dissenting, in which HATCHETT, Chief Judge, and BARKETT, Circuit Judge, join:
The district court understood this case. It held that Warren’s compilation of selected data concerning cable television operations, in the form of data-reporting units with each unit named for a principal community within the unit, was original and creative. Its decision should be affirmed.
I. Introductory
The Copyright Clause of the Constitution provides that Congress has the power to secure to authors “the exclusive Right to their respective Writings.” U.S. Const, art. I, § 8, cl. 8. Therefore originality — authorship — is a constitutional requirement. By 17 U.S.C. § 102 Congress provided for copyright protection to original works of authorship. 17 U.S.C. § 103 provides that § 102 includes compilations. 17 U.S.C. § 101 defines compilation:
A “compilation” is a work formed by the collection and assembling of preexisting materials or of data that are selected, coordinated, or arranged in such a way that the resulting work as a whole constitutes an original work of authorship.
Thus, originality is also a statutory requirement.
Warren’s compilation was held by the district court to be protected as an original work of authorship “selected” pursuant to § 101. It is a work containing data on cable television operations nationwide, issued annually in book form, entitled Television and Cable Factbook, and the volume in question is the 1988 edition. It contains collected data selected and assembled into reporting units each of which comprises a functional “cable system,” which Warren defines as: “an entity composed as one or more communities that are offered the same service by the same cable system owner at the same price.” Each “cable system” bears the name of a “lead” or “principal” community within the system. That name identifies the cable system, and data for the system is presented under that name.1 To simplify the evidence the parties have accepted that evidence concerning cable television operations in the state of Illinois is representative.
Understanding Warren’s compilation, and this case, requires one to understand that cable television service exists by authority of franchises granted by organized governmental units, usually cities and counties. The district court found how, prior to Warren’s work, the various compilers of industry data commonly compiled and arranged information concerning cable television operations:
Warren Publishing admits that the cable system information coordinated and arranged by the various compilers in the industry is commonly organized alphabetically by state and then alphabetically by community within the states.
Dist.Ct.op. p. 10. This common form of organizing and presenting data is not surprising since franchises to operate sprang from discrete communities. Moreover, this accords with Federal Communications Commission definitions. FCC defines a “cable television system” as a facility that provides cable service to subscribers “within a community.” 47 C.F.R. § 76.5(a). Also, it defines a “cable television system” as one that “operates ... within a separate and distinct community.” 47 C.F.R. § 76.5(dd). Type and extent of service, rates, commonality of service with other communities, and sharing of facilities or equipment or staff or management are not elements of FCC definitions of cable systems.
As the industry developed innumerable new cable operations were franchised and activated, some contiguous to existing franchises, others disassociated and far distant from previously franchised communities, *1522some operators with a single franchise, others with more than one. Over time cable operations were sold, merged, expanded in area, mechanical equipment was shared, and staff and servicing combined or shared. Geographic areas of service changed. But the industry norm for selecting and presenting data remained the community.
Against this background one must examine what, in a general sense, a compiler does and, in a specific sense, what Warren as compiler did. The creator of a compilation responds to a perceived need for information, and that response may be a highly creative act but at this initial stage it is only an idea and clearly not copyrightable. William S. Strong, Database Protection After Feist v. Rural Telephone Co., 42 J. Copyright Soc’y U.S.A. 39, 47 (1994). Responding to the perceived need the compiler must choose the facts it wants and devise a framework for the data to be assembled, which includes formulating rules and identifying categories that may be highly selective but are not necessarily so. Id. Categories desired may be limited or dictated by their utility or by the marketplace and hence involve no originality, or they may be original to the compiler. It is at this identification/formulation of categories stage that the compiler moves from uncopy-rightable idea to acts of selection that are the expression of his ideas.
Warren grasped the “perceived need for information” reflecting the present nature of the cable television industry and the past practices of the industry for selecting and presenting data. It then chose the facts it wanted to compile. The Supreme Court has recognized this choice of facts as part of a compiler’s authorship: “The compilation author typically chooses which facts to include, in what order to place them, and how to arrange the collected data so that they may be used effectively by readers.” Feist Publications, Inc. v. Rural Service Co., Inc., 499 U.S. 340, 348, 111 S.Ct. 1282, 1289, 113 L.Ed.2d 358 (1991). See also Key Publications, Inc. v. Chinatown Today Publishing Enters., Inc., 945 F.2d 509, 513 (2d Cir.1991), “Selection implies the exercise of judgment in choosing which facts from a given body of data to include in a compilation.”
