Full opinion text
MEMORANDUM AND ORDER ■ KEETON, District Judge. By Memorandum and Order of March 20, 1992, the court dismissed the parties’ motions for summary judgment in this copyright infringement action and invited new motions compatible with rulings therein announced. Each party has renewed its motion for summary judgment and filed further submissions (Docket Nos. 168-190). A hearing on these motions was held on May 19, 1992, and additional submissions were filed after that hearing. The reader may find background information in two earlier documents issued by this court: the first, the Opinion in a related case involving the Lotus 1-2-3 program at issue here, Lotus Dev. Corp. v. Paperback Software Int'l, 740 F.Supp. 37 (D.Mass.1990), and, the second, a Memorandum issued in this case, Lotus Dev. Corp. v. Borland Int’l Inc., 788 F.Supp. 78 (D.Mass.1992) (Memorandum and Order of March 20). In the Memorandum and Order of March 20, I concluded that Lotus had failed to frame adequately its contentions with respect to the infringement of elements of its user interface less than the whole interface. In its renewed motion for summary judgment, Lotus asserts specifically that Borland has copied expressive elements of the 1-2-3 interface, including “menu commands,” “menu structure,” “long prompts,” and “keystroke sequences.” Although Lotus continues to argue that its entire user interface beyond dispute was copied, I adhere to the view that on the present record a reasonable jury could find that Borland copied less than the whole 1-2-3 user interface. Nevertheless, based upon the parties’ most recent submissions, I conclude that, beyond genuine dispute, Fed.R.Civ.P. 56, Borland copied parts of the 1-2-3 user interface. For the reasons stated herein, I deny Borland’s motion for summary judgment and grant, in part, Lotus’ motion for summary judgment. I. DEFINITIONS AND PREMISES The Memorandum of March 20 presented for possible use in this case the following form of jury interrogatory concerning the extent to which Borland copied the Lotus 1-2-3 user interface in creating its Quattro programs: Question 1 (a) Do you find that the Quattro Pro user interface as a whole was copied from the Lotus 1-2-3 user interface as a whole? _ YES _ NO (b) Do you find that the part of the Quattro Pro user interface called the “emulation interface” (also called the “1-2-3 compatible interface”) as a whole was copied from the Lotus 1-2-3 user interface as a whole? _ YES _ NO (c) If NO, do you find that some part, and, if so, which of the following part or parts of the Lotus 1-2-3 user interface were copied into some part of the Quattro Pro “emulation interface” (also called the “1-2-3 compatible interface”)? (1) The menu commands _YES _NO (2) The menu structure _YES _NO (3) The command sequence _YES _NO (4) The long prompts _YES _NO (5) The macro facility _YES _NO The Memorandum of March 20 noted that it was not clear that Lotus was making a claim of the sort addressed in part (a) of proposed Question 1 and that ambiguity remained regarding the meaning of the terms used in parts (c)(l)-(5). Lotus has in its recent submissions clarified its contentions. First, Lotus acknowledges that the “native” modes of the Quattro programs have user interfaces that differ from that of 1-2-3. Thus, the question posed in part (a) of Question 1 is not in genuine dispute. Second, Lotus has defined, as it uses them, the terms “menu commands,” “menu structure,” “keystrokes,” “keystroke, sequences,” “long prompts,” and “macro language.” In general, except for some blending of argument with definitions, the parties appear not to be in dispute about the meaning of these and other terms defined below. The definitions that I use in this Memorandum, for the purpose of explaining and analyzing the contentions of the parties, are consistent with the submissions of both Lotus and Borland, as I understand them. “Command” refers to an abbreviated description of a direction that a user of a software program (whether Lotus 1-2-3, Borland’s Quattro Pro, or another program) may invoke to cause some operation to be performed. “Menu” refers to a display on the computer monitor of a limited number of commands available to the user at a given moment. “Menu command” refers to a command that appears in a menu. In Lotus 1-2-3, a menu command is ordinarily a single English-language word. In rare instances, it is instead a representation of an English-language pronunciation (such as “Xtract”). Menu commands are displayed on the computer monitor by the 1-2-3 program in a succession of menus. The menus communicate to the user, in sequence, the spreadsheet operations available to the user. “Command structure” refers to the organization of the menus and menu commands. (Other phrases used with essentially the same meaning include “menu command structure,” “menu hierarchy,” and “menu command hierarchy.”) In Lotus 1-2-3, menu commands are organized so that less than a dozen related menu commands are displayed at any given moment. This display communicates to the user the spreadsheet operations immediately available. Each menu of less than a dozen commands is linked to preceding/succeeding menus by the operation of menu commands. All command menus are ultimately linked to a single main (root/trunk) menu to form a “menu tree.” “Keystroke sequence” refers to a sequence of keystroke entries that a user may invoke. Keystroke sequences may be generated as one navigates the menu command hierarchy performing sequential spreadsheet operations. “Long prompt” refers to a displayed multi-word English-language description of a “highlighted” menu command. A “highlighted” menu command appears on the computer monitor as a block of inverse video — that is, on a monochrome monitor with a black background on which characters are lit, a highlighted word appears as black letters within a lit block. “Macro language” refers to a feature by which a user may define a very short keystroke sequence as equivalent to a longer keystroke sequence. Thus, a user may invoke the short keystroke sequence (a “macro”) as a substitute for the longer keystroke sequence. In stating this definition, I omit a sophisticated programming capability available in 1-2-3 through its macro language feature that Lotus, as I understand its submissions, does not contend is involved in its claim of infringement in this action. Having stated the definitions of the components of the user interface that I will use in this Memorandum, I now state additional points that I conclude are not in dispute about the relations among these definitions and associated matters. The keystroke sequences and macro language have functionality. Typing (“inputting,” in jargon) the first character of a command word invokes the command and causes the operation associated with the command word to be performed. (In many instances, a submenu associated with the command word is displayed.) The menu command hierarchy is a fundamental part of the functionality of keystroke sequences and the macro language. For example, the keystroke sequence “/RFC” directs the computer to format a range of numbers to appear as currency values because the character “/” initiates a command sequence, the character “R” implements the “Range” command, the character “F” implements the “Format” command, and the character “C” implements the “Currency” command. It may be necessary to enter additional information to invoke a spreadsheet operation