping windows are not covered by the 1985 license as Microsoft’s advertising laid some stress on the fact that Windows 1.0 had “No Overlapping,” so “you never ‘lose’ a window.” 709 F.Supp. at 930. Microsoft’s comparative advertising should set some limit on what it can claim was licensed. In any event, the Macintosh interface at most combines a possibly unlicensed switching technique, the only other means of expression, to that plainly licensed, splitting. Accordingly, the scénes á faire doctrine ought also to apply. Finally, use of overlapping windows in the Xerox, Lilith and-Perq systems in the early 1980s shows beyond question that this concept is not original to Apple. See Microsoft videotape A. The court’s April ruling on Al stands. A8 The window design and layout in Macintosh includes windows which may appear partly on and partly off the screen. The court previously ruled that this feature was unprotectible under the merger and scénes á faire doctrines. Microsoft demonstrated that the feature was used in both Xerox’s Smalltalk and in the Perq system prior to its use by Apple, evidence that it is a standard treatment in the industry. Microsoft also argues that the idea of overlapping windows, and even the broader, unitary idea of the “desktop metaphor” that Apple urges upon the court, requires this feature, because free movement of the items on the work desk cannot be achieved in the absence of this feature. Apple argues that this feature is not indispensable to a graphical user interface employing a “desktop metaphor,” for at least four systems shown in Microsoft’s videotape A apparently lack this feature— Star, Sun View, View Point, and Metaphor. The ability to move windows partly on and partly off the screen is, like tiled windows, a product of splitting the screen between that portion of the window left on the screen and all the other images displayed simultaneously, and switching off that portion of the image dragged off the screen. See Figure 11. This also represents an expression of very limited means — sometimes showing one portion of the window while at other times showing another part and, of course, switching off that portion of the window dragged off the screen. Again, while the appearance of a window may represent protectible expression, the ability to move it part on or off the screen is indispensable to the convenient manipulation of information contained in these sub-screens and thus is unprotectible. This portion of the April ruling stands. B1 The design and layout in Macintosh includes the top-most overlapping window displayed as the active window. This feature was previously found unprotectible by the court due to merger of idea and expression and as a standard industry feature, under scénes á faire. It appears that 17 of the 27 systems shown in Microsoft’s videotape A have this feature. Indeed, it is hard to imagine the usefulness of an obscured window being the active window, under the so-called desktop metaphor or any other functional design, although apparently some systems have this option. The active top window idea was first used at Xerox PARC prior to its incorporation in the Lisa system by Apple. These reasons suggest that making the topmost window active is a standard industry feature. Although Apple contends that the top window need not be the active window, citing the systems in the scénes á faire videotape in which other windows can be active, feature B1 clearly appears to be indispensable to the useful employment of the overlapping windows idea. Making the topmost window the sole active window serves the utilitarian function of helping to avoid the accidental input or output of information or manipulation of icons in a window for which this was not intended. This feature also overcomes the constraint imposed by the relatively fixed amount of screen space generally available in most monitors; operating memory has increased greatly in relation to the screen space able to accommodate it, making necessary sophisticated means to manage the information in memory. Furthermore, making the topmost window the active one is no more expression than reading the page to which a book is opened or working on the paper which is on the top of a stack. Instead, this is the description of a process that is inseparable from the idea of overlapping windows, and is thus but an unprotectible idea. Finally, the only expression that can be said to be involved, the graphics that indicate that a particular window is the active window, is feature A4, and covered by the 1985 license. 717 F.Supp. at 1433-34. B2 Macintosh design animation brings a window to the top of a stack of overlapping windows when the mouse is clicked on it. Feature B2, like feature Bl, does not seem to implicate any expression at all. This feature was previously found unprotectible under merger and scénes á faire, a ruling that is itself a bit of an understatement. Clicking the mouse to bring a window to the top of a stack establishes a hierarchical or priority relationship between or among the windows visible on the screen, a concept of great generality or abstraction. By hiding or obscuring the other windows, this process emulates a third dimension in the otherwise two-dimensional computer screen and “defines a partial ordering among graphical objects.” Tomihisa Kamada, Visualizing Abstract Objects and Relations at 44 (World Scientific, 1989). Apple understandably does not expressly refute this in its papers, and the prior ruling stands. D1 Macintosh design animation drags a gray outline of the window along with the cursor when the mouse is pressed on a window’s title bar. Feature D1 was held unprotectible under scenes á faire and due to the limited number of ways to express its idea. Microsoft offers the declaration of Professor James D. Foley to support its contention that hardware constraints dictate this feature. Moving merely the outline of a window, instead of the entire window and its contents, requires far less computing power, and is a common solution of systems running on less-powerful computers. Foley Decl II 24. Nineteen systems employing this feature are displayed in Microsoft’s videotape A. The developers of the Lisa and of Windows 2.03 and 3.0 attempted to move the entire window or a filled-in rectangle, and found this impractical due to technical constraints. Tesler 2/27/90 Dep at 135:25-136:2; Daniels Dep at 87:10-23; Konzen Dep at 120:21-23; Barrett Dep at 97:20-98:16. Apple nonetheless contends that there are other ways of expressing this idea, and that many were considered by their designers. As one example, Smalltalk-76 moves a single corner of the window. Systems that move the entire window or a shaded rectangle are also cited, although it is clear that some, if not all, of these systems have more computing power at their disposal. The movement of just a corner of a window is not compatible with the idea of manipulating multiple windows for users’ convenience in organizing information. Users must know what area the “piece of paper” or workspace they are moving will cover. The outline of the window is the obvious graphic for this idea, and its use in over a dozen systems reinforces this conclusion. The ruling on this item stands. D2 Macintosh design animation moves a window to the new position as soon as the mouse is released after dragging the window’s outline. Like B2, this describes only the process by which the visual displays are moved. Using the mouse in this way is standard in the industry, and is inseparable from the idea of direct manipulation by mouse of multiple windows. The prior ruling under scénes á faire and merger stands. D3 Macintosh design redisplays all newly exposed areas on the screen after a window is moved. The court previously ruled this feature unprotectible under merger and scénes á faire. This feature appears in 25 of the systems in Microsoft’s scénes á faire videotape Exhibit A. At least two of the systems, the Lilith and the Perq, preceded Apple in use of this feature. Microsoft argues that this is a functional feature that is vital to the idea of multiple, overlapping windows, as well as an idea in and of itself. Apple asserts that the “graphic holes” that are left in areas of windows previously obscured, such as in Smalltalk-76, show that it is not necessary that these blank spaces be immediately refreshed. The absence of information, however, cannot be a useful feature in a system designed to convey information to its users, and such “graphic holes” would represent technological shortcomings and hardly serve aesthetic considerations. The prior ruling on this feature stands. B. Iconic Representation. The Apple works in suit use icons that represent familiar objects from the office environment to facilitate the organization of information stored in the computer’s memory. At the outset, iconic representation is not original to the Macintosh interface. In the context o