At the initial stage of choosing the facts that it wanted Warren moved from idea to intellectual expression through selection. The selection of facts it wanted were not the facts that previously the industry had compiled in terms of community. Rather Warren chose to select and present facts that reflected the way the industry is currently actually operating. Its choice was reflected in functional service/operations/management terms. The building block, the data-reporting unit, for selection and presentation of industry data was the “cable system” as newly defined by Warren, “an entity composed as one or more communities that are offered the same service by the same cable system owner at the same price.” Warren had, as Strong, supra, has described it, devised a framework for the data to be assembled and had formulated selective rules and categories. Reporting data by a functional unit was a new and original concept, and the implementing definition of “cable system” was new to the industry and crafted by Warren.
Next, it was necessary for Warren to define and identify the universe of raw data from which it would select and present information. It chose a universe composed of all geographic communities (in the state of Illinois, the representative state) having cable television service. This defined universe was itself new. It consisted of 1,000 plus geographic communities (1,017 by one count, 1,045 by another). It included cities, towns, and villages, and also included counties and townships, which historically were not usual franchise-granting units. The FCC maintained its own list of cable systems (as it defined them), composed of cities, towns and villages, that is, franchise-granting units. FCC’s universe was 724 communities. Warren’s functional/operational definition swept in non-franchising geographic areas receiving service. Its universe of raw data was thus new in concept and some 40% larger in number of communities than the FCC universe.
As its next step Warren identified and selected from its universe 406 data-reporting units in Illinois, each a “cable system” pursuant to its functional definition. Then, drawing from the 1,000 plus universe, Warren had *1523to identify and properly locate 'within the proper unit of the 406 each geographic community enjoying cable service. More than half of the 406 Illinois cable operations turned out to be single-system operations (SSO’s), that is, each served only a single community. A multiple community system (MSO) served more than one community. Each SSO, because of its singularity, fell within Warren’s same operator/same service/same price definition of a cable system. The name under which its data was presented was necessarily that of the single community it served. Having located within the proper cable system (MSO or SSO) each community served, for MSO’s Warren had to merge or combine the operating data for each community into one unitary body of operating data to be reported for the system. Data relating to each geographic community served was no longer independently listed community-by-community but instead was included in the unitary system data. The name of an individual (nonprincipal) geographic community whose service was operated and managed as part of a cable system appeared but without data and was cross-referenced to the system where its data was included in the unitary data. This referencing was necessary, of course, because data-reporting was unitary rather than individual.
As part of Warren’s acts of selection it was necessary for it to choose a name by which each cable system would be listed and identified and under which the system data would be set out. For this purpose Warren elected to use a geographic name, and the type of geographic name it chose was that of the “lead” or “principal” geographic community within the system. Obviously, for an SSO the name of the single community served was selected. When these acts were concluded Warren’s selection (406 units) consisted of 45% fewer data-reporting units than the FCC’s listing of 724.
II. Originality and creativity
We are faced in this case with what Feist described as the “undeniable tension” between two well-established propositions— that facts themselves are not copyrightable but compilations of facts generally are. 499 U.S. at 344-45, 111 S.Ct. at 1286-87. A compilation draws its originality from its selection and arrangement.
Factual compilations, on the other hand, may possess the requisite originality. The compilation author typically chooses which facts to include, in what order to place them, and how to arrange the collected data so that they may be used effectively by readers. These choices as to selection and arrangement, so long as they are made independently by the compiler and entail a minimal degree of creativity, are sufficiently original that Congress may protect such compilations through the copyright laws. Nimmer § § 2.11[D], 3.03; Denicola 523, n. 38. Thus, even a directory that contains absolutely no pro-tectable written expression, only facts, meets the constitutional minimum for copyright protection if it features an original selection or arrangement. See Harper & Row [v. Nation Enterprises], 471 U.S., [539] at 547, 105 S.Ct., [2218] at 2223, [85 L.Ed.2d 588]. Accord, Nimmer § 3.03.
Id. at 348, 111 S.Ct. at 1289. The originality necessary to render Warren’s work copyrightable lies in its selection of data as provided by § 101. The selection must be done “in such a way” that it possesses the necessary originality.