fully. For example, in order to implement an operation formatting a range of numbers as currency values, it is necessary to delimit the range of spreadsheet cells to be formatted. In addition, it is necessary to specify the number of decimal places to be displayed. Variables such as range of cells and number of decimal places, for which values must be input each time an operation is to be performed, are called “parameters.” The authors of Lotus 1-2-3 made certain predictions about the value of each of the relevant parameters likely to be input for use with certain spreadsheet operations. Those predictions have been incorporated into Lotus 1-2-3 as suggestions; a user failing to specify a value for a parameter where it is necessary to supply one accepts the suggestion of 1-2-3’s authors by default. The “/RFC” command set, for example, has associated with it two “default parameters.” The default for range is “current cell,” and the default for number of decimal places is two (i.e., dollars and cents). The user who prefers a different format (for example, whole dollars rather than dollars and cents) may enter a different number, zero rather than two (the default) to so indicate. In similar fashion, the user may supply a range different than the default value. The menu structure will not permit the command “Currency” to be executed without first proceeding through the “Range” and “Format” commands. Indeed, inputting a “C” at a different point in the menu structure may cause a different command, such as “Copy,” to execute. The foregoing description identifies one way in which a menu command may be invoked — that is, by pressing the letter key corresponding to the English-language name of the command. A second way of invoking a menu command is to make use of the highlighting around a menu command. The user may use arrow keys on the keyboard to move the block of highlighting (“cursor”) to an adjacent menu command. Depressing the “Enter” key (or “Return” or “< — ¡ ” key, depending on the keyboard) invokes a highlighted command. Because the macro language plays such a central role in the parties’ contentions, and because it is an extraordinarily sophisticated element of Lotus 1-2-3, I recite some further examples of the use of the macro language. A user may define a keystroke sequence with a macro by inputting the keystroke sequence in a spreadsheet cell and assigning a macro label to that cell. For example, if a user enters the sequence “Hello~” (the tilde, stands in for the “Enter” or “Return” or “< — j” key) in spreadsheet cell Al and assigns the macro label “\H” to cell Al, then the user may cause the word “Hello” to appear in any other cell by invoking the abbreviated keystroke sequence “\H” instead of the longer sequence “H,” “e,” “1,” “1,” “o,” “Enter.” (The reverse slash, “\,” signifies the “Alt” key, pressed and held in place while another key is pressed. Thus, the keystroke sequence “\H” consists of pressing the “h” key while the “Alt” key is simultaneously depressed.) The capability of the program to enter, for example, the word “Hello” in any spreadsheet cell when the “\H” keystroke sequence is entered is functional. It is not protectable by copyright, 17 U.S.C. § 102(b), and it is not the subject of this case. At a more sophisticated level, a user may construct a macro that implements menu commands. For example, a user may input the keystroke sequence “/RFC” in cell A1 and attach a macro label such as “\C” to that cell. This defines “\C” in the macro language — subject to later redefinition — as “/RFC~“.” Invoking the “\C” keystroke sequence in any cell will cause the spreadsheet to format that cell to display numeric values as currency values. The ability of the computer to format the cell for currency values is functional. It is not copyrightable, id., and it is not the subject of this case. The spreadsheet program is instructed to format the cell by the keystroke sequence “R,” “F,” “C," “Enter,” “Enter.” That sequence invokes the menu commands “Range,” “Format,” and “Currency” (and accepts, by operation of the “Enter” key twice, two default parameters). Thus, the menu commands are an important part of the functionality of the macros. Lotus contends that the menu commands and the command structure are copyrightable expressive elements of the 1-2-3 user interface and that they are copyrightable expressive aspects of the macros. Borland contends that the macros, in their entirety, are an uncopyrightable system. These competing contentions, as well as others, are addressed below. II. COPYING In Borland’s Mem. in Supp. of Renewed Mot. for Summ. J. (Docket No. 168), Borland states the following: Furthermore, it is undisputed in the record that Borland did not copy the 1-2-3 menu command hierarchy directly from any 1-2-3 version, including Release 2.01. Rather Borland employees reviewed books about 1-2-3, Release 2.01, written by third-parties, which books contain schematic or menu-tree type representations of the 1-2-3 menu command hierarchy. Borland used these third-party menu trees to construct 123-compati-ble menu hierarchies in their own products. But Borland employees did not copy those menu trees, even from the third-party books. Rather, they viewed the menu trees and implemented into their own products the relationship of functions depicted in those menu trees. Id. at 15 (footnote omitted). Borland contends that, on these facts, indisputably Borland did not “copy.” Borland’s contention, however, is based on its idiosyncratic use of the word “copy,” and is fundamentally wrong. Instead, Borland’s admissions establish beyond dispute that Borland did copy the menu commands and command structure of Lotus 1-2-3. Borland argues that the menu command hierarchy is a “set of functional relationships” that is nowhere displayed in the 1-2-3 user interface. Thus, Borland argues, it did not and could not have “copied” the menu command structure. That argument simply fails. It is an argument about a fact — copying or not. Its premise bears instead not on the fact — copying or not— but upon the legal issue of copyrightability. That is, the premise of Borland’s argument is that the menu command structure must be fixed in a tangible medium to be copyrightable. See 17 U.S.C. §§ 101-102. The admitted fact that the Quattro programs duplicate the set of “functional relationships” of Lotus 1-2-3 and were designed to do so is conclusive against Borland on the issue of copying that set of functional relationships. Thus, Borland has admitted that it intentionally incorporated into its user interface the 1-2-3 menu commands and menu command hierarchy. Moreover, I reject Borland’s tangible-medium argument as applied to this case. The argument would be relevant, if at all, to copyrightability rather than copying, but because Borland has made the argument as if it had a bearing on copying, I will digress briefly to address it here. To be the subject of copyright protection, an expression must be fixed in a tangible medium. Id. § 102. All that is required in this regard is that the expression be embodied in a copy “by or under the authority of the author” in a form “sufficiently stable to permit it to be perceived, reproduced, or otherwise communicated for a period of more than transitory duration.” Id. § 101. The output of a computer program, at least insofar as it is typical of the program, predictable from it, and directed by the operation of the program, satisfies these requirements. The