What does “originality” mean? The selection must be made independently by the compiler, not copied, and must owe its origin to the author. Novelty is not required. But selection must entail a minimal degree of creativity.2
How much originality is required? Feist tells us: “a modicum of intellectual labor,” 499 U.S. at 347, 111 S.Ct. at 1288; “independent creation plus a modicum of creativity,” id. at 346, 111 S.Ct. at 1288; “at least some minimal degree of creativity,” id. at 345, 111 S.Ct. at 1288; “the requisite level of creativity is extremely low; even a slight amount will suffice,” id. at 345, 111 S.Ct. at 1288. *1524Nimmer expresses the degree of originality this way:
It has been said that all legal questions are in the last analysis questions of degree, requiring judicial line drawing. Certainly, copyright law is replete with such questions. The determination of the quantum of originality necessary to support a copyright presents such a question. It is not, however among the more troublesome questions of degree inherent in copyright law, as the line to be drawn includes almost any independent effort on the side of sufficient originality.
Melville B. Nimmer and David Nimmer, Nimmer on Copyright § 2.01[B], at 2-13 (1996). And
[Ojriginality for copyright purposes amounts to ... little more than a prohibition of actual copying.
Nimmer, § 2.01B[, p. 2-14 (quotes and internal quotes omitted)]. Accord, Key Publications, supra, at 513.
Warren’s selection entails more than the required degrees of originality and creativity. Warren saw the heed, chose the facts it wanted to compile, chose how it wanted to arrange them in gathering points for data rather than by individualized presentations. It employed a new concept of gathering cable data into a smaller number of units and, for this purpose, it devised a new concept of a cable system as functionally defined and a new concept (and new title) of “principal community.” It is sufficient if there is a “small spark of distinctiveness,” but this is no small spark. It is a fundamental change in reporting data of a changing and developing industry. The fact that some of the data-reporting units were SSO’s does not diminish the fact of Warren’s acts of selection or of the originality and creativity of the selection, which required Warren to determine whether each of the 1,000 plus systems was a single community system (SSO) or part of a multiple community system (MSO) and to assemble and report system data accordingly-
Along with originality of selection and arrangement is a related but different requirement. The Constitution authorized protection of the work of an author. The claimant to copyright protection must be an author, not a mere discoverer of facts.
No one may claim originality as to facts. Facts may be discovered, but they are not created by an act of authorship. One who discovers an otherwise unknown fact may well have performed a socially useful function, but the discovery as such does not render him an “author” in either the constitutional or statutory sense.
Nimmer, § 2.11[A], p. 2-172.16 (footnotes omitted). But by hypothesis a compiler collects and assembles the work of others, and his compilation is a “work formed by the collection and assembly of preexisting material.” See 17 U.S.C. § 101.
In the tension between facts and compilation of facts there are some facts that cannot trigger copyright-ability. In a narrow range of circumstances facts themselves may be of such character that a work relating to them is incapable of meeting the requirement of a “work of authorship” referred to by the Constitution and by the statute. These might be called “public domain facts,” known to or available to the world at large. Feist, at pp. 347-48, 111 S.Ct. at pp. 1288-89, refers to census data, scientific and historical and biographical facts, and news of the day. Professor Nimmer refers to scientific facts as to the nature of the physical world, historical facts, and contemporary news events. Nimmer, § 2.03[F], at p. 2-36. Regulations covering “Registration of Claims to Copyright” provide in 37 C.F.R. § 202.1:
Material not subject to copyright
(d) Works consisting entirely of information that is common property containing no original authorship, such as, for example: Standard calendars, height and weight charts, tape measures and rulers, schedules of sporting events, lists or tables taken from public documents or other common sources.
Another narrow range of facts do not fit neatly within the “public domain” category but nevertheless are so obvious or trivial that no creativity will transform mere selection of them into copyrightable expression. Nim-mer § 201[B], p. 2-14. In Feist, the tele*1525phone company’s white page directory alphabetically listed telephone users by name, town and telephone number. 499 U.S. at 362, 111 S.Ct. at 1296. The subject matter was not original with- the telephone company, and the company’s use of the facts through alphabetical listing was not only unoriginal but practically inevitable. Id. at 363, 111 S.Ct. at 1297. The Supreme Court “ultimately reversed [in Feist ] on the ground that plaintiffs white pages directory was not copyrightable at all.” Jane C. Ginsburg, “No Sweat ”? Copyright and Other Protections of Information after Feist v. Rural Telephone, 92 Colum. L.Rev. 338, 342. In BellSouth Adv. & Pub. Corp. v. Donnelley Info. Pub., Inc., 999 F.2d 1436 (11th Cir.1993) (en banc), BAPCO’s heading structure, for example, “Attorneys” or “Banks”, represented such obvious labels for the entities appearing beneath that they lacked the required originality for copyright protection.
III. The opinion of this court
Apart from two lesser points discussed in Parts V and VI below, the opinion of this court has these main premises:
(1) The Factbook does not come within the “selection” prong of the § 101 definition of a “compilation” because no selection has been made (by anybody), since the Factbook lists all geographic communities having cable service. (Mss. pp. 22-26.)
(2) Assuming that the Factbook is sufficiently creative and original to be copyrightable, Warren’s claim of protection fails because:
(a) Warren seeks copyright protection for mere techniques for discovery of facts.
(b) The selection of principal communities was made by cable operators and not by Warren. Therefore, Warren does not meet the constitutional requirement that it be the “author” of the compilation, rather it is engaged in mere discovery of facts.
I take these up in the above sequence.
(1) The premise that the Factbook contains no selection at all because it lists the universe composed of all geographic communities having cable service.
It is puzzling that this argument is seriously advanced. It is a play on words such as “listing” and “including” and it confuses the universe of data with the data drawn from the universe. As noted in the opinion of this court, (Mss. p. 26), the district court itself recognized that a list of a universe (in that reference, FCC’s universe) would of itself not be original and, therefore, not copyrightable. Warren claims no copyright on the universe, and the district court found no copyright on such a list. Warren claims, and the district court found, a copyright on the selection of data drawn from the universe. Definition of a universe of data was an essential initial step in selection, but no claim is made that the universe by itself is copyrightable as a selection. The district court’s references to “selected communities” are plainly references to communities as selected and presented through Warren’s 406 cable-system data-presenting units.
Key Publications, supra, tells us that selection implies the exercise of judgment in choosing which facts from a given body of data to include in a compilation. 945 F.2d at 513. There the compiler’s universe consisted of a multitude of businesses that she thought of interest to Chinese-Amerieans. The in-fringer urged that the compiler had made no selection but had included every business of which she had information. The court found that she had excluded businesses she thought might not remain open for very long, and this alone indicated the necessary thought and creativity. Id. The compiler did not list the universe, only the selected businesses. This court suggests that Warren has no copyright protection because it “included” its universe as well as its selected data. This misconceives the work of selection. All communities were selected, some identified and located in MSO’s, others in SSO’s.
(2) The.premise that, assuming that the Factbook is sufficiently creative and original to be copyrightable, Warren’s claim of protection fails because:
(a) The premise that, as in BellSouth, Warren seeks copyright protection for mere techniques for discovering facts.
This is a baffling premise. In BellSouth the district court had described acts that *1526BAPCO performed as alleged “acts of selection” — geographic limits, closing dates for entries, requiring yellow page subscribers to use business telephone service, and use of marketing techniques such as free listings and on-site visits. On appeal this court found that the district court had erred in not considering whether these alleged acts of selection met the level of originality, therefore it examined the acts. 999 F.2d at 1441. This court then held that through these strategies and marketing techniques BAPCO had learned that subscribers described then-businesses in particular manners in yellow page listings and would pay for listings under certain business categories. The strategies and techniques used by BAPCO were not selected facts at all in the copyright sense but were merely creative means used to discover the facts it wanted to learn, merely industrious means for collecting data. These “uncopyrightable formative acts used to generate [the] listings were not entitled to copyright protection.” Id. at 1441.
Warren seeks no copyright on the means it used to find out facts. It has no strategies or marketing techniques. As acts of selection it collected facts in the old-fashioned way.3 It collects data from trade publications, FCC records and reports, newspaper and magazine clipping sendees. Each year it sends “thousands and thousands” of questionnaires to over 10,000 cable operators in the country, which are used to identify changes from the preceding year and to update. If a system does not respond or responds inadequately Warren telephones the operator to obtain update data. It follows leads to new systems. A staff of over 20 people spend the entire year gathering data, inputting, checking, conferring and updating. It confirms with some operators the geographic areas they are currently serving. It contacts some operators to inquire what community is considered to be the lead or principal community. These are all fact-gathering techniques. None is claimed to enjoy copyright protection.