menu command hierarchy is part of the 1-2-3 program’s output, is directed by the program, is identical each time the program runs, and may be perceived (and, as in this case, duplicated). It is irrelevant that the hierarchy cannot be perceived in its entirety at one moment (for example, in one screen display), just as it is irrelevant that the plot of a novel cannot be perceived from viewing one page. The menu command hierarchy is copyrightable subject matter. Borland’s argument that it copied from third-party sources rather than Lotus is equally without merit. Borland argues this point on the strength of the court’s suggestion that Borland might prove that it did not copy Lotus’ interface by demonstrating that instead it copied someone else’s. If Borland had copied a third party’s independently created menu command hierarchy that “happened” to duplicate Lotus’ interface, Borland might be excused (at least from liability to Lotus, though not perhaps to the third party). Here, however, Borland has admitted copying from sources that, with or without permission from Lotus, copied from Lotus. Borland was aware that those sources copied from Lotus — Borland admits verifying the accuracy of the menu structure it generated from third-party sources by comparing the Quattro programs to 1-2-3. The fact that Borland used third-party sources as a means of. copying the Lotus 1-2-3 menu command hierarchy in no way excuses Borland’s deliberate imitation of the Lotus menu structure. Based on the foregoing, I conclude that, beyond genuine dispute, Borland copied the menu commands and menu command structure of Lotus 1-2-3. Moreover, Borland admits to copying the functionality of the keystroke sequences and macro language. By its own assertions, Borland’s reason for copying the menu command structure was to obtain the benefit of its functionality. It follows that in fact Borland has admitted copying aspects of the keystroke command sequences and macro language that Lotus contends are expressive and copyrightable, although Borland chooses to describe what it did in another way. Lotus has identified at this stage one additional element of its user interface that it claims Borland copied: the long prompts. It is clear from the record that the long prompts appearing in the Quattro programs differ in many instances from the long prompts in 1-2-3. On the other hand, in many other instances there is little or no difference. There is evidence that a Borland employee wrote the Quattro long prompts and did not “copy” from 1-2-3; however, she admits to looking at 1-2-3 to ensure that she correctly understood the Lotus commands. A reasonable jury could find on the basis of this evidence either way — that Borland did or that Borland did not copy the long prompts of 1-2-3. Therefore, I conclude that whether the long prompts were copied is a question for the jury. Moreover, because of this conclusion, it is apparent that a reasonable jury could find that Borland did not copy the 1-2-3 interface as a whole. III. COPYRIGHTABILITY Having concluded that Borland copied the menu commands and menu command hierarchy as well as the keystroke sequences and macro language, I now proceed to determine whether those aspects of the 1-2-3 user interface, taken together, are copyrightable. A. Potential Fact Questions I stated earlier my tentative conclusion that the application of the copyrightability standard is for the court and not a jury. Borland, 788 F.Supp. at 96. I invited counsel to respond to that tentative conclusion. Lotus has fully endorsed the conclusion that copyrightability issues, at least in this case, are for the court. Borland, however, contends (in the alternative to Borland’s own motion for summary judgment) that questions of copyrightability are for a jury. In response to the court’s invitation, Borland proposes the following questions to be answered by a jury: 1. Does the Lotus 1-2-3 menu command hierarchy comprise a system, procedure or method of operation? 2. Is the 1-2-3 menu command hierarchy designed and used as a system for performing tasks using a spreadsheet program? 3. Are the command words of 1-2-3 and their order an inseparable part of a system for performing spreadsheet tasks? 4. Does the 1-2-3 menu command hierarchy enable a person to map out and execute a procedure for performing a particular spreadsheet task? 5. Is the 1-2-3 menu command hierarchy a means for issuing commands to the computer program to perform spreadsheet tasks? 6. Does the 1-2-3 menu command hierarchy provide a procedure or method of operating a spreadsheet program? 7. Was the 1-2-3 menu command hierarchy designed and arranged using functional rules or principles? 8. Was the 1-2-3 menu command hierarchy designed and arranged to maximize its efficiency and usefulness? 9. Is the 1-2-3 menu command hierarchy fundamental to a user’s ability to execute macros written using 1-2-3? 10. Do the command words of 1-2-3 convey information to the user other than the choices of functions that are available? 11. Does the 1-2-3 menu command hierarchy explain to the user how to use the 1-2-3 program to perform spreadsheet tasks? Liability Questions for the Trier of Fact (Docket No. 95) at 2-5. Under my rulings, it is irrelevant that the 1-2-3 interface includes functional elements or “comprises a system” so long as it also includes separable expressive elements. Thus, five of Borland’s first six questions are irrelevant. Question 3, though relevant, is a mixed law-fact question that will not be asked of a jury for reasons previously stated. Borland, 788 F.Supp. at 94-96. The seventh and eighth questions are more problematic. Those questions may have a bearing on whether the expressive elements of 1-2-3 are in fact separable from the functional aspects of the interface. As stated in the Memorandum of March 20, “[i]f the menu commands or menu command structure were dictated solely by functional concerns, then those elements may not be copyrightable.” Borland, 788 F.Supp. at 97 (citing Brandir Int’l, Inc. v. Cascade Pac. Lumber Co., 834 F.2d 1142, 1145 (2d Cir.1987)). Lotus argues that Brandir is not “strictly applicable” because computer programs are classified as “literary works” and not “pictorial, graphic, and sculptural works” by the Copyright Act. The law draws heavily on analogy, however, and computer programs, whatever their formal classification, like pictorial, graphic, and sculptural works, are useful articles. Moreover, though insisting that the burden on this question falls on Borland, Lotus acknowledges that any elements of its program that were functionally dictated are not copyrightable. Pl.’s Mem. of L. in Supp. of Renewed Mot. for Summ. J. on Infringement (Docket No. 171) at 34-35 n. 47. Nevertheless, I conclude that no reasonable jury could find that the menu command hierarchy was limited to one or even several alternate designs at the time it was created. On the basis of the evidence before me, a factfinder could conclude that some — but only some — subelements of the menu command hierarchy were functionally dictated. For these reasons, it is my tentative conclusion that there may be genuine fact disputes regarding the subject matter of Borland’s seventh and eighth questions. Questions seven and eight, as formulated, however, are “evidentiary” rather than decisive or “ultimate” issues appropriate for use on a verdict form. It remains to be determined whether an “ultimate question” on the subject