(b) The premise that the selection of principal communities was made by cable operators and not by Warren, therefore, Warren does not meet the constitutional requirement that it be the “author” of the compilation, rather it is engaged in mere discovery of facts.
First, as a matter of fact did Warren delegate to operators the choice of principal communities? The district court found “[t]he principle community used to represent the entire system, is then selected by contacting the cable operator to determine which community is considered the lead community within the cable system.” (Op. 10.) This court draws upon that statement to conclude that Warren has made no selection of principal communities or that it accepts as conclusive operators’ consideration of what are principal communities of their respective cable systems. This single sentence by the district court does not bear the weight of this court’s conclusions. The phrase “contacting the cable operator to determine” the principal community was used by a witness. Elsewhere a witness explained that a call might be made to an operator to determine in conjunction with the operator the identity of the principal community. Moreover, this court’s conclusions are inconsistent with Mi-crodos’ position. Before the panel Microdos asserted that the choice of the principal community is controlled by external objective factors. It urged that the principal community is the one with the greatest number of subscribers. It has suggested the principal community is the “largest,” which it infers to be the most populous because, Microdos says, it will generate the most subscribers. It has asserted that the lead community is the site of the headend (the location of equipment used to process television signals for redistribution to cable subscribers).4 Also, *1527Microdos has strenuously urged that Warren draws the identity of the principal community from data it finds in Federal Communications Commission reports. Additionally, in its petition for rehearing en banc it has called the court’s attention to Atlanta as the principal community for its cable service, chosen, Microdos says, because it is the dominant municipal area served and everybody knows that it is the principal community. It is obvious that these objective factors are relevant to determining the identity of the principal community of a cable system, and Micro-dos accordingly has relied upon them.5 But this court has laid all these aside as having no significance, indeed as though never uttered, in favor of its own conclusion that it is the cable operators, not Warren, that determine in the case of a multiple community systems, the community name under which to list the pertinent data for the entire cable system. (Mss. p. 33.)
Alternatively, this court proposes (n. 29) that a cable operator is “likely” to designate as its principal community the community with the most subscribers, therefore no exercise of judgement was required by Warren to select the principal communities. I have pointed out the many factors asserted by Microdos itself as relevant to selection of the principal community. Neither the record, nor Microdos, supports the “likelihood” that the principal community will be the one with the most subscribers, nor the statement (n. 26) that every MSO has one “obvious” principal community. That may be so as to Atlanta (“everybody knows it is the principal community”), but a study of some 406 principal communities in Illinois, most of them smaller towns and cities, discloses no such “obvious” character.
This court (Mss. p. 31) analogizes Warren’s contacts with operators to what it describes as the “similar” selection of headings made by telephone users in BellSouth. But, as BellSouth noted, the headings offered to BAPCO subscribers did not originate with BAPCO but were obvious and unoriginal labels for business categories such as “Attorneys” or “Banks.” 999 F.2d at 1445. The BAPCO subscriber ordered from an unoriginal menu of business categories the menu item it liked and would pay to be listed under in the yellow pages. Warren’s category, “principal community,” is neither obvious nor unoriginal. The cable operator in this case was asked for operational information about how his business was currently functioning, to be listed in an operations directory. The extensive objective factors advanced by Mi-crodos itself demonstrate relevant criteria that bear on this industrial directory listing.
Moreover, this court has focused upon the selection of the principal community, whose name the system will bear, as though that is all that the case is about. The acts of selection carried out by Warren were a stream of events, beginning with its choice of the facts it wanted and the construct of a functional methodology in which to develop and present them. The use of a geographic name for each cable system, and the choice of the names of principal communities as identifiers, and the decision on a particular name, were not isolated acts of selection like Athena springing full grown from the brow of Zeus, or a decision made by a snap of someone’s fingers, or a mechanical decision from a single telephone call, or by numerous calls. They were parts of the stream of acts of selection that I have described. This court does not, however, refer to Warren’s exercise of judgment in creating this structure of selection and in choosing the facts to be reported and how to report them. Yet these acts of selection are independent expressions of the author, part of the overall “work of authorship.”6 This court does not hold them *1528to be unoriginal or non-creative. Instead, it ignores them and treats this ease as turning on the single fact of the source of information about principal communities. This trivializes what this case is about.
Laying aside the foregoing, I turn to this court’s conclusion that Warren does not meet the requirements of authorship because it is a “mere discoverer” of facts. The difference between mere discovery of facts by Warren and authorship by Warren cannot be based on the single fact that Warren is engaged in collecting information. All compilers are collectors of facts collected from some other source. If the fact of collecting data from an original source deprives a compiler of authorship status, all the vitality is drained out of the congressional provision for copyright of compilations in §§ 101 and 102. Warren’s status as author versus mere discoverer requires examination of the nature of the facts discovered. If they are “public domain” facts, or such facts as by their nature cannot support originality, Warren is not an author. If, however, Warren has collected facts that are capable of supporting originality, and it meets the statutory requirements for selection and presentation, then both statute and the constitutional provision for originality (authorship) are met. The linchpin of Feist is the nature of the underlying facts (names, towns and telephone listings) that would not support copyrightability. BellSouth has the same linchpin: headings, such as “Attorney” and “Banks” that were not original expressions of an author but mere facts, obvious if inevitable classifications drawn from the public domain. In both cases, Feist and Bell-South, the compiler was gathering information that would not support copyrightability.
Warren’s facts do not fit into these narrow categories of uncopyrightable facts. Data on how businesses in a growing and changing industry are owned, operated and managed is not public' matter like today’s news event, or the speed of a falling object, or the face of the calendar, nor is it unoriginal subject matter open to and utilized by the world at large like the telephone listings of Feist. Its facts are functional data of a changeable and changing industry, structured in a new and original format. Its gathering of these facts from original sources is authorship, not mere discovery.
U.S. Payphone, Inc. v. Executives Unltd. of Durham, Inc., 18 U.S.P.Q.2d 2049, 1991 WL 64957 (4th Cir.1991), is a per curiam with Justice Powell on the panel. The compiler assembled and summarized public information on state tariffs regulating fees payable to telephone utilities by owners and operators of pay telephones. The summarized information was presented in the format of one sheet for each state. The court found:
Payphone’s selection and organization of the state tariff material was sufficiently subjective and original to make the Tariff Section copyrightable material.
18 U.S.P.Q.2d at 2051, 1991 WL 64957.
IV. BellSouth does not ring for this case
It is understandable that judges of this court wish to be faithful to the en bane decision in BellSouth. But we need not extend it. I gather in one place the reasons this decision is not controlled by BellSouth.
(1) BellSouth concerned uncopyrightable facts, obvious headings drawn from the public domain. This case does not.
(2) The district court in BellSouth found that the compiler’s establishment of the geographic limits of its directory and of a closing date for listings were acts of selection. This court found these were uncopyrightable acts common to compilations. 999 F.2d at 1441. There are no such acts in the present case.
(3) In BellSouth this court found that the district court had erred in treating as copy*1529rightable facts that were not copyrightable facts at all but merely techniques for the discovery of facts — marketing techniques and sales strategies. Warren has no such techniques and strategies and it relies upon selection and presentation of facts.
(4) Much of the BellSouth decision concerns BAPCO’s claims of originality based upon the coordination and arrangement provisions of § 102. 999 F.2d at 1442-44. These issues are not present in this case. The district court ruled against Warren on coordination and arrangement, and this holding is not an issue in this appeal.
(5) BAPCO failed to establish that its structure of headings was “original expression,” that is, that it was the author of the headings such as “Attorneys” and “Banks.” Without question Warren is the creator of the heading “principal community,” a name previously unknown to the industry and implicating the concept of a data-reporting unit previously unknown. BellSouth found that an expressive act of dividing such obvious categories as “Attorneys” into subeategories (such as bankruptcy lawyers and criminal lawyers) merged into the idea of listing in a directory the subtitles as a class of business. 999 F.2d at 1444. There are no such subdivisions in this case, and, as discussed below in Part VI, merger does not fit, indeed does not even come close, to this case.