matter of the seventh and eighth of Borland’s suggestions can be formulated as an adjudicative fact question of the type that is normally the province of a jury and does not „,pose the substantial risks described in my earlier Memorandum, see Borland, 788 F.Supp. at 94-96. Question nine assumes a premise that is fundamentally in error — an error that has been systemic in Borland’s arguments throughout the course of this litigation. The problem, which may be described as a “chicken and egg” problem, is addressed in greater detail in Section C, below. At this point, I simply state my view that question nine should not be asked of a jury. The answers to the last two questions are irrelevant. A “no” answer to question ten does not preclude a determination that the command words are expressive even though the expression may be limited — i.e., the words communicate the functions to which they are assigned. Similarly, the answer to question eleven is immaterial. B. Legal Issues on Copyrightability In the Memorandum and Order of March 20, 1992, I concluded that, “absent further guidance from higher authority before the date of trial,” Borland, 788 F.Supp. at 89,1 would apply the following standard for deciding copyrightability issues: FIRST, in making the determination of “copyrightability,” the decisionmaker must focus upon alternatives that counsel may suggest, or the court may conceive, along the scale from the most generalized conception to the most particularized, and choose some formulation, some conception of the “idea,” “system,” “process,” “procedure,” or “method”— for the purpose of distinguishing between the idea, system, process, procedure, or method and its expression. SECOND, the decisionmaker must focus upon whether an alleged expression of the idea, system, process, procedure, or method is limited to elements essential to expression of that idea, system, process, procedure, or method (or is one of only a few ways of expressing the idea, system, process, procedure, or method) or instead includes identifiable elements of expression not essential to every expression of that idea, system, process, procedure, or method. THIRD, having identified elements of expression not essential to every expression of the idea, system, process, procedure, or method, the decisionmaker must focus on whether those expressive elements, taken together, are a substantial part of the allegedly copyrightable “work.” Id. at 90 (quoting Lotus Dev. Corp. v. Paperback Software Int'l, 740 F.Supp. 37, 60-61 (D.Mass.1990)) (all emphasis omitted). The Second Circuit has recently issued an opinion that bears significantly on issues of copyrightability and substantial similarity. In Computer Assocs. Int’l, Inc. v. Altai, Inc., 1992 WL 139364, No. 91-7893, 1992 U.S.App. LEXIS 14305 (2d Cir. June 22 1992), the court announced an “Abstraction-Filtration-Comparison” test for determining “substantial similarity.” The first two steps, “abstraction” and “filtration,” are designed to define the idea of the program and to eliminate it (as well as other noncopyrightable subject matter) from further consideration. The test “serves ‘the purpose of defining the scope of plaintiff’s copyright.’ ” Id. (quoting Brown Bag Software v. Symantec Corp., 960 F.2d 1465, 1476 (9th Cir.1992)). Thus, the first two steps of the Second Circuit’s “substantial similarity” test concern what other courts and commentators have called “copyrightability.” Only the third step, “comparison,” addresses similarities between the copyrighted work and the allegedly infringing work. To what extent does the Second Circuit’s ‘ ‘Abstraction-Filtration-Comparison’ ’ test differ, either substantively or in methodology, from the combination of copyrightability and substantial similarity tests tentatively adopted for this case in my Memorandum and Order of March 20? I conclude that the standard for determining copyrightability stated in my Memorandum and Order is compatible with the abstraction-filtration portion of the Second Circuit’s test. The Second Circuit founded its abstraction step on the opinions of Judge Learned Hand that were also the foundation of the first step of the copyrightability test stated in my Memorandum and Order. The second step of that copyrightability test parallels the Second Circuit’s “filtration” step. The third step of the Second Circuit test, “comparison,” serves two functions. The first concerns the issue addressed in the third step of the “copyrightability” test I have tentatively adopted for this case— whether the expressive elements of the allegedly copyrightable work are a substantial part of it. I conclude that in this respect the two tests are compatible substantively though different in methodology. The other function that the Second Circuit’s “comparison” step serves is emphasized in the term used to identify it — “comparison.” The comparison is between the relevant portions of the allegedly infringing work and the expressive elements of the allegedly copyrightable work to ascertain whether any part of the allegedly infringing work is similar to expressive elements of the allegedly copyrightable work that are a substantial part of the allegedly copyrightable work (i.e., whether there is substantial similarity in the mixed law-fact sense). I conclude, again, that in relation to this comparison, the Second Circuit’s test and the combination of the “copyright-ability” and “substantial similarity” tests I have adopted tentatively are compatible substantively, though different in methodology. Borland argues that the decision and reasoning in Computer Associates are contrary to the Paperback decision, and, as well, to my Memorandum and Order of March 20. In one respect only, however, did the Second Circuit explicitly so indicate. It criticized an incentive-based reason stated at one point in the Paperback opinion. The criticized argument, however, was by no means essential to the outcome in Paperback, and acceptance or rejection of that argument is not likely to affect the outcome in this case. In other respects, my reasoning in Paperback and in my Memorandum and Order of March 20, 1992 in this case was substantively consistent with the opinion of Computer Associates, as I understand that opinion, though, as I have explained above, different in methodology. A particular example of significance to this case concerns Borland's argument that the Second Circuit’s treatment of “compatibility” militates against copyrightability of the 1-2-3 interface. Borland extracts from the Second Circuit’s opinion a determination that aspects of computer software cannot be subject to copyright if they are greatly circumscribed by the hardware or software with which they are designed to interact. That proposition, even if accepted as Borland has stated it, does not apply to 1-2-3. Borland’s argument to the contrary must be rejected for reasons that I explain in Part C, immediately below. C. Which Came First? A familiar childhood riddle asks: Which came first — the chicken or the egg? As folk riddles often do (and lawyer’s questions on cross-examination sometimes do), this riddle strongly suggests that any answer but one of two identified options is out of bounds. Perhaps from another perspective, however, one may recognize that the chicken and egg are one type of organism in different stages of a life cycle. Over a long span of time, the form of that life cycle has evolved from an earlier form in which the chicken and the egg