(6) At the heart of BellSouth is the single fact of the subscriber’s selecting an appropriate heading from the menu of obvious and unoriginal headings. The selection involved in the present case is a stream of original and nonobvious acts only one aspect of which involves choice of names for data-reporting units.7
V. Use of the word “system” does not bar eopyrightability
The district court used the word “system” in referring to Warren’s acts of selection, and the panel opinion by this court fell into the same phraseology. 17 U.S.C. § 102(b) tells us that copyright protection for an original work of authorship does not extend to a “system.” This court relies upon § 102[b] as a bar to eopyrightability. Neither district court nor the panel addressed § 102(b), and one may infer that both courts used the word “system” in a generic, everyday sense and not as a word of art under § 102(b).
In any event, § 102(b) is not, as this court describes it, a “limiting principle.” In the leading case, Toro Co. v. R. & R. Products Co., 787 F.2d 1208, 1212 (8th Cir.1986), the claimant asserted a copyright on its use of a “system” of numbering in its catalog replacement parts for lawn care machines. The district court denied eopyrightability on the ground that the claim was for a “system.” The court of appeals rejected the view that literal use of the term “system” from § 102(b) is a “limiting principal.”
[Section 102(b)] is nothing more than a codification of the idea/expression dichotomy as it developed in the case law prior to passage of the 1976 Act. H.R.Rep. No. 1476, 94th Cong., 2d Sess. 57, reprinted in 1976 U.S.Code Cong. & Ad.News 5659, 5670 (“Section 102(b) in no way enlarges *1530or contracts the scope of copyright protection under the present law. Its purpose is to restate ... that the basic dichotomy between expression and idea remains unchanged.”) (Emphasis added.)
Id. at 1212. The court held the claimant’s copyright not valid, based on the idea/expression dichotomy; i.e., claimant could not copyright the idea of using numbers to designate replacement parts. And its expression of that idea simply drew numbers from the public domain and, without rhyme, reason, or judgment, arbitrarily assigned them to parts. The expression of the idea did not meet the requirement of originality. See Nimmer § 203(D), p. 2-35, to the same effect as Toro. It seems beyond argument that Warren does not seek copyright protection on the idea of gathering and selecting data and reporting it in a manner that responds to the perceived needs for functional data of a changeable and changing industry. Rather it seeks a copyright on its expression of that idea.
The opinion of this court recognizes that the use of the term “system” does not preclude copyrightability. But, the court says, Warren’s acts of selection were in fact a “system,” and that fact creates a bar. (n. 21). In the first place, the district court made no such finding, nor does the evidence address it. Second, this contention is contrary to what Congress itself has said. See quotation, above, H.R. Report No. 1476. If what the copyright claimant has done is an expression of sufficient originality that it is entitled to copyright, calling it a “system” does not strip it of copyrightability.
VI. The doctrine of merger does not bar copyrightability
By footnote this court suggests the merger doctrine as an alternative ground for denying copyrightability. Merger operates where there is only one or so few ways of expressing an idea that protection of the expression would effectively accord protection to the idea itself. The court suggests that Warren’s principal community presentation is one, if not the only, commercially useful way of organizing a compilation of information on the cable television industry, so the presentation is merged with the idea of a cable television directory.
This is another puzzling point. As has been said repeatedly, Warren’s reporting data by principal community units is a total departure from prior methods utilized in the industry. No one — industry or government — has previously assembled and presented functional data drawn from the way the industry presently operates and is managed. Presumably, for many users, Warren’s way of selecting and presenting data is the most useful way. But there are many ways.
The television industry is driven by the advertising dollar, and advertisers place their dollars by numbers and types of viewers, based in part on information that includes numbers of homes reached by cable. Seekers of cable data may wish to utilize data compiled community-by-community, as, for example, an equipment salesman will wish to know whether discrete communities within a service area utilize differing equipment. A compiler may wish to organize cable television data by counties, by areas of the state (“upstate” and “downstate”), by adjoining communities, by agricultural areas, by urban and rural areas, by big systems and smaller systems, large cities and small towns, high income and low income areas, sports-oriented areas and less interested areas. Nor does cross-referencing change the picture. Warren cross-references and groups data by service/management. Another compiler may group and cross-reference agricultural areas or high income areas. It may group and cross-reference all cable operations that use a particular manufacturer of equipment.
Warren’s selection of data is original, creative and useful. To suggest it is the only conceivable useful way is astonishing. The FCC listed 724 communities in Illinois versus Warren’s selected 406, based on different criteria. The Broadcast Yearbook, another recognized directory of the industry, listed 243 communities in Illinois. Different organizations create lists different in structure, scope and number that may be useful for different readers for varying purposes. Merger does not fit.
*1531VIL Conclusion
The district court correctly decided this case, and we should affirm its decision. Our statutes provide rational and economically useful copyright protection for compilations. If that protection is to be narrowed and cabined the choice is for Congress, not the courts.
. As the district court succinctly put it, “how one defines a 'cable system’ will dictate the communities selected to represent those systems [i.e., the principal communities].'' Dist.Ct.op. p. 10.
. It is of only semantic significance whether originality is defined as embodying creativity or whether creativity is regarded as a necessary adjunct to originality. But it is clearer to refer to them as separate elements. Nimmer, § 201[B], p. 2-15.
. This court suggests (n. 27) that Warren's acts of selection merely show industriousness, which is not relevant to copyrightability. To the contrary, Warren’s acts of selection are examined, just as this court in BellSouth examined BAPCO’s acts of selection, 999 F.2d at 1441, to determine whether those acts met the level of originalily to extend copyright protection.
. Television signals may be received by satellite, by microwave tower, or by telephone lines from television stations. Microwave towers generally are located on high ground that may be unrelated to other facilities of the operator and not necessarily even in the area served by the cable system. Many systems have multiple headends. It has been suggested that “lead" community *1527means the site of the managerial headquarters of the system, where a customer, salesmen for cable equipment, or a potential advertiser may seek the manager or the engineer or the sales manager. But managerial headquarters is not necessarily even in the area served, and examination of Illinois systems in the Factbook shows that frequently it is not.
. Other relevant factors are miles of cable and numbers of homes passed in a particular community.
. The selections made of categories — "cable system,” "principal community,” — are by themselves, acts of selection that meet requirements of originality. CCC Information Services, Inc. v. Maclean Hunter Mkt. Reports, Inc., 44 F.3d 61 (2d Cir.1994), held copyrightable a compilation *1528of the compiler’s predictions of used car valuations based upon market data and the compiler’s judgment and expertise. One of the elements of originality held to pass Feist's threshold was the use of the abstract concept coined by the compiler, of the "average” vehicle in each category. Id. 44 F.3d at 67. The Second Circuit held the compilation was protected and that the district court erroneously applied a higher standard of originality than Feist. Kregos v. Associated Press, 937 F.2d 700 (2d Cir.1991), concerned a form that displayed statistics on the recent past performances of baseball pitchers scheduled to start the next day’s games. The compiler’s selection of statistics survived summaiy judgment motion alleging lack of originality.
. While I do not suggest that BellSouth be abandoned, it has drawn considerable criticism. Wood, Ethan L., Copyrighting the Yellow Pages: Finding Originality in Factual Compilations, 78 Minn. L.Rev. 1319, 1335 (1994): "The Eleventh Circuit's approach directly contradicts Key Publications” [discussed above in text].... "The Eleventh Circuit opinion in BellSouth is much more hostile than Key Publications to claims of copyright infringement of the yellow pages.” Id. at 1333. "The Eleventh Circuit’s BellSouth decision used a standard of originality that is inconsistent with the Supreme Court's approach in Feist.” Id. at 1336. "IT]he Eleventh Circuit has raised the threshold of required originality higher than the Feist decision established." Id. at 1337. And finally, "[U]nlike the Eleventh Circuit, the Second Circuit, which has traditionally been the most influential in developing copyright law, properly follows the Feist approach.” Id. at 1339 (footnote omitted).
See also Nimmer § 3.04[B], p. 3-31 (footnotes omitted):
Most applications of Feist have recognized the circumscribed sphere to which its holding applies, ruling that it invalidates the copyright only in the most banal of works, such as the white pages of a copybook.
Other post-Feist decisions cannot be squared with BellSouth. See CCC Information Services, Inc. v. Maclean Hunter Mkt. Reports, Inc., 44 F.3d 61 (2d Cir.1994), in footnote 5, supra.
See also U.S. Payphone, Inc. v. Executives Unltd. of Durham, Inc., 18 U.S.P.Q.2d 2049, 1991 WL 64957 (4th Cir.1991), discussed above in Part III.