were not so distinct as they now appear to be. History, science, and philosophy may provide another answer, or many others, different from the two suggested in the riddle. Borland’s brief poses a conundrum for this case in a form analogous to the chicken-and-egg riddle. Borland asserts that the 1-2-3 interface is not copyrightable because the menu command hierarchy “was dictated by the nature of the user macros with which it was designed to interact.” Supp.Mem.Re: Additional Authority (attached to Docket No. 189) at 9 (quoting Computer Assocs. Int’l v. Altai, Inc., 1992 WL 139364, at *24, No. 91-7893, slip op. at 50, 1992 U.S.App. Lexis 14305, at *67 (2d Cir. June 22, 1992)). An implicit premise behind this argument is that the menu command hierarchy was designed to fit the macros. The subtle suggestion is that the macros came first — that they were preexisting. Necessarily implicit in Borland’s argument is the assertion that neither the final version of the menu command hierarchy nor any substantial part of it was preexisting when the macro language was created. The thrust of the discussion of “compatibility” in Computer Associates, however, relies upon proof that what the program was designed to fit was already in existence before the program was designed to fit it. Thus, a program designed to fit hardware specifications cannot be protected by copyright unless the program contains expressive elements not substantially dictated by the hardware. Similarly, a program designed to interact with preexisting software, such as the operating system at issue in Computer Associates, is not entitled to protection to the extent that it is constrained by the need for compatibility with the preexisting software. Thus, the rule makes sense if the premise of a preexisting functional limitation is valid. In this case, however, there is a very simple answer to the question “Which came first?” The Lotus 1-2-3 interface — or at least a version of it — was written first. All user macros derive from it. Thus, Borland is simply wrong factually to argue that the 1-2-3 interface was constrained by the macros. On this issue, there is no genuine dispute of fact. I assume in Borland’s favor that, like the chicken and the egg, the macro language (as opposed to macros defined by users using the macro language, which necessarily came later) evolved simultaneously with the menu commands that delimit it. Nevertheless, it is beyond dispute that the macro language did not evolve first. Borland has argued, also, that the need to ensure that the menu commands in any given menu begin with a different letter is a functional constraint. Even that argument fails, however, because of the availability of other symbolic tokens; e.g., A, B, C, ... or 1, 2, 3,---- It is no doubt true that the macros have functional significance. Moreover, as this court found in Paperback, the menu “system” is a fundamental part of the functionality of the macro language and the macros. As the Second Circuit has recognized, however, the fact that a form of expression takes on functional character does not remove it from the protection of copyright. Brandir Int'l, Inc. v. Cascade Pac. Lumber Co., 834 F.2d 1142, 1147 (2d Cir.1987); see also Paperback, 740 F.Supp. at 58. Lotus used the slash character (“/”) to initiate a command sequence; the character “R,” to specify the command “Range”; the character “F,” to specify the command “Format”; and the character “C,” to specify the command “Currency.” A user inputting the characters “/,” “R,” “F,” and “C” in sequence (and supplying necessary additional parameters) will cause the computer to perform the operations associated with the specified commands. If the keystroke sequence “/RFC-” were stored in a spreadsheet cell that had been assigned the macro “\C,” then a user inputting the “\C” keystroke sequence would cause the computer to perform the same operations as would be invoked by the six-keystroke sequence “/,” “R,” “F,” “C,” “Enter,” “Enter.” Because the “\C” keystroke sequence is invoked by pressing the “c” key with the “Alt” key depressed as one would use a “Shift” key, it involves two keys but perhaps only one “keystroke.” Thus, depending upon one’s definitions, a saving of either five or six keystrokes would result. Had Lotus preferred, it could have chosen, for example, the command “Scope” instead of “Range,” “Appearance” instead of “Format,” and “Money” instead of “Currency.” Then the user would invoke the sequence “/,” “S,” “A,” “M” to establish a “money” appearance ($xxx.xx) for a scope of cells. In that case, the parallel to Borland’s argument would be that the existence of “/SAM""” macros in users’ files (spreadsheet files with cells defining macros may be saved for future'use) “dictated” the use of the “Scope,” “Appearance,” and “Money” commands in the menu command hierarchy. Obviously, the argument is without merit — “Range,” “Format,” and “Currency” are demonstrably acceptable choices. The fact that Lotus chose one command set for its first version may have made it “necessary” that future versions adopt the same command set; however, the initial choice of the command set was a free choice. It bears repeating here that Lotus (though not Borland) was entitled to incorporate in version 2.01 the menu command hierarchy that it employed in earlier versions of 1-2-3. Thus, Borland’s contention that the macros preceded and dictated the menu command hierarchy of version 2.01 fails because that fact is irrelevant. This is so because Lotus’ exercise of creative expression that is at issue in this case (the menu command structures of versions 1A and 2.0) manifestly preceded the users' macros that incorporate the 1-2-3 command hierarchy. It may be argued that the macros, by themselves, are functionally dictated by the 1-2-3 menu command hierarchy and so are not copyrightable independently of the copyrightability of the menu command hierarchy. Moreover, even if some macros are copyrightable, it may be that the owner of any copyright in a macro is the user who authored the macro and not Lotus. I am not faced with those questions in this case. I conclude only that Borland’s argument rests on a false proposition. Thus, to decide this case I need not and do not decide whether Borland is prohibited from reading and interpreting the macros that have been created by users of 1-2-3. Had Borland created a program that read users’ 1-2-3 macros and converted them to macros for use in the Quattro programs’ native modes, so that they could be interpreted, executed, modified, debugged, etc. by resort to Borland’s command hierarchy, that would have presented a different case from the one now before me. Borland did not obtain the right to expressive aspects of Lotus’ command hierarchy merely because — if it be the case — the 1-2-3 program revolutionized the spreadsheet market. The menu command hierarchy has a functional aspect when incorporated into the keystroke sequences and macros. That functional aspect is separable from the expressive aspect that preceded it. Borland cannot obtain the right to use the macros and keystroke sequences just because the only means of doing so is by infringing expressive features of the Lotus 1-2-3 macro language and keystroke sequences. The error of Borland’s argument may be demonstrated by a simple hypothetical. If an uncopyrighted movie is made from a copyrighted novel (under the authority of an appropriate license), the public is free to copy all aspects of expression unique to the movie; however, the novel does not enter the public domain. Similarly, if the macros have uncopyrightable functional aspects, Borland does not infringe the copyright in Lotus 1-2-3 version 1A by duplicating the functionality in any way that does not copy the expressive elements of Lotus 1-2-3 version 1A, but it must not infringe upon expressive elements. Borland’s argument that Baker v. Selden, 101 U.S. 99, 25 L.Ed. 841 (1879), establishes a different rule that controls this case is rejected. This is not a case like Baker v. Selden in which the system depends on the use of the copyrighted matter. Borland has, in fact, designed a system (Quattro Pro’s native mode), using macros and keystroke sequences and using an alternate command hierarchy, that is fully functional. D. More on Computer Associates v. Altai Some further points regarding the Second Circuit’s opinion bear upon its applicability to issues in the present case. 1. Beyond Program Code to Nonliteral Expression. The reasoning underlying the Second Circuit’s “Abstraction-Filtration-Comparison” test extends beyond program code to nonliteral expression. That is, although the issue in that case concerned program code and its structural aspects, the Second Circuit based its conclusions on reasons of broader application. In fact, the court explicitly approved determinations with respect to the copyrightability of certain non-literal, noncode (nonstructural) aspects of the 1-2-3 spreadsheet in Paperback. See Computer Assocs., 1992 WL 139364, at *16, No. 91-7893, slip op. at 34, 1992 U.S.App. LEXIS 14305, at *46. Thus, from statements made in the Second Circuit’s opinion, albeit obiter dicta, I draw support for my conclusion that certain expressive elements of the 1-2-3 user interface may be protected by copyright. 2. Criticism of Whelan. Borland’s argument that the Second Circuit, with one stroke, knocked out the “conceptual underpinnings” of Whelan Assocs., Inc. v. Jaslow Dental Lab., Inc., 797 F.2d 1222 (3d Cir.1986), cert. denied, 479 U.S. 1031, 107 S.Ct. 877, 93 L.Ed.2d 831 (1987) and of Paperback cannot be sustained. First, Whelan remains good authority in the Third Circuit and may provide guidance for this court in the same way as the Second Circuit’s opinion in Computer Associates. That is, for courts in the First Circuit, each of these decisions from other federal circuits is instructive and neither is controlling beyond the persuasive force of its reasoning. Second, and more to the point, the Second Circuit criticized the Whelan decision for reasons that in large part do not apply to the rulings I made in Paperback or to the rulings I have made and now make in this case. The Second Circuit was critical of Whelan on two principal grounds — that it failed to recognize that programs may have more than one idea, and that it mistakenly asserted that all that is not idea is expression. To resolve those problems, the Second Circuit formulated its “Abstraction-Filtration^ Comparison” test. In both these respects, the rulings I have made in the present case are compatible substantively (though not entirely in methodology) with the Second Circuit’s views rather than those of the Third Circuit, where those two circuit’s differ. For example, as indicated in the Paperback opinion, I recognize that the 1-2-3 user interface can be described as incorporating more than one idea. Thus, the Paperback opinion identifies the idea of an electronic spreadsheet, the idea of having a readily available method of invoking the menu command system (which I concluded merged with its expression as the command), and the idea of the two-line- moving cursor menu, among others. 3. Baker Not Alone Controlling.- In its most recent submissions, Borland once again insists that 17 U.S.C. § 102(b) and Baker v. Selden control this case, now asserting that the Second Circuit decision supports Borland’s argument. Once again, Borland overstates the point. Borland is correct in observing that the Second Circuit treats Baker as the “starting point” in analyses of utilitarian works, including computer programs. That view is compatible with my earlier Memorandum and Order. See Borland, 788 F.Supp. at 92-93. Borland fails to note, however, that the Second Circuit went on to state: While Baker v. Selden provides a sound analytical foundation, it offers scant guidance on how to separate idea or process from expression, and moreover, on how to further distinguish protectable expression from that expression which “must necessarily be used as incident to” the work’s underlying concept. Computer Assocs., 1992 WL 139364, at *11, No. 91-7893, slip op. at 24, 1992 U.S.App. LEXIS 14305, at *32. The Second Circuit’s decision cannot fairly be characterized as holding that Baker controlled the outcome in Computer Associates. The Second Circuit was sensitive not only td its duty of fidelity to precedent, but as well to its duty of fidelity to congressional mandates that came into existence long after Baker was decided. 4. “Substantial Similarity” and “Copyrightability.” The Second Circuit referred to its test as one of “substantial similarity” and did not use the term “copyrightability” for any part of the test. I do not understand this difference of terminology to have substantive implications, however, and more especially, not for this case. Nor do I understand the Second Circuit’s three-step test to be . meant as a rigid barrier to alternate methods of analysis and decision. The court “advised” rather than mandated that the courts within the Second Circuit adhere to the three-step test announced. Moreover, a court in the First Circuit must take account of some degree of dissonance between the First and Second Circuits regarding the methodology of infringement analyses. Compare Computer Assocs., supra, with Concrete Mach. Co. v. Classic Lawn Ornaments, Inc., 843- F.2d 600 (1st Cir.1988). Although Concrete Machinery outlines a methodology different from the three steps of the Second Circuit’s test, I do not understand the First Circuit to have mandated the order of analysis it described. See Borland, 788 F.Supp. at 86. I adhere to the view stated in my Memorandum and Order of March 20 that, under existing precedents, prudential concerns about case management may have a bearing on the order in which a court proceeds with its analysis in a complex copyright infringement case. Id. Thus, I conclude that I may and should proceed to a copyrightability analysis at this time even though some potentially material factual questions are still unresolved — facts regarding the long prompts, and facts regarding the extent to which the menu commands and command hierarchy may have been dictated by functional considerations. E. Applying the Standard 1. The First Step. FIRST, in making the determination of “copyrightability,” the decisionmaker must focus upon alternatives that counsel may suggest, or the court may conceive, along the scale from the most generalized conception to the most particularized, and choose some formulation, some conception of the “idea,” “system,” “process,” “procedure,” or “method”— for the purpose of distinguishing between the idea, system, process, procedure, or method and its expression. One may describe a number of conceptions of the 1-2-3 user interface. A non-exclusive list, commencing with the most abstract and moving toward the particular, includes: (1) Lotus 1-2-3 is an electronic spreadsheet. (2) It is a menu-driven electronic spreadsheet. (3) Its user interface involves a system of menus, each menu consisting of less than a dozen commands, arranged hierarchically, forming a tree in which the main menu is the root/trunk of the tree and submenus branch off from higher menus, each submenu being linked to a higher menu by operation of a command. (4) Its user interface involves a system of menus, each menu consisting of less than a dozen commands, arranged hierarchically, forming a tree in which the main menu is the root/trunk of the tree and submenus branch off from higher menus, each submenu being linked to a higher menu by operation of a command, so that all the specific spreadsheet operations available in Lotus 1-2-3 are accessible through the paths of the menu command hierarchy. (5) Finally, one may conceive of the interface as that precise set of menu commands selected by Lotus, arranged hierarchically precisely as they appear in 1-2-3. Under this conception, the interface comprises the menu of commands “Worksheet,” “Range,” “Copy,” “Move,” “File,” “Print,” “Graph,” “Data,” “System,” and “Quit,” linked by operation of the command “Worksheet” to the menu of commands “Global,” “Insert,” “Delete,” “Column,” “Erase,” “Titles,” “Windows,” “Status,” and “Page,” etc. (The completion of this proposed statement of the “idea,” listing all of the more than 400 commands for which “etc.” stands, would require several dozen more lines of text.) Borland argues that the appropriate conception of the “idea” of the 1-2-3 interface is the fifth option. If that were the case, of course, there would be no elements of expression in the menu commands and menu command hierarchy and therefore no copyrightable aspects in them. The premise of Borland’s argument is that an “idea” of Lotus 1-2-3 version 2.01 is complete compatibility with earlier versions of 1-2-3, and more precisely with macros generated for use with earlier versions. Borland argues that the precise menu commands and menu structure are necessary to such functional compatibility. Thus, the argument goes, the entire interface of version 2.01 is a functional system or “idea” and is not copyrightable. This argument is essentially tautological. As applied to any case involving a useful article, an argument of this kind would always define the idea to incorporate all the specifics of the particular expression of that idea in the allegedly copyrightable work. Nothing would be copyrightable under this methodology of analysis. The argument is an attempt to win by definition without focusing at any time on any substantive issue concerning the separation of idea and expression. To select, at the opposite extreme, the very abstract statement of the idea of 1-2-3 as “an electronic spreadsheet” would be to draw an inappropriately abstract boundary between idea and expression. Thus, I concur in a fundamental principle of the Computer Associates opinion and reject the contrary proposition in Whelan. Arguably, my Opinion in the Paperback decision, where no sharper focus was essential to the outcome, is consistent with accepting a conception of the idea that falls between the second and third formulations above. See Paperback, 740 F.Supp. at 67 (electronic spreadsheet having “menu structure”). In any event, I now explicitly recognize that for decision of the issues now before me the selection of functional operations that the spreadsheet performs must be considered part of the idea of the program. Copyrightability depends on expression distinct from the selection of the set of spreadsheet operations that can be performed. I conclude that an appropriate conception of the “idea” or “system” of the 1-2-3 interface is the fourth of the five alternative conceptions stated above. 2. The Second Step. SECOND, the decisionmaker must focus upon whether an alleged expression of the idea, system, process, procedure, or method is limited to elements essential to expression of that idea, system, process, procedure, or method (or is one of only a few ways of expressing the idea, system, process, procedure, or method) or instead includes identifiable elements of expression not essential to every expression of that idea, system, process, procedure, or method. Does the Lotus 1-2-3 user interface include identifiable elements of expression? For reasons stated below, I conclude that it does. A very satisfactory spreadsheet menu tree can be constructed using different commands and a different command structure from those of Lotus 1-2-3. In fact, Borland has constructed just such an alternate tree for use in Quattro Pro’s native mode. Even if one holds the arrangement of menu commands constant, it is possible to generate literally millions of satisfactory menu trees by varying the menu commands employed. This may be easily demonstrated. Recall the ten commands that appear in Lotus’ main menu: “Worksheet,” “Range,” “Copy,” “Move,” “File,” “Print,” “Graph,” “Data,” “System,” and “Quit.” One can imagine an entirely plausible spreadsheet in which the “Worksheet” command has been named, quite naturally, “Spreadsheet.” Of course, this might require changing the “System” command to avoid two commands abbreviated “S,” perhaps to “DOS.” The “Quit” command could be named “Exit” without any other modifications. The “Copy” command could be called “Clone,” “Ditto,” “Duplicate,” “Imitate,” “Mimic,” “Replicate,” and “Reproduce,” among others (in some cases requiring modification of other commands in the menu). Additional possibilities include “Output” for “Print,” “Draw” or “Chart” for “Graph,” “Figures” or “Information” for “Data,” “Scope” for “Range,” and “Transfer” or “Relocate” for “Move.” Just these potential modifications of the main menu yield over 250 combinations of commands in the main menu with ten distinct first letters. Changes in submenus increase the number of possible menu hierarchies geometrically. Since there are dozens of independent submenus, the number of possible menu hierarchies is extremely large. Borland argues that “[t]o hold that an idea, plan, method or art described in a copyright is open to the public but that it can be used only by the employment' of different words and phrases which mean the same thing, borders on the preposterous.” Borland’s Resp. to PL’s Renewed Mot. for Summ. J. (Docket No. 183) at 19 (quoting Crume v. Pacific Mut. Life Ins. Co., 140 F.2d 182, 184-85 (7th Cir.), cert. denied, 322 U.S. 755, 64 S.Ct. 1265, 88 L.Ed. 1584 (1944)). This case, however, unlike Crume, is not a case in which the system “can be effected solely by the employment of words descriptive thereof.” Crume, 140 F.2d at 184. Use of just the initial letters of command words (together with long prompts) or of other symbolic tokens would have been a sufficient alternate method of implementing the system. In this case, the command words chosen are not necessary to expression of the system nor are they necessarily incident thereto. See Computer Assocs., 1992 WL 139364, at *13, No. 91-7893, slip op. at 30, 1992 U.S.App. LEXIS 14305, at *40. Lotus argues that a large number of substantially different arrangements (hierarchies) could also have been effected. Looking again at just the main menu, is there any reason that the commands “Copy” and “Move,” for example, could not have been arranged in the opposite order? Borland argues that the arrangement was necessary, citing evidence that Lotus arranged the menu commands in order of the expected frequency of use. It is clear that certain command words have been grouped according to function; e.g., there are eighteen commands that affect the display mode of spreadsheet cells that are grouped together under the “Format” command. Thus, a jury could find that at least some aspects of the arrangement of command words, as opposed to the specific choice of command words, was guided by functional concerns. This is, however, a disputable fact question that may affect only the scope of relief in this case. See ABKCO Music, Inc. v. Harrisongs Music Ltd., 508 F.Supp. 798 (S.D.N.Y.1981) (court awarded damages based upon contribution to success of infringing work of copyrighted material) (for subsequent history, see 944 F.2d 971 (2d Cir.1991)). I conclude that it cannot be genuinely disputed that a large part of the structure and arrangement of the menu commands is not driven entirely by functional considerations. There are sufficient non-functional aspects that at least hundreds and perhaps thousands of different expressions of the function were possible when Lotus chose the particular structure of menu commands incorporated into Lotus 1-2-3. This may be demonstrated by examining more closely Borland’s argument that the menus were arranged according to the predicted frequency of use of the commands. I assume the truth of Borland’s assertion that Lotus predicted before marketing its spreadsheet that the “Copy” command would be used more often than the “Move” command. Nevertheless, that is merely a prediction of frequency of use. It did not require Lotus to list “Copy” before “Move.” A user can type a “C” or an “M” with equal ease no matter which command is listed first. If a user prefers to invoke a command by first highlighting it and then typing “Enter,” “Move” is only one keystroke from “Copy”; moreover, the same “right arrow” key that takes the cursor from “Worksheet” to “Copy” moves the cursor from “Copy” to “Move.” Thus, the order in which commands are listed in a menu has very limited functional value. In addition, a prediction of frequency of use depends upon who is the predicted user and what the predicted uses are. For example, a user may work on a spreadsheet without printing any of the work performed on that day. Yet, the user will ordinarily invoke the “Quit” command appearing at the end of the main menu at the end of each session. Nevertheless, the “Print” command appears before the “Quit” command on the menu. Many users may rarely invoke the “System” command available in 1-2-3. That is, the need to invoke a disk operating system (“DOS”) “shell” from within the spreadsheet program may be for most users of 1-2-3 a rare event. If the commands in the main menu of 1-2-3 are listed in order of predicted frequency of use, why does the “System” command “precede” the “Quit” command? In fact, what does it mean to say “precede”? “Q” neither precedes nor follows “S” on the keyboard. “System” precedes “Quit” as one moves from left to right within the main menu using the “right arrow” key. It is also true that “Quit” precedes “System” as one moves from right to left using the “left arrow” key. The 1-2-3 menus are circular (in jargon, they have a “wraparound” feature)— moving the cursor one step beyond the “end” of the menu causes the cursor to come to rest at the opposite “end.” Thus, from the default cursor position on the “Worksheet” command, arguably “Quit” precedes “System.” For all these reasons, any ex ante prediction of frequency of use is itself of limited usefulness. It follows that the arrangement of menu commands according to predicted frequency of use is not a major functional limitation on the number of arrangements of menu commands. The menu command hierarchy is an integral part of the functionality of the macros and of the keystroke sequences. Nevertheless, the fact that the macros and keystroke sequences incorporate the menu command hierarchy into their functionality does not remove the menu command hierarchy from the scope of copyright, if otherwise subject to copyright protection. Moreover, the macros and keystroke sequences are protected to the extent that it is necessary to infringe a copyright to use them. Of course, as I have stated above, it was not necessary to copy expressive aspects of the macro language and keystroke sequences to copy their function. 3. The Third Step. THIRD, having identified elements of expression not essential to every expression of the idea, system, process, procedure, or method, the decisionmaker must focus on whether those expressive elements, taken together, are a substantial part of the allegedly copyrightable “work.” The question posed by this element of the copyrightability test is whether the creativity involved in establishing the menu commands, menu command hierarchy, macro language, and keystroke sequences was more than trivial. No reasonable jury could find otherwise. As Borland has itself acknowledged, at least implicitly, Lotus 1-2-3 was a dramatic change and improvement over what was available on the market at the time Lotus was created. Although a large portion of that improvement relates to the functional aspects of Lotus 1-2-3, the features that I have now concluded are expressive also played a substantial role. Borland has maintained that those features are part of an uncopyrightable system (an argument I now reject for the reasons stated), but Borland has never argued that they were trivial, nor could it do so persuasively. F. Ashton-Tate v. Ross Borland argues that the Ninth Circuit’s decision in Ashton-Tate Corp. v. Ross, 916 F.2d 516 (9th Cir.1990), aff'g 728 F.Supp. 597 (N.D.Cal.1989) militates in favor of a conclusion that the menu commands and command hierarchy are not copyrightable. The Ninth Circuit held that a list of menu commands was not copyrightable for reasons stated by the district court. The district court, in turn, held that a document bearing “a list of labels for user commands, many of which are common commands that were already available on other software programs” was not innovative or novel. Ashton-Tate Corp. v. Ross, 728 F.Supp. 597, 602 (N.D.Cal.1989). The relevance of that conclusion to this case is in some doubt in view of the fact that Lotus 1-2-3 is one of the programs on which the commands were already available. There is no evidence in this case that the commands available in 1-2-3 where common commands at the time of Lotus’ authorship. The district court also held that there was nothing innovative in the order in which the commands were listed. Id. Borland has submitted in this case a copy of the list of commands at issue in Ashton-Tate. Borland’s Mem. in Supp. of Cross-Mot. for Summ. J. (Docket No. 141) at 85-86. That list contains well under one hundred commands scrawled on one sheet of paper segregated into under one dozen functional groups. In form, detail, arrangement, and content, it bears almost no resemblance to the hundreds of menu commands arranged in Lotus’ dozens of menus. The relevance of the court’s conclusion to the case at bar extends, if at all, only to the copyrightability of subelements of the 1-2-3 menu hierarchy, and depends upon factual inquiries that remain to be made. Finally, the court concluded that in the absence of any contribution to the user interface other than the command list (including, especially, any contribution to the program code), the contribution of the command list was a mere contribution of ideas. As Judge Learned Hand observed, however, and